Topical References Assignment Example | Topics and Well Written Essays - 1250 words

Topical References - Assignment Example The motivation behind this examination is to comprehend the different kinds of disarranges that are secured inside the setting of Child and Adolescents Psychiatry while the exploration is based to respond to this inquiry which sort of mediations might be used to upgrade efficiency of youngsters and youths experiencing emotional well-being issues and furthermore how these scatters sway the profitability. References Angold, A. what's more, Costello, E. J. (2000). The Child and Adolescent Psychiatric Assessment (CAPA). Diary of the American Academy of Child and Adolescent Psychiatry 39 (1): 39â€48. Dynamic: To portray the Child and Adolescent Psychiatric Assessment (CAPA). The base meeting covers the age run from 9 to 17 years. The meeting glossary that gives itemized operational meanings of side effects and seriousness evaluations is portrayed, and psychometric information and further advancements of the meeting are introduced. Across 5,962 parent-kid meets, the center areas of the CAPA (mental indications, utilitarian weakness, socioeconomics, family structure and working) took on normal 59 minutes for kids and 66 minutes for guardians. Test-retest unwavering quality for analyze ran from ? = 0.55 for lead issue (CD) to ? = 1.0 for substance misuse/reliance. Legitimacy as decided by 10 unique rules was acceptable. Advancements of the CAPA incorporate a shorter â€Å"gateway† rendition utilizing center side effects as screen things, a Spanish adaptation, and forms for twin investigations, use with youthful grown-ups (YAPA), and preschool-age kids (PAPA). There is a spot in both research and clinical settings for a thoroughly operationalized meet, (for example, the CAPA) that centers around guaranteeing that respondents comprehend what is being solicited and on unmistakably characterizing levels from side effect seriousness and utilitarian impedance. [Abstract from author] Classi, P. M., Le, T. K., Ward, S., and Johnston, J. (2011). Quiet attributes, como rbidities, and medicine use

A Theory of Dividends Dissertation Example | Topics and Well Written Essays - 10500 words

A Theory of Dividends - Dissertation Example The fundamental profit arrangements for all intents and purposes embraced by the greater part of the organizations worldwide can be essentially partitioned into four classes. Consistent profit framework best the rundown followed by steady profit proportion framework, profits equivalent to free income framework and clingy profit framework (Internet, Introduction to Dividend Policy and Theory, two issues, 1). In the steady profit framework, organizations deliver whole consistent sum as profit each year regardless of the benefit go while in the consistent proportion framework, they receive a steady proportion on earnings.â The last is otherwise called the EPS or DPS meaning income per offer or profit per share. In the third framework, quantum of free income manages the installment. In the last one, otherwise called the Lintner Model, organizations receive a clingy strategy in which profits are first paid in a low level and afterward expanded after certain years. This is essentially do ne to upgrade the certainty of speculators and shareholders. Dividend theoriesWhy should organizations announce profits? It is the duty of the administrations to demonstrate some money related advantage to the financial specialists who have contributed their reserve funds. Organizations lead their business with the speculations offered by the investors and it is their bounden obligation to offer profits consequently to them. In any case, as the portions of the organizations are exchanged the financial exchanges, it is normal for showcase players to connect the profits to the offer costs. The facts demonstrate that there are a few organizations that don't deliver profits.

Your College Application in 650 Words or Less

Your College Application in 650 Words or Less Last August, Sarah Vander Schaaff interviewed me about working with college students on their application essays. Here is her article, edited for The Essay Expert and updated for 2015-16! In 650 words or less, describe a place or environment where you are perfectly content. What do you do or experience there and why is it meaningful? Or In 650 words or less, recount an incident or time when you experienced failure. How did it affect you, and what lessons did you learn? Or In 650 words or less, discuss an accomplishment or event, formal or informal, that marked your transition from childhood to adulthood within your culture, community, or family. You are now sitting in the shoes of a high school senior, thank you very much. These questions are a few of the prompts on the 2015-16 Common Application for college. They are slightly changed from last years’ says Brenda Bernstein, the professional resume writer, personal statement coach, and business copywriter behind the Essay Expert, and for some students facing the questions, answering them is hardly simple. “High school students don’t all have life coaches standing by,” Brenda told me when I spoke to her on the phone this summer. “They don’t step back and reflect on their lives.” Sometime between August and five days before applications are due, Brenda gets a call for help. If the student has already written an essay, she takes a look at it and determines the level of help it might needâ€"anything from some minor fixes to  an overhaul.  And it’s here when Brenda goes back to the basic creed that good writing requires good thinking. “By the end of the first conversation they have an outline for a new essay and new thoughts. I have the students tell me what they see as the way the essay is going to go.” Bernstein says students struggle when they “… know stuff that happened in their lives and they’ve drawn broad conclusions or they have conclusions that they think they should draw.” Parents divorcing, for example, is a major life event that sometimes deserves a deeper look. “They know they got a lesson from it,” she said. But the true lesson is not always the one they initially thought. Bernstein says she got an early start thinking about college applications. Her father interviewed applicants for Yale as part of the school’s alumni committee. Later, he’d let her read the interview reports. “I would read a lot of them,” she said. While at Yale herself, friends applying to graduate schools often asked her to edit their essays. Later, while working as a public interest lawyer in New York, she started posting flyers offering her editing services. And in a move that would make a pretty good essay itself, she decided she didn’t want to practice law, sought advice from an astrologer, and moved to Madison, Wisconsin where she still lives seven years later. Her site offers help not only for students, but also for professionals seeking a better resume or LinkedIn profile. The Essay Expert’s essay consultations range in price from $697 for Common Application essay help to $247 for a personal statement reviewâ€"far less than some boot camps described in The New York Times that cost $14,000 for 4 days. Essays, Bernstein says, are 6th on a list of 10 criteria colleges look at for admissions. “If students can’t write this type of essay without help, should they get into the college?” I asked Bernstein. She said pretty much all students get help, whether that’s from a parent, a teacher, or a paid professional (26 percent employing the latter). In school, students also have access to writing help if they need it. “A smart studentâ€"and professionalâ€"will get input and support on every important piece of writing they submit, especially when so much is riding on it. What’s important is that they write it in their own voice.” Furthermore, explains Bernstein, “I find that as they work with me they become much better writers.” Part of what she teaches them, she added, was to look at what about their lives and experiences fit together as a story. What The Essay Expert is doing appears to be the counterpoint to the technique explained by Frank Bruni in his June opinion piece in The New York Times, Naked Confessions of the College-Bound: Oversharing in Admissions Essays. Bruni tells of one Yale applicant’s essay.  A highly qualified student wanted to highlight her complete absorption with French and her teacher, “she described their one-on-one conversation at the end of a school day. And then, this detail: During their talk, when an urge to go to the bathroom could no longer be denied, she decided not to interrupt the teacher or exit the room. She simply urinated on herself.” Bruni cites similar examples, and quotes author Sally Rubenstone, who calls this type of behavior, “…the Jerry Springer-ization of the college admissions essay.” Coaches like Bernstein can help rein in the conflicting messages our children receive when they mistake the shock-factor for good writing. But the real work for these essays begins long before the fall of senior year. As parents, we need to give our children moments to reflect on their own lives, and genuine experiences to mull about as individuals without the constant pressure to document every minute of it. When I taught public speaking, I asked students to give a biographical speech touching on three major aspects of their lives. Many talked about a sport, or a hobby, or a pet. But almost all included this as part of those three defining elements of biography: a family tradition. The next time you fret about not giving your kid every opportunity to fill his or her resume, think about what it takes to be able to reflect on one’s own life. As the Essay Expert asks, “How well did you know yourself at the end of high school?”

Gender Roles Of Women And Women - 1536 Words

Men and women have very different expectations and roles in the work force. In the past almost all corporations have discriminated against gender, especially women. Whether it is women not being allowed certain positions and/or promotions at a job, or women not being allowed into the field of work that they desire at all. Women have also been discriminated by being payed lower wage and made to work longer hours in unsafe and unhealthy work environments. Gender roles set expectations that are extremely incorrect. A woman is just as capable at any job that a man is and should not be discriminated against because of her gender. Culture and society have a huge impact on gender roles, specifically among women. â€Å"Among the ancient Celts women rulers and warriors were so common that when a group of Brigantian captives was brought to Rome in the reign of Claudius they automatically assumed his wife†¦Ã¢â‚¬  (Irish, Shawn). Women in some cultures like the ancient Celts had women in leadership roles such as governing the people of their land. However, some cultures see women as nothing more than someone to care for the children and keep the house clean. â€Å"Before World War II, the division of workers into different jobs on the basis of gender was universally accepted† (Irish, Shawn). Meaning that society found it morally right that you could not allow women certain jobs due solely to the fact that they were women. Also, in ancient Rome, a common job for women was to work as a â€Å"hawker†. HawkingShow MoreRelatedGender Roles Of Women And Women1524 Words   |  7 PagesGender refers to the masculine and feminine attributes assigned to either sex. Culture and biology determine the moral, intellectual, and emotional differences between males and females. Biology partially determines gender constructions of men and women. A woman’s ability to conceive or a man’s natural strength are all examples of biological gender. However, I believe a larger proportion of personal gender is formulated through cultural believes and values. Cultures create gender through social practicesRead MoreGender Roles Of Women And Women1535 Words   |  7 PagesEveryday you hear women and men being compared in life. A majority of the time it is in a form of men being more superior than women rather it s stereotypes, appearances, or money; Men usually end up more dominant. Society tends to assume that men are more capable than women in all aspects throughout life. These postulations have commenced before we were even born. Men are expected to constantly be working and providing the home for the women and children; Women are assumed to be cleaning, cookingRead MoreGender Roles Of Women And Women1630 Words   |  7 Pagesallow freedom for women, freedom for men, freedom from those sharply defined gender roles â€Å" (Fred Ward). I agree with this quote because in today’s society gender defines the roles of what a male and a female can do. What is gender? Gender, according to the Merriam Webster dictionary, is the state of being male or female (typically used with reference to social and cultural differences rather than biological ones). The term gender is only used when we describe how men and women are different in aRead MoreGender Roles Of Women And Women1677 Words   |  7 PagesGender is one of the greatest factors that can determine how we define ourselves and others. Although gender is determined at birth for many, it does not influence one s identity of self until a later age. How society views the roles of each gender influences how a perso n views themselves. As the views changes, so does one’s identity. In the past, the roles of men and women were defined and separated. Men were expected to provide income for the family, while women were expected to take care of theRead MoreGender Roles Of Women And Women1512 Words   |  7 Pages1.a Gender roles are how a society believes a man or women should act based on their gender. Each society has its own gender roles, where a man might be expected to do something in one society a woman might be expected to do it in another one. An example of gender roles affecting women is that women are expected to take care of the housework and so they do chores around the house even if they have a job. Men are expected to be the breadwinner and support his family financially. Men are not expectedRead MoreGender Roles Of Women And Women892 Words   |  4 Pages Gender roles are a set of behaviors that indicate that a person should fit in to a category of either being male or female . This is society’s concept of how men and women are expected to behave. These gender roles were commonly accepted before the twenty-first century. Society viewed males as being the main financial providers for their family whereas women were viewed as the maker of the home, raising and taking care their children. Men were expected to be strong, athletic,aggressive,and dominantRead MoreGender Roles Of Women And Women2066 Words   |  9 PagesThroughout history, gender roles were portrayed based on their physical appearance and the ability to do things. Men were always considered the caretakers and protectors of their homes and never had to worry about what kind of outfit they had to wear the next day. Men were free to roam around in whatever clothing that appealed to them and no one seemed to care. However, in the case of women, they were the ones who did the house work all day and if she wanted to get an education that was somethingRead MoreGender Roles Of Women And Women23 59 Words   |  10 PagesMen and women are consistently sexualized and misrepresented in the media. These misrepresentations have serious implications on society and both sexes are constrained by these ideals. These ideal depictions of masculine and feminine are threatening to the already distorted gender roles. The media standards for beauty are outrageous and simply unattainable, putting immense pressure on women today. Moreover media targeting men present a constant theme of hyper masculinity that is normalized inRead MoreGender Roles Of Women And Women1662 Words   |  7 Pages Gender Roles â€Å"Gender is the costume, a mask, a straightjacket in which men and women dance their unequal dance† (Kearl). Throughout history, women have been treated with unequal and unjust standards in comparison to their male counterparts. This paper will discuss gender roles in society and the stereotypes placed on men and women. As well as gender role development and the different influences that lead to these roles and stereotypes. Also, it will outline the history and progress of women’sRead MoreGender Roles Of Women And Women876 Words   |  4 PagesA gender role is simply a behavior that may be learned by a person that is so called â€Å"accepted† into society or a cultural normality. For example, men are known for supporting the family the most such as keeping the food on the table and being most of the income that comes into the household. Women on the other hand are normally looked upon as house cleaners, childcare givers, and less responsible jobs inside the house. Gender stereotypes are a little bit different; this is a way of seeing someone

Constant Critics of Monopolies Free Essays

Monopolies are under constant critics from the public and other producers of being polutive, straining to competition and they are accused of worsening resource allocation. Whether this is true or not, depends on the specific company, but certain characteristics are possible to define. It is these I will describe in the following, and hence conclude if monopolies worsen or improve It is important to distinguish between competition and monopoly before describing advantages and disadvantages of both. We will write a custom essay sample on Constant Critics of Monopolies or any similar topic only for you Order Now Many monopolies are government owned. This means that the incentive to trive for more profit, better conditions etc. is gone. This is due to the fact that, if there is a loss, the government will cover it, and government owned companies seldom strive to achieve maximum profits. A lot of the characteristics are also seen in privately owned monopolizing firms. When they become so big, that competition is practically gone, the incentive to make even more profits, and In a competitive industry this is not the case. The fear of loosing your job, not being able to compete, your products becoming obsolete etc. re important factors, which stimulate productivity. It is therefor obvious that the competitive industry will try harder to allocate their resources in the most efficient way. To land, the external costs in a competitive industry will often be pollution, seeing that the firm will strive hard to diminish their costs resulting in the firm ignoring ‘unnecessary’ costs. The monopoly owned by the government, would never be able to ignore such a serious matter, and they would have to pay the costs. A monopoly would also have to be careful not to damage its image, seeing that is, in many cases, already is unpopular. Capital, on the other hand, is often to the benefit of a monopoly, since they produce at a large scale. To fully utilize capital, a lot of labour is needed, labour which a monopoly is expected to have, and a smaller competitive firm may lack. For example, a blast furnace might need a crew of 24 men working night and day, to fully utilize it. The monopolizing company may be able to provide the men, but the smaller firm might not have the money to hire all the 24 men at night, seeing wages are much higher at then. The question then is if the competitive company is so much more efficient due to ard work, that they still can produce more than the monopoly. When it comes to labour, it is obvious that a competitive industry will strive to utilize the workers at a maximum level, due to the desire of minimizing costs, and workers will in general be very efficient due the reasons mentioned above. The workers in a monopoly, often loose the feeling, that their work makes a difference in the firm, making it hard for managers to fully In my opinion, the characteristics described above are not as valid any more. Companies, which enjoyed monopoly status in the eginning of the 80’s, like IBM, are now realizing that nothing lasts forever, and they have be innovative, even if the competition is not a great threat. Bill Gates, owner of MicroSoft, has very admirable policies concerning this. His firm is not a monopoly, but it is definably a cutting-edge firm, which is shaping the future. One rule he has, is that every six months the bottom five percent of the company’s workforce (in terms of performance) get firedo. It is his goal to make his own products obsolete, not letting others do it, and it seems he is achieving that goal. How to cite Constant Critics of Monopolies, Essay examples

Tom Clancy Essays - Tom Clancy, Red Storm Rising, Red Storm

Tom Clancy English Tom Clancy's genius The Cold War and post Cold War eras have brought with them many interesting aspects. New technologies initially meant for mass destruction filter down into the civilian world, making current lives easier. One example of this is the anti-lock braking systems of today's cars. Originally designed to slow fighter-planes on landing without skidding, these systems make it safer for parents to take their children on vacation. One less noted advancement the eras brought is a considerable amount of exciting and forewarning fiction. While most authors chose to warn of nuclear and post nuclear holocaust, one significant author chose a different approach. Tom Clancy chose to write of conventional warfare and sometimes unconventional enemies. Between his novel Red Storm Rising and Debt of Honor, Tom Clancy makes evident the changing face of America's enemies and threats, while staying true to issues that keep people interested in his books. Published in 1986, Red Storm Rising is Tom Clancy's second novel dealing with the former Soviet Union as a potential enemy. This was a time when America's finest tank and infantry units went on exercises in Germany fully armed with the expectation that the Russians could attack them at any time. This was also a time when the Soviets did the same exercises with the same amount of live ammunition. Therefore there was reason enough to worry about potential conflicts. Deep within the ocean waters, submarines played similar cat and mouse games with other submarines and surface ships. However some of these submarines were more dangerous then a whole army because they were fully loaded with nuclear missles. These facts were well know to the American public and made Red Storm Rising all the more real when it combined land and ocean warfare in a way that captivated millions of readers. The book begins as the Soviet Union's ability to provide their own oil is cut off by a terrorist attack. Right away it is no ted that two very frightening events have just happened. Terrorism, for one, is a major scare tactic that can and does strike fear into millions. This was demonstrated by two suspected attacks in the U.S. recently (Bombing of Flight 800 and the Olympic Park bombing). Secondly, the threat of losing petroleum resources is enough to drive governments to drastic measures. This fact is evident in the world's participation in the 1991 Gulf War. The leaders of the Soviet Union decided that the only way to prevent the total collapse of their economy and country was to seize the oil rich Middle East. They also realized that the countries that make up the North Atlantic Treaty Organization (NATO), in particular the United States would not stand for this hostile action. Consequently the Soviets determine that it will be necessary to neutralize NATO conventionally; that is to say without nuclear weapons. Of course, throughout the Cold War the many themes of the U.S.S.R. attacking the U.S a re presented by various authors. All of these had the same result: nuclear holocaust. One exception is that Red Storm Rising is the first to present it (theme of U.S.S.R. attacking the U.S.) in a non-nuclear scenario. This is very intriguing to examine the possibilities which include all the new technological weapons in the American and Soviet arsenals. Red Storm Rising captivates audiences with its techno-wizardry of smart bombs and satellite guided cruise missles. ?It was like an arcade game. Big, slow-moving blips denoted the aircraft. Smaller, quicker blips were the Mach-2 missiles (Clancy 178).? This was seen by a radar operator who was under attack during Red Storm Rising. However it is not the high tech gadgets that appeal to audiences of Red Storm Rising. There is a personable feel as the reader becomes better acquainted with the characters and sympathizes for them and the decisions they make. This is not the story of machines run by artificial intelligence, these are r eal people, friends, and neighbors of the reader. Bob Toland was a middle-level analyst at the National Security Agency. He'd left the Navy after six years whey the adventure of uniformed service had palled, but he remained an active reservist. His work at NSA dovetailed nicely with his

Apply for the CPP Retirement Pension

Apply for the CPP Retirement Pension The application for the Canada Pension Plan (CPP) retirement pension is quite simple. However, there are a lot of things to learn and decide  before you apply. What is the CPP Retirement Pension? The CPP retirement pension is a government pension based on workers earnings and contributions. Just about everybody over the age of 18 who works in Canada (except in Quebec) contributes to the CPP. (In Quebec, the Quebec Pension Plan (QPP) is similar.) The CPP is planned to cover about 25 percent of pre-retirement earnings from work. Other pensions, savings and interest income are expected to make up the other 75 percent of your retirement income. Who is Eligible for a CPP Retirement Pension? In theory, you must have made at least one valid contribution to the CPP. Contributions are based on employment income between a set minimum and maximum. How much and how long you contribute to the CPP affects the amount of your pension benefits. Service Canada maintains a Statement of Contributions and can provide an estimate of what your pension would be if you were eligible to take it now. Register for and visit My Service Canada Account to see and print a copy. You can also get a copy by writing to: Contributor Client ServicesCanada Pension PlanService CanadaPO Box 9750 Postal Station TOttawa, ON K1G 3Z4 The standard age to start receiving a CPP retirement pension is 65. You can receive a reduced pension at the age of 60 and an increased pension if you delay starting your pension until after the age of 65. You can see some of the changes that are taking place in the reductions and increases in CPP retirement pensions in the article Canada Pension Plan (CPP) Changes. Important Considerations There are numerous situations that can affect your CPP retirement pension, and some may increase your pension income. Some of those are: Child rearing provision  can be requested if you stopped working or received a lower income as the primary caregiver of your children under the age of seven, which could increase your retirement pension.Pension sharing with your spouse or common-law partner could mean tax savings for you.Credit splitting  after a divorce or separation  allows CPP contributions made by you and your spouse or common-law partner to be equally divided.International social security agreements  may make you eligible for a pension if youve lived and worked in certain countries. How to Apply for the CPP Retirement Pension You must apply for the CPP retirement pension. It is not automatic. For your application to be eligible You must be at least a month past your 59th birthdayYou must have contributed to the CPPYou must want your pension payments to begin within 11 months. You can apply online. This is a two-part process. You can submit your application electronically. However, you must print and sign a signature page that you then must sign and mail to Service Canada. You could also print and complete the ISP1000 application form and mail it to the appropriate address. Dont miss the detailed information sheet that comes with the application form. After You Apply for the CPP Retirement Pension You can expect to receive your first CPP payment approximately eight weeks after Service Canada receives your application. Service Canada has other  useful information to be aware of once you start receiving your benefits.

Yeats and The Symbolism of Poetry

Yeats and 'The Symbolism of Poetry' One of the greatest poets of the 20th century and a recipient of the Nobel Prize, William Butler Yeats spent his early childhood in Dublin and Sligo before moving with his parents to London. His first volumes of poetry, influenced by the symbolism of William Blake and Irish folklore and myth, are more romantic and dreamlike than his later work, which is generally more highly regarded. Composed in 1900, Yeatss influential essay The Symbolism of Poetry offers an extended definition of symbolism and a meditation on the nature of poetry in general. The Symbolism of Poetry    Symbolism, as seen in the writers of our day, would have no value if it were not seen also, under one disguise or another, in every great imaginative writer, writes Mr. Arthur Symons in The Symbolist Movement in Literature, a subtle book which I cannot praise as I would, because it has been dedicated to me; and he goes on to show how many profound writers have in the last few years sought for a philosophy of poetry in the doctrine of symbolism, and how even in countries where it is almost scandalous to seek for any philosophy of poetry, new writers are following them in their search. We do not know what the writers of ancient times talked of among themselves, and one bull is all that remains of Shakespeares talk, who was on the edge of modern times; and the journalist is convinced, it seems, that they talked of wine and women and politics, but never about their art, or never quite seriously about their art. He is certain that no one who had a philosophy of his art, or a theory of ho w he should write, has ever made a work of art, that people have no imagination who do not write without forethought and afterthought as he writes his own articles. He says this with enthusiasm, because he has heard it at so many comfortable dinner-tables, where some one had mentioned through carelessness, or foolish zeal, a book whose difficulty had offended indolence, or a man who had not forgotten that beauty is an accusation. Those formulas and generalisations, in which a hidden sergeant has drilled the ideas of journalists and through them the ideas of all but all the modern world, have created in their turn a forgetfulness like that of soldiers in battle, so that journalists and their readers have forgotten, among many like events, that Wagner spent seven years arranging and explaining his ideas before he began his most characteristic music; that opera, and with it modern music, arose from certain talks at the house of one Giovanni Bardi of Florence; and that the Plà ©iade laid the foundations of modern French literature with a pamphlet. Goethe has said, a poet needs all philosophy, but he must keep it out of his work, though that is not always necessary; and almost certainly no great art, outside England, where journalists are more powerful and ideas less plentiful than elsewhere, has arisen without a great criticism, for its herald or its interpreter and protector, and it may be for this reason that great art, now that vulgarity has armed itself and multiplied itself, is perhaps dead in England. All writers, all artists of any kind, in so far as they have had any philosophical or critical power, perhaps just in so far as they have been deliberate artists at all, have had some philosophy, some criticism of their art; and it has often been this philosophy, or this criticism, that has evoked their most startling inspiration calling into outer life some portion of the divine life, or of the buried reality, which could alone extinguish in the emotions what their philosophy or their criticism would extinguish in the intellect. They have sought for no new thing, it may be, but only to understand and to copy the pure inspiration of early times, but because the divine life wars upon our outer life, and must needs change its weapons and its movements as we change ours, inspiration has come to them in beautiful startling shapes. The scientific movement brought with it a literature, which was always tending to lose itself in externalities of all kinds, in opinion, in declamation, in pic turesque writing, in word-painting, or in what Mr. Symons has called an attempt to build in brick and mortar inside the covers of a book; and new writers have begun to dwell upon the element of evocation, of suggestion, upon what we call the symbolism in great writers. II In Symbolism in Painting, I tried to describe the element of symbolism that is in pictures and sculpture, and described a little the symbolism in poetry, but did not describe at all the continuous indefinable symbolism which is the substance of all style. There are no lines with more melancholy beauty than these by Burns: The white moon is setting behind the white wave,And Time is setting with me, O! and these lines are perfectly symbolical. Take from them the whiteness of the moon and of the wave, whose relation to the setting of Time is too subtle for the intellect, and you take from them their beauty. But, when all are together, moon and wave and whiteness and setting Time and the last melancholy cry, they evoke an emotion which cannot be evoked by any other arrangement of colours and sounds and forms. We may call this metaphorical writing, but it is better to call it symbolical writing, because metaphors are not profound enough to be moving, when they are not symbols, and when they are symbols they are the most perfect of all, because the most subtle, outside of pure sound, and through them one can the best find out what symbols are. If one begins the  reverie  with any beautiful lines that one can remember, one finds they are like those by Burns. Begin with this line by Blake: The gay fishes on the wave when the moon sucks up the dew or these lines by Nash: Brightness falls from the air,Queens have died young and fair,Dust hath closed Helens eye or these lines by Shakespeare: Timon hath made his everlasting mansionUpon the beached verge of the salt flood;Who once a day with his embossed frothThe turbulent surge shall cover or take some line that is quite simple, that gets its beauty from its place in a story, and see how it flickers with the light of the many symbols that have given the story its beauty, as a sword-blade may flicker with the light of burning towers. All sounds, all colours, all forms, either because of their preordained energies or because of long association, evoke indefinable and yet precise emotions, or, as I prefer to think, call down among us certain disembodied powers, whose footsteps over our hearts we call emotions; and when sound, and colour, and form are in a musical relation, a beautiful relation to one another, they become, as it were, one sound, one colour, one form, and evoke an emotion that is made out of their distinct evocations and yet is one emotion. The same relation exists between all portions of every work of art, whether it be an epic or a song, and the more perfect it is, and the more various and numerous the elements that have flowed into its perfection, the more powerful will be the emotion, the power, the god it calls  among  us. Because an emotion does not exist, or does not become perceptible and active among us, till it has found its expression, in colour or in sound or in form, or in all of the se, and because no two modulations or arrangements of these evoke the same emotion, poets and painters and musicians, and in a less degree because their effects are momentary, day and night and cloud and shadow, are continually making and unmaking mankind. It is indeed only those things which seem useless or very feeble that have any power, and all those things that seem useful or strong, armies, moving wheels, modes of architecture, modes of government, speculations of the reason, would have been a little different if some mind long ago had not given itself to some emotion, as a woman gives herself to her lover, and shaped sounds or colours or forms, or all of these, into a musical relation, that their emotion might live in other minds. A little lyric evokes an emotion, and this emotion gathers others about it and melts into their being in the making of some great epic; and at last, needing an always less delicate body, or symbol, as it grows more powerful, it flows out, with all it has gathered, among the blind instincts of daily life, where it moves a power within powers, as one sees ring within ring in the stem of an old tree. This is maybe what Arthur OShaughnessy meant when he made his poets say they had built Nineveh with their sighing; and I am certainly never certain, when I hear of some war, or of some religious excitement or of some new manufacture, or of anything else that fills the ear of the world, that it has not all happened because of something that a boy piped in Thessaly. I remember once telling a seer to ask one among the gods who, as she believed, were standing about her in their symbolic bodies, what would come of a charming but seeming trivial  labour  of a friend, and the form answering, the devastation of peoples and the overwhelming of cities. I doubt indeed if the crude circumstance of the world, which seems to create all our emotions, does more than reflect, as in multiplying mirrors, the emotions that have come to solitary men in moments of poetical contemplation; or that love itself would be more than an animal hunger but for the poet and his shadow the priest, for unless we believe that outer things are the reality, we must believe that the gross is the shadow of the subtle, that things are wise before they become foolish, and secret before they cry out in the  market-place. Solitary men in moments of contemplation receive, as I think, the creative impulse from the lowest of the Nine Hierarchies, and so make and unmake mankind, and even the world itself, for does not the eye altering alter all? Our towns are copied fragments from our breast;And all mans Babylons strive but to impartThe grandeurs of his Babylonian heart. III The purpose of rhythm, it has always seemed to me, is to prolong the moment of contemplation, the moment when we are both asleep and awake, which is the one moment of creation, by hushing us with an alluring monotony, while it holds us  waking  by variety, to keep us in that state of perhaps real trance, in which the mind liberated from the pressure of the will is unfolded in symbols. If certain sensitive persons listen persistently to the ticking of a  watch,  or gaze persistently on the monotonous flashing of a light, they fall into the hypnotic trance; and rhythm is but the ticking of a watch made softer, that one must  needs  listen, and various, that one may not be swept beyond memory or grow weary of listening; while the patterns of the artist are but the monotonous flash woven to take the eyes in a  subtler  enchantment. I have heard in meditation voices that were forgotten the moment they had  spoken; and  I have been swept, when in more profound meditatio n, beyond all memory but of those things that came from beyond the threshold of waking life. I was writing once at a very symbolical and abstract poem, when my pen fell on the ground; and as I stooped to pick it up, I remembered some  phantastic  adventure that yet did not seem  phantastic, and then another like adventure, and when I asked myself when these things had happened, I found, that I was remembering my dreams for many nights. I tried to remember what I had done the day before, and then what I had done that morning; but all my waking life had perished from me, and it was only after a struggle that I came to remember it again, and as I did so that more powerful and startling life perished in its turn. Had my pen not fallen on the ground and so made me turn from the images that I was weaving into verse, I would never have known that meditation had become trance, for I would have been like one who does not know that he is passing through a wood because his eyes are on the pathway. So I think that in the making and in the understanding of a work of art, and the mo re easily if it is full of patterns and symbols and music, we are lured to the threshold of sleep, and it may be far beyond it, without knowing that we have ever set our feet upon the steps of horn or of ivory. IV Besides emotional symbols, symbols that evoke emotions alone,and in this sense all alluring or hateful things are symbols, although their relations with one another are too subtle to delight us fully, away from rhythm and pattern,there are intellectual symbols, symbols that evoke ideas alone, or ideas mingled with emotions; and outside the very definite traditions of mysticism and the less definite criticism of certain modern poets, these alone are called symbols. Most things belong to one or another kind, according to the way we speak of them and the companions we give them, for symbols, associated with ideas that are more than fragments of the shadows thrown upon the intellect by the emotions they evoke, are the playthings of the allegorist or the pedant, and soon pass away. If I say white or purple in an ordinary line of poetry, they evoke emotions so exclusively that I cannot say why they move me; but if I bring them into the same sentence with such obvious intellectual symbols a s a cross or a crown of thorns, I think of purity and sovereignty. Furthermore, innumerable meanings, which are held to white or to purple by bonds of subtle suggestion, and alike in the emotions and in the intellect, move visibly through my mind, and move invisibly beyond the threshold of sleep, casting lights and shadows of an indefinable wisdom on what had seemed before, it may be, but sterility and noisy violence. It is the intellect that decides where the reader shall ponder over the procession of the symbols, and if the symbols are merely emotional, he gazes from amid the accidents and destinies of the world; but if the symbols are intellectual too, he becomes himself a part of pure intellect, and he is himself mingled with the procession. If I watch a rushy pool in the moonlight, my emotion at its beauty is mixed with memories of the man that I have seen ploughing by its margin, or of the lovers I saw there a night ago; but if I look at the moon herself and remember any of her ancient names and meanings, I move among divine people, and thing s that have shaken off our mortality, the tower of ivory, the queen of waters, the shining  stag  among enchanted woods, the white  hare  sitting upon the hilltop, the fool of  faery  with his shining cup full of dreams, and it may be make a friend of one of these images of wonder, and meet the Lord in the air. So, too, if one is moved by Shakespeare, who is content with emotional symbols that he may come the nearer to our sympathy, one is mixed with the whole spectacle of the world; while if one is moved by Dante, or by the myth of Demeter, one is mixed into the shadow of God or of a goddess. So too one is furthest from symbols when one is busy doing this or that, but the soul moves among symbols and unfolds in symbols when trance, or madness, or deep meditation has withdrawn it from every impulse but its own. I then saw, wrote Gà ©rard de Nerval of his madness, vaguely drifting into form, plastic images of antiquity, which outlined themselves, became definite, and seemed to represent symbols of which I only seized the idea with difficulty. In an earlier  time  he would have been of that multitude, whose souls austerity withdrew, even more perfectly than madness could withdraw his soul, from hope and memory, from desire and regret, that they might reveal those processions of symbols that men bow to before altars, and  woo  with incense and offerings. But being of our time, he has been like Maeterlinck, like Villiers de IIsle-Adam in  Axà «l, like all who are preoccupied with intellectual symbols in our time, a foreshadower of the new sacred book, of which all the arts, as somebody has said, are beginning to dream. How can the arts overcome the slow dying of mens hearts that we call the progress of the world, and lay their hands upon mens heartstrings again, without becoming the garment of religion as in old times? V If people were to accept the theory that poetry moves us because of its symbolism, what change should one look for in the manner of our poetry? A return to the way of our fathers, a casting out of descriptions of nature for the sake of nature, of the moral law for the sake of the moral law, a casting out of all anecdotes and of that brooding over scientific opinion that so often extinguished the central flame in Tennyson, and of that vehemence that would make us do or not do certain things; or, in other words, we should come to understand that the beryl stone was enchanted by our fathers that it might unfold the pictures in its heart, and  not to  mirror our own excited faces, or the boughs waving outside the window. With this change of substance, this return to imagination, this understanding that the laws of art, which are the hidden laws of the world, can alone bind the imagination, would come a change of style, and we would cast out of serious poetry those energetic rhythms, as of a man running, which are the invention of the will with its eyes always on something to be done or undone; and we would seek out those wavering, meditative, organic rhythms, which are the embodiment of the imagination, that neither desires nor hates, because it has done with time, and only wishes to gaze upon some reality, some beauty; nor would it be any longer possible for anybody to deny the importance of form, in all its kinds, for although you can expound an opinion, or describe a thing, when your words are not quite well chosen, you cannot give a body to something that moves beyond the senses, unless your words are as subtle, as complex, as full of mysterious life, as the body of a flower or of a woman. The form of sincere poetry, unlike the form of the popular poetry, may indeed be sometimes obscure, or ungrammatical as in some of the best of the Songs of Innocence and Experience, but it must have the perfections that escape analysis, the subtleties that have a new meaning every day, and it must have all this whether it be but a little song made out of a moment of dreamy  indolence,  or some great epic made out of the dreams of one poet and of a hundred generations whose hands were never weary of the sword. The Symbolism of Poetry by William Butler Yeats first appeared in  The Dome in April 1900 and was reprinted in Yeats Ideas of Good and Evil, 1903.

Critical Literature Review Essay Example | Topics and Well Written Essays - 1500 words

Critical Literature Review - Essay Example Therefore, they should be sheltered and protected under all circumstances. Moreover, the best way to protect them is to keep them away from the social life, locked inside the walls of the house for the rest of her lifetime (Brà ¼ckner, pp. 44-46, 2004). There are so many examples all over the world and throughout the history that indicate the gender disparity, gender inequality, sexism, or stratification of genders. Despite the fact that over the past decades, the situation has improved a bit but still there is a long way to travel. It is very difficult to talk about a world where there would be equal opportunities in relation to gender (Pasero, pp. 219-228, 2003). In fact, this is virtually impossible to implement all around the globe. There are several reasons for this strong but bitter assumption. We live in a world where, consciously or unconsciously, we have created some strong and irrevocable social perception and outlook of gender. In fact, gender is not just a symbol of any one’s reproductive role but it has become a symbol and determinant of one’s behavior, attitudes, and role expectations (Tomm, Hamilton, Mackie & Calgary Institute for the Humanities, pp. 159-167, 1998). For example, a female school teacher, if announces that she would be late for a class or would not be able to make an appearance the next day; this statement from her would not get so much of attention if the same statement is put forward by a male teacher. Actually, after hearing the same from the father we would start thinking about different things and would make various assumptions in out minds. That he would be a single father, or may be he is too nice or may be his wife is too busy with other stuff. This happens because as individuals grow up in human societies they quickly learn ab out their roles and expected behaviors (Chaneles, pp. 85-92, 1984). For example, both men

Conflict in a case study about conflict intervention in somalia Essay

Conflict in a case study about conflict intervention in somalia from1992 to 2010 - Essay Example Addressing these issues would also mean the promotion of broad objectives like the U.S. foreign policy to ensure regional stability in the whole of Africa in terms of trade and investments, supporting human rights and democracy and removing situation and circumstances which provides protected havens for the international terrorists and drug traffickers. The anarchy that followed the collapse of Somalia as a nation state in 1991 fathered large-scale migration of populations; the nation was facing economic chaos and the incidences of mass violence. The persistence of these problems triggered political dangers as groups emerged whose main aim was to initiate war and empower leaders. (Ohaegbulam, 2004, p. 65) Root cause of the conflict During the late 1980s the coalition of the three clans- the Marreexaan, Ogaadeen and the Dhulbahante – popularly known as the â€Å"MOD† under the leadership of President Maxamad Siyaad Barre was a declaration of war against the Isxaaqs of th e previous British colony in the Northwest which was organized into the Somali National Movement- the SNM. He also organized the Majeerteens in the northeast as the Somali Salvation Democratic front, the SSDF and the Hawiyas to Mogadishu’s direct west and south, which was organized as the United Somali Congress, known as the USC. ... The SNM were angry about the fact that their army could not make it to Mogadishu- the major land holdings of the Isxaaqs before the armies of the USC. The SNM were reluctant of returning home in the north but they had to, so while returning they created a rump state. In the meantime the USC’s military leader Maxamad Faarax Aideed from the Habar Gidir subclan challenged Mahdi’s right to the presidency. Hence these all resulted in the inter-clan war in 1991 to gain a control over Somalia. Beside this an inter clan war also took place to have a control over Mogadishu. The absence of the central government placed the country at a disadvantage, as there was nobody to negotiate on the behalf of the state and both the Somali population and the international relief workers were vulnerable to attack. (Lahneman, 2004, p. 31; Webersik, 2004, p. 518; Somalia Civil War, n.d; The civil war in Somalia, n. d) The Social Economic and Political crisis After the collapse of the Somali sta te in January 1991 the country faced disorganization in their education and learning system, which was an impact of the division of the clan in the country. In the process of the social breakdown the schools, technical education centers and the university faced the worst consequences. The infrastructures of the institutes were destroyed. The major victims of the civil war were the children, as nearly 3000 of them were dying every day in the early days of 1992. (Abdi, 1998, pp. 327-328) The situation at that time had worsened in the South especially in Mogadishu where the instances of well armed youths along with their armed Land Rovers roamed around the city and the roadways plundering and looting masses, extorting

Ordinary Control Freaks in Judith Guests Ordinary People Essay

Ordinary Control Freaks in Ordinary People The Most interesting form of Literature is American Literature. Usually dealing with a struggle that must be overcome, American literature deals with real-life situations which one can empathize with. One of the most interesting novels written by an American author is Ordinary People, by Judith Guest. Ordinary People tells the story of an ordinary family struggling to cope with the loss of a family member following a boating accident. Brilliantly written, the novel consists of two narrators- Conrad, the Jarrett family's only son left after the boating accident, and Calvin, Conrad's father. By using two narrators, both Calvin and Conrad's thoughts are revealed. In the novel, Calvin struggles to cope with the loss of his son Buck in the boating accident, and, afterward, Conrad's attempt to commit suicide. Calvin's personality conflicts with his wife's and his peers'. He desires control and order, but, to his dislike, doesn't always have it. Calvin's history was eventful, growing up in a foster home without a father and a mother he scarcely saw, which plays a critical role in his need for control. Growing up in a foster home, Calvin's childhood provided little opportunity to control his life. As a grownup, Calvin desires the control he never had. A good example of Calvin's desire for control is evident in the scene, during breakfast, when Calvin is talking to Conrad about Dr. Berger, a psychiatrist recommended by Conrad's former psychiatrist, Dr. Crawford. Also identifying Calvin's desire for control is Calvin's ambition to be a good father. His desire to be a good father stems from his childhood, where he never had a father. His career was no... .... Calvin's mindset at work has rubbed off on him in his personal life. Many examples can support this claim. A good example of Calvin's work mindset effecting his personal life can be found by his thoughts about Conrad's orthodontist trips in the quote "Strive, strive. Correct all defects" (9). In short, Calvin's desire for control is a trait which is easily identifiable in his personality. Evidence supporting the claim that Calvin desires control and order, but, to his dislike, doesn't always have it is numerous in the novel Ordinary People. Calvin's struggle to overcome his obstacles, including his foster home childhood, the death of his son Buck, and Conrad's suicide attempt, along with the challenges faced by Conrad and Beth allow for very complex and dynamic characters in this excellent example of American literature at its finest.

Hybrid Network Security

ACCEPTED FROM OPEN CALL SECURITY ISSUES IN HYBRID NETWORKS WITH A SATELLITE COMPONENT AYAN ROY-CHOWDHURY, JOHN S. BARAS, MICHAEL HADJITHEODOSIOU, AND SPYRO PAPADEMETRIOU, UNIVERSITY OF MARYLAND AT COLLEGE PARK ABSTRACT Satellites are expected to play an increasingly important role in providing broadband Internet services over long distances in an efficient manner. Most future networks will be hybrid in nature — having terrestrial nodes interconnected by satellite links. Security is an imporSSSL tant concern in such networks, since the session 2 Proxy satellite segment is susceptible to a host of attacks, including eavesdropping, session hijacking and data corruption. In this article we address the issue of securing communication in satellite networks. We discuss various security attacks that are possible in hybrid SSSL session 1 SSSL handshaking and satellite translation at client proxy (RPA) networks, and survey the different solutions proposed to secure data communications in these networks. We look at the perforMost future networks mance problems arising in hybrid networks due to security additions like Internet Security Prowill be hybrid in tocol (IPSec) or Secure Socket Layer (SSL), and suggest solutions to performance-related nature — having problems. We also point out important drawbacks in the proposed solutions, and suggest a terrestrial nodes hierarchical key-management approach for interconnected by adding data security to group communication in hybrid networks. satellite links. Security is an important concern in such networks, since the satellite segment is susceptible to a host of attacks, including eavesdropping, session hijacking and data corruption. INTRODUCTION With the rapid growth of the Internet, satellite networks are increasingly being used to deliver Internet services to large numbers of geographically dispersed users. The primary advantage of satellite networks is their wide broadcast reach — a satellite can reach users in remote areas where terrestrial connectivity is not available. Satellite networks are also easily and quickly deployed, and can be a more cost-effective solution in areas where laying ground fiber networks would be too expensive. Although satellite networks offer great potential, they also present significant challenges that need to be addressed. Security is becoming an increasingly important aspect of all network. In this article we focus on the challenges that need to be addressed in order to make satellite networks more secure while maintaining seamless interoperability with terrestrial networks. These security-related challenges include the following considerations: †¢ Satellite channels are wireless broadcast media, which makes it possible for an unauthorized user to receive the signal and eavesdrop on the communication, if it is not encrypted. †¢ Without proper security mechanisms, any sufficiently well-equipped adversary can send spurious commands to the satellite and jam or disrupt the communication. †¢ Satellite channels can occasionally have high bursty errors (for example, during heavy rain) that result in packet loss. Satellite networks also suffer from long propagation delays (for example, 0. 5 seconds for geostationary satellites). Therefore, security systems should add minimal delays to the communication and have mechanisms to recover from loss in security information. Incorporating security solutions originally designed for terrestrial networks, such as Internet Security Protocol (IPSec) or Secure Socket Layer (SSL), into satellite networks can cause severe performance penalties. In this article we consider some of these issues. We focus on data security for IP-based commercial networks, and discuss the performance problems that arise due to the encryption of the Transmission Control Protocol (TCP) header and payload when popular unicast security protocols like IPSec or SSL, originally designed for terrestrial connections, are applied to satellite networks without incorporating changes necessitated by the unique characteristics of satellite networks. We also look at the protocols proposed for secure group communication in hybrid satellite networks, and describe a hierarchical approach to group key management that is robust, scalable, and suitable for the characteristic topology of hybrid networks. The rest of the article is organized as follows. We describe the hybrid satellite-network topology and features that make it different from terrestrial networks. We discuss security needs for the hybrid network. We discuss the current approach to provide end-to-end unicast security in hybrid networks, and describe the performance problems arising as a result. We survey 50 1536-1284/05/$20. 00  © 2005 IEEE IEEE Wireless Communications †¢ December 2005 NOC Internet Gateway Proxy Client Proxy NOC (a) Internet Web server Web server (b) n Figure 1. Commercial direct-to-home network topology: a) case 1; b) case 2. the proposals for key management for secure group communication in satellite networks. We describe a possible solution to secure unicast communication without sacrificing performance and highlight our key-management approach to security for group communication in satellite networks. We conclude the article by pointing to future research directions. highly susceptible to the delay-bandwidth product and exhibits very poor performance in satellite channels. Satellite TCP connections need large transmit windows to fully utilize the available bandwidth. However, due to the TCP slowstart algorithm and large propagation delay in the satellite channel, it takes much longer for satellite TCP connections to reach the target window size, in comparison to terrestrial TCP connections. Also, the window is very vulnerable to congestion due to the multiplicative decrease strategy of TCP. The problem is compounded by the fact that TCP misinterprets link-layer corruption (which is the prevalent source of loss in satellite links) as congestion (which is rare) and consequently reduces the window. The PEP provides an efficient solution to the above problem. In satellite networks, a PEP agent is installed at the satellite gateway between the satellite network and the Internet. The PEP agent inspects every TCP packet that flows through the network. For data packets, the PEP sends back premature acknowledgments to the TCP senders, without waiting for the TCP segments to be actually delivered to the receivers. These premature acknowledgments are specially formatted to be indistinguishable from real acknowledgments and they considerably shorten the perceived round-trip delay. Studies have shown that this technique is critical for the performance improvement of satellite networks [2–4]. Hence, TCP PEPs have been widely deployed in satellite networks today. Commercial networks also employ HTTP proxy servers to improve the speed of responses to Web-browser requests. When a user browses through content on the Internet, the application layer protocol in use is HTTP. A typical HTTP exchange involves a request by the browser for a Web page (â€Å"GET†), and a response from the Web server, which contains the hypertext markup language (HTML) text of the requested Web page. A typical HTML page would also contain multiple embedded â€Å"objects† such as images, embedded media or scripts, and so forth. Each embedded object has to be retrieved with a separate HTTP request-and-response exchange. Therefore, a Web page that contains n – 1 embedded objects takes n * RTT time to load fully, where RTT is one round-trip time. This can be extremely costly in a satellite network, where the RTT is usually high. COMMERCIAL HYBRID SATELLITE NETWORK ARCHITECTURE The network topologies we consider are illustrated in Fig. 1. In both topologies, we assume that there is one geostationary satellite with multiple spot-beams covering a large geographical area. Each spot-beam covers a subset of the total user set. We assume that future satellites will have an IP stack, be capable of onboard processing, and switch the data between supported spotbeams. The satellite therefore acts as an IP router-in-thesky. The Network Operations/Control Center (commonly known as NOC or NCC) connects to the satellite through the hub satellite gateway. The NOC is also connected to the Internet through high-speed terrestrial links. Terrestrial users can be either standalone machines (Fig. 1a), or a cluster of machines at each location, such as a local area network (LAN) (Fig. 1b). Terrestrial LANs can be either wired or wireless. Each user or LAN is connected to a local satellite terminal. The users receive traffic from the satellite via the forward channel (satellite downlink). The users can also communicate with the satellite via the return channel (uplink). There is no terrestrial connectivity between the users or the LANs. Usually, in commercial satellite networks that transfer Internet traffic, a split-connection Transmission Control Protocol (TCP) Performance Enhancing Proxy (PEP) is implemented to reduce the negative effects of the satellite link on the Internet connection [1]. Satellite channels provide large bandwidth (which can be as high as 90 Mb/s in the downlink), but also suffer from long propagation delay in comparison to terrestrial links. The delay can be as high as 500 ms (round-trip) for a geostationary satellite link. The propagation delay can have a severe adverse impact on the delivery of Internet traffic. Most of the Internet traffic uses the TCP, which is IEEE Wireless Communications †¢ December 2005 51 SSSL encryption New IP header ESP Original TCP header IP header header TCP payload (SSL record) IPSEC encryption ESP trailer proxy (user side). There is a hub proxy server located at the NOC with the hub satellite gateway — this proxy server represents the gateway proxy for both TCP and HTTP performance enhancements. SECURITY THREATS Similar security attacks can be launched against different hybrid satellite network topologies, but the impact of attacks would differ depending on the type of network and the applications supported by the network scenario. In the following, we list some of the important security threats in the hybrid network described above, and highlight the importance of the threats for the different network scenarios. Confidentiality of information: For networks that require information privacy, a primary threat is unauthorized access to confidential data or eavesdropping. Since the satellite is a broadcast medium, any entity on the ground with the right equipment can receive the satellite transmission. If the data is broadcast in the clear, then adversaries can be privy to the information that is flowing in the network. Data confidentiality can be achieved by message encryption. This requires that the senders and receivers are concurrently aware of the correct cryptographic keys used in the encryption/ decryption operations. This is a twofold problem: the problem of selecting suitable cryptographic algorithms for doing encryption so that overall network performance is not affected, and the problem of coordinating keys between users, that is, key management. Sending spurious commands: An adversary with the right equipment can send spurious control and command messages to the spacecraft, thus making the spacecraft perform operations different from their intended use. This can disrupt legitimate operations and communication in the network. This attack can be prevented if the sources of the messages are properly authenticated by every receiver. This would require suitable mechanisms for authentication, such as digital signatures [5]. The level of security required would dictate the authentication policy, for example, whether only the end users should authenticate each other, or whether authentication should happen on a per-hop basis. The latter might be necessary for scenarios where the satellite should not broadcast spurious information. If the satellite authenticates the source of every message it receives, it will transmit only those messages for which source authentication occurs correctly. Message modification attack: When the traffic goes over open networks, an adversary who is listening on the path can intercept both control and data messages. The adversary can modify the messages and send them to the destination, which can be the spacecraft, the ground terminals, or the end users. When the message reaches the intended destination, it would think that the corrupt message is coming from the true source, but the message content might be different from that expected or required for normal network operation. Message modification can be prevented by SSL record HTML page n Figure 2. IPSec and SSL encryption on a packet. The HTTP proxy server (also known by various other names, depending on the vendor) is implemented in satellite networks to overcome this problem. In a typical implementation, this requires a local Web proxy server at each user location, and a remote proxy server at the central hub facility of the satellite network (i. e. , the NOC). The Web browser at the user location should be able to recognize the local proxy (which can be either software on the client machine, or a separate hardware connected inbetween the client machine and the local satellite terminal). When the browser makes a request for a Web page, the HTTP GET request is sent to the local Web proxy, which forwards the request to the destination Web server. The Web server responds with the requested base HTML page. This page is intercepted by the proxy server at the network hub facility. The hub proxy server reads the base HTML page and sends multiple GET requests to the destination Web server for all the embedded objects in the base HTML page. This exchange occurs over a high-speed terrestrial connection between the hub and the Internet, thereby saving the time each request would have needed for a round trip over the satellite link. As the objects of the Web page are retrieved by the hub, they are immediately forwarded to the proxy at the user location. As the user browser receives the base HTML documents, it generates appropriate GET requests to fetch the objects corresponding to the links embedded in the document. The browser GET requests are terminated at the Web proxy server, which forwards the prefetched documents to the user browser immediately. The net result is that only a single â€Å"GET† request from the user browser traverses the satellite link, while a set of rapid responses quickly deliver the requested Web page and associated elements to the browser. The need for satellite capacity is also reduced, which is the most costly element of a satellite network. In terms of the user’s experience, the user sees a brief pause after the original Web-page request (corresponding to the round-trip time it takes for the request to the forwarded to the destination server, and the response to be received by the browser, over the satellite link), followed by near-instantaneous delivery of all content residing on the requested page. The trade-off is additional hardware at the user location and the central-hub facility. In Fig. 1a, the proxy server at the user represents both the PEP (user side) and the HTTP 2 IEEE Wireless Communications †¢ December 2005 appending message-integrity check mechanisms to every message, for example, message authentication codes (MACs) [6] or digital signatures. Security requirements and policies can dictate whether message authentication should happen only at the communication end points, or whether interme diate nodes should also verify the integrity of every message. Denial-of-service attack: Some attacks on security can be facilitated if strong security mechanisms are put in place for performing message-integrity checks or authenticating users. Consider the case where the satellite does authentication and integrity checks on all messages before broadcasting. An adversary can send a large number of spurious messages to the satellite, thus making the satellite spend significant computational cycles processing the spurious messages, which could be better spent broadcasting legitimate messages. Since the satellite has limited processing power, such an attack can be very effective, especially if strong cryptographic mechanisms like digital signatures are used for authentication and message integrity. This is a denial-of-service (DOS) attack. Although this DOS attack can be launched against any node in a network, a satellite network can be particularly susceptible to such an attack, since the satellite is a single point of failure and can be easily overwhelmed if made to perform too much computation. New IP header ESP Original IP TCP header header header TCP payload ESP trailer Encryption with K1 Original IPSEC ESP tunnel mode encryption New IP header ESP Original IP TCP header header header TCP payload ESP trailer Encryption with K2 Encryption with K1 Layered IPSEC ESP tunnel mode incryption n Figure 3. IPSec and layered IPSec encryption. Key K1 is shared between endpoints only. Key K2 is shared between endpoints and TCP PEPs. 1 embedded objects takes n * RTT to be loaded, an increase in delay by a factor of n. IPSEC FOR SECURITY AT THE NETWORK LAYER Several proposals for data confidentiality and authentication in satellite networks call for use of IPSec, which has been widely adopted by the Internet Engineering Task Force (IETF) for security at the network layer. IPSec and SSL are used independently of each other. IPSec creates an end-to-end tunnel at the network layer for the secure transfer of traffic. The two end-points in the communication negotiate security parameters known as the security association (SA) before traffic can be encrypted. Once the SA has been established in the handshake phase, the IP packets are encrypted using the algorithms and the keys specified in the SA. This is done when the IP-encrypted security payload (IPSec ESP) [9] is used. The IPSec ESP provides for both data encryption and authentication. IPSec provides strong security for data confidentiality and authentication, but it has a heavy byte overhead — in the ESP mode, IPSec adds 10 bytes of overhead to the header and trailer. In addition, if authentication is used, ESP adds 16 bytes or more for the integrity check value, and another 8 bytes or more of initialization vector (IV) if the encryption algorithm uses an IV. Also, IPSec has been designed primarily to secure point-to-point communication; it s not well suited for group communication, due to the lack of the dynamic key-establishment procedure necessary to for secure communication in groups where the membership changes with time. In addition, IPSec does not allow for authentication at intermediate nodes, but this might be useful in some security situations. A widely researched problem when using IPSec in satellite networks is its inability to coexist with PEPs. The keys used for encryption in the IPSec ESP are known only to the two endpoints and therefore any intermediate node in the network cannot decrypt the traffic. IPSec ESP has two modes of operation — tunnel mode and transport mode. In tunnel mode, the entire IP packet is encrypted and a new IP header and ESP header are generated and attached to the encrypted packet (Fig. 3), which adds an extra SECURING END-TO-END UNICAST COMMUNICATION USING IPSEC OR SSL Research on satellite security has focused on using the existing standardized technology, originally designed for terrestrial networks, to fix well-known security holes in satellite networks. Two such protocols that are widely used for secure unicast communication are IPSec [7] and SSL [8]. Figure 2 illustrates the encryption regions of SSL and IPSec. SECURE SOCKET LAYER FOR SECURE WEB TRAFFIC The SSL protocol secures the Web-browsing connection on an as-needed basis. When the client requests a secure connection or the server demands one, SSL is activated to secure the HTTP connection. The resulting connection is popularly known as secure HTTP (or HTTPS) and it encrypts the application-layer HTTP data end-to-end between the client and the server. In the protocol stack, the SSL layer sits between the application and the transport layers. Therefore, SSL encryption hides the TCP payload from all nodes in the network, except the client and the server. SSL encryption does not allow the HTTP proxy to function correctly. The HTML Web page encrypted into the SSL records is readable only by the client and the server who have the decryption keys. The keys are not available to the proxy, and therefore the proxy cannot read the HTML Web page. Consequently, the hub proxy server cannot send requests to the Web server for the embedded objects in the page and, therefore, HTML object prefetching cannot take place. The net result is that a Web page with n – IEEE Wireless Communications †¢ December 2005 53 The HTTP proxy also cannot function when the IPSec ESP is used. Since the HTML page is encrypted end-toend, the HTTP proxy cannot read the Web page in order to prefetch the embedded objects. Therefore, use of IPSec leads to a severe degradation in performance for both the TCP PEP and HTTP proxy. SSSL translation at hub proxy (HPA) SSSL session 2 SSL session 3 Proxy Client Internet Gateway NOC SSSL session 1 SSSL handshaking and translation at client proxy (RPA) Proxy Web server Figure 4. The SSL Internet Page Accelerator concept for efficient HTTPS over satellite. 20 bytes of overhead in addition to the overhead mentioned above. Encrypting the original IP header provides very strong security by disabling attacks (such as traffic analysis, etc. ). In transport mode, the payload portion of the IP packet is encrypted and a new ESP header is attached to the packet after the original IP header, which is in the clear. In either mode, the IP packet payload, which includes the TCP header, is encrypted with keys known only to the end points. Therefore, a TCP PEP, which is an intermediate node in the communication path, cannot read or modify the TCP header, since the PEP does not know the keys. Consequently, the PEP cannot function, thus leading to degradation in the performance of the TCP protocol. The HTTP proxy also cannot function when the IPSec ESP is used. Since the HTML page is encrypted end-to-end, the HTTP proxy cannot read the Web page in order to prefetch the embedded objects. Therefore, use of IPSec leads to a severe degradation in performance for both the TCP PEP and HTTP proxy. It is important to note that the problems that arise from the use of the SSL protocol or the IPSec ESP are independent of one another. It is conceivable that both protocols are used simultaneously, for example, when a secure Web page is accessed via a secure VPN tunnel. However, in such cases the performance issues do not change and the effect would be equivalent to using the IPSec ESP alone. On the other hand, if SSL alone is used, then the performance would be better, since the TCP PEP can function correctly in this scenario. [10] and layered IPSec [11], the idea is to encrypt different regions of the IP packet using different keys (Fig. ). The TCP payload is encrypted with key K1, which is shared only between the endpoints. The original IP header and the TCP header are encrypted with key K2, which is shared between the end points and also with intermediate authorized nodes such as the TCP PEP. Therefore, the TCP PEP can decrypt the header portion of the ESP packet with K2 and read the TCP header to do its performance optimizations. But the PEP cannot read the TCP payload and thus cannot access the actual data, since it does not posses the key K1. The layered IPSec approach allows TCP PEPs to function effectively. However, the method does not solve the problem of HTTP proxy servers. The HTML page is encrypted with key K1 as part of the TCP payload, and K1 is not shared with any intermediate node. Therefore, the Web page is not accessible to the HTTP proxy and no object prefetching can be accomplished. Olechna et al. [12] have suggested two solutions to the IPSec problem. In the first approach, the paper proposes moving the TCP PEP gateways to the endpoints. The TCP optimizations are done on the traffic in the clear, and then the traffic is encrypted using IPSec. There is no TCP PEP at the satellite hub. This approach improves the performance, but when a packet is lost or received in error TCP goes into congestionavoidance phase and the transmission is reduced by half. The second proposed approach, which deals effectively with this problem, is to split the secure connection into two at the satellite gateway. One connection is between the client and the gateway, and the second connection is between the gateway and the Internet server. This allows the gateway to decrypt the IPSec packet and read the headers and thereby do performance optimizations. This requires trust in the satellite gateway, which can now read all the traffic. This might be unacceptable to users who require strong end-to-end security. Several modified TCP protocols have been proposed that perform better than the original specification in the event of channel errors or delay, or when IPSec is used. A discussion of PROPOSED SOLUTIONS TO MITIGATE PERFORMANCE PROBLEMS WITH SSL OR IPSEC Several proposals have been made in academia and industry to deal with performance problems that arise from using IPSec and SSL in satellite networks. The concept of breaking up IPSec encryption into multiple encryption regions or zones on a single packet has been proposed independently in [10, 11]. Although the finer details in the two approaches are different, the basic idea is the same. Known as multilayer IPSec (ML-IPSec) 54 IEEE Wireless Communications †¢ December 2005 Group keys (TEK) K1,8 Internal keys (KEK) K1,4 K5,8 Path of keys for M8 K1,2 K3,4 K5,6 K7,8 Leaf keys Members Group key K2 K1,12 Group key K1 K1,8 K9,12 K1 K2 K3 K4 K5 K6 K7 K8 M1 M2 M3 M4 M5 M6 M7 M8 (a) K1K2 K3 K4K5 K6 K7K8 K9 K10K11K12 M1M2M3M4M5M6M7M8 G1 G2 G3 G4 Members Gateways (b) Figure 5. Logical key hierarchy and its extension to satellite networks: a) with eight members; b) ML-IPSec integrated LKH tree with users and gateways. these TCP enhancements can be found in [13]. The problem of HTTP proxy performance when SSL is used has been addressed within the industry by breaking up the end-to-end single SSL connection between client and server into m ultiple SSL connections [14]. In this solution, the client browser creates a secure HTTP connection with the remote page accelerator (RPA) at the client satellite terminal, a second connection is created between the RPA and the hub page accelerator (HPA), and a third connection is between the HPA and the server (Fig. 4). The RPA performs all necessary handshaking with the client browser. The HPA can decrypt the SSL traffic from the server and perform the desired object prefetching. Taken together, this allows delivery of secure Web content with little performance degradation and with little change to the standard protocols. The major drawback to this scheme is that it requires a high level of trust in the intermediate nodes. The HPA, which is a third-party entity, can read all the sensitive Web traffic that passes between the client and the server. This might be unacceptable when absolute end-to-end security is desired. is O(logN) (where N is the number of members in the group), which is less than the O(N) keys required if the GC arranged the members in a flat topology. To allow PEPs to function correctly when network-layer security is used, [15] proposes the use of ML-IPSec. The paper proposes using a single LKH tree to manage the group key K2, used to encrypt the transport layer header (known to end users and trusted gateways), and the group key K1, known only to the end users and used for encrypting the transport layer data. As shown in Fig. b, users M 1 †¦M 8 are leaf nodes in a subtree of degree three, and gateways G 1 †¦G 4 are leaf nodes in a subtree of degree two. The root key of the member node subtree, K 1,8 , is used to encrypt the transport payload. The root of the overall key tree, K1,12, is used to encrypt the transport header. All member nodes know both K1,8 and K1,12, but the gateways know K 1,12 only (apart from the internal keys in the gateway subtree). How the LKH tree would be managed is not stated in [15]. This is important, since the users and the gateways might not be in the same administrative or security domain. The paper also considers all users and gateways as a â€Å"flat† network for key distribution purposes, rather than taking into account the hierarchical nature of the network topology. The use of LKH for key management in satellite links has also been proposed in [18], which suggests algorithms for dynamically managing the LKH tree in case of member joins and leaves. Duquerroy et al. [19] proposed â€Å"SatIPSec,† for key distribution and secure communication for both unicast and multicast in a satellite network. The solution is based on IPSec, with the addition of flat multicast key exchange (FMKE) to support key management for secure group communication. Management of SAs for both unicast and multicast communication is integrated into the FMKE protocol. FMKE also incorporates reliability mechanisms so as to guarantee reliable key distribution in the lossy satellite setting. However, FMKE manages SAs between the satellite terminals or gateways only and does not extend to the end users. Therefore, end-to-end security is not provided when using SatIPSec. The RPA performs all necessary handshaking with the client browser. The HPA can decrypt the SSL traffic from the server and perform the desired object prefetching. Taken together, this allows delivery of secure Web content with little performance degradation and with little change to the standard protocols. KEY MANAGEMENT PROPOSALS FOR SECURE GROUP COMMUNICATION IN HYBRID NETWORKS Some research has been done with individual algorithms that serve as tools in building keymanagement protocols in order to facilitate secure group communication in hybrid satellite networks. Howarth et al. [15] have proposed the use of logical key hierarchy (LKH) [16, 17] for efficient key management for multicast groups in a satellite network. LKH makes use of a centralized key manager or group controller (GC), which constructs a logical key tree with the group members as the leaves of the tree (Fig. 5a). The internal nodes of the tree are the key encrypting keys (KEK), which are used to securely transport key updates to the group. The root of the tree is the session key or traffic-encrypting key (TEK), which is used to encrypt the session traffic. The number of keys that need to be updated when a member node joins or leaves the group IEEE Wireless Communications †¢ December 2005 55 New IP header ESP Original TCP HTML header IP header header object links Base HTML page Encryption with K2 ESP trailer Encryption with K1 n Figure 6. Layered IPSec with modifications for HTTP optimization. Also, FMKE treats all the satellite terminals it services (which are called SatIPSec clients) in a â€Å"flat† topology, and establishes separate secure channels to all SatIPSec clients. This will not scale when there are a large number of clients. Also, SatIPSec does not consider the dynamic joins and leaves of members in the group communication setting; a client needs to be preauthorized for all the groups it wants to take part in. The protocol also requires complete trust in the group controller and key server (GCKS), which is a third party that is responsible for managing the SAs between the clients. All clients need to have preshared secrets with the GCKS. IPSEC AND SSL IN HYBRID NETWORKS: OUR APPROACH We look at separate solutions to the performance problem arising out of using SSL and IPSec in hybrid networks, and also consider how the two approaches can be combined. HTTP OVER IPSEC TUNNEL One viable method is to break up the end-to-end IPSec tunnel into multiple connections. This is similar to the solution proposed in [12]. But while their approach looks at only the TCP enhancements, we add the use of the HTTP proxy as well. In our approach, the IPSec connection from the client is terminated at the client proxy. The proxy creates its own IPSec connection to the gateway TCP proxy. A third IPSec connection is created from the gateway TCP proxy to the Web server. Schematically, this is similar to Fig. , with IPSec connections replacing the SSL connections in the figure. The IPSec handshaking between the client and the server is spoofed by the client proxy on the client end, and by the TCP hub proxy on the server end. In this model, the Web traffic can be read completely by the client proxy and the hub proxy. The two proxies are able to perform the TCP enhancements because they can read the TCP header. In addition, the hub HTTP proxy can perform HTM L object prefetching from the server because it can read the base HTML page as it is returned to the client on a HTTP request. When the client browser generates staggered requests for the embedded objects upon receiving the base HTML page, the client proxy is responsible for returning local acknowledgments to the requests, and sending all the objects to the client browser at one time. The design is therefore fully able to maintain the functionality of the TCP and HTTP proxies. It also encrypts the traffic so that it can be seen only by the client, the server, and the two intermediate proxy servers. The design also makes minimal changes to existing standard protocols. However, the design also requires that there be full trust in the proxy servers. Also, there is additional overhead in setting up three IPSec connections, as opposed to one (as in the end-to-end case). The overhead in encryption/decryption also increases by a factor of three for every IP packet, since the intermediate proxies need to decrypt the TCP header and the HTML content. When the security requirement is that the traffic be unreadable to intermediate nodes, the above approach will not work. In this situation, we propose extending the layered IPSec approach in order to allow portions of the HTML content to be also accessible to the proxy servers. Assume for layered IPSec that the keys are K1 and K2. K1 is known only to the client and the server, while K2 is known to the client, the Web server, and the intermediate proxy servers at the client and the gateway. When the client makes HTTP requests, the requests are encrypted using K2, so that the client proxy server can read the requests and send local acknowledgments. Additional software at the Web server parses the requested HTML page so as to obtain all the embedded object links. These object links are collated into a new HTML page that contains only the object links, and this new page is encrypted with K2. The base HTML page that contains all the information and the object links is encrypted with K1. Both the encrypted base HTML page and the encrypted object links HTML page are sent in reply. Therefore, the encrypted ESP packet looks as it is depicted in Fig. 6. Upon receiving the IPSec packet from the Web server, the hub proxy is able to read the object links (since it has K2) and therefore do prefetching for the embedded links. In addition, the hub proxy can also read the TCP header and perform TCP enhancements. However, the HTML base-page data cannot be read by the hub proxy, since it does not have K1. The encrypted base HTML page can only be read by the client when the IPSec packet reaches the destination. This design allows the TCP and HTTP proxies to perform effectively while maintaining a high level of end-to-end security. However, the security is not as strong as in traditional IPSec, since the intermediate proxies do get some information insofar as they can read the links of the embedded objects, even though they cannot read the application data. This is the major trade-off necessary to achieve acceptable performance in this design. In addition, the model requires changes to be made to the IPSec protocol so that layered IPSec is supported with the HTTP performance additions. A major issue in the above model is the handshaking mechanism required to set up the layered IPSec connection. To maintain a high level of security, we propose that the connection be set up primarily between the client and the server, who negotiate both K1 and K2, apart from other parameters of the security association. The handshaking mechanism then provides K2 securely to both the client and the hub proxy servers. The client and the hub proxy servers are required to authenticate themselves correctly before they can receive the secondary key or access the IPSec traffic. 56 IEEE Wireless Communications †¢ December 2005 DSSL main mode (K1) DSSL main mode (K2) Internet Gateway Proxy (K2) Client (K1, K2) NOC DSSL primary and secondary modes Original TCP IP header header Proxy SSL record Primary SSL record Proxy (K2) Web server (K1,K2) Encryption with K2 Encryption with K1 IP packet format for DSSL n Figure 7. Dual-mode SSL for HTTP optimization. HTTP OVER SSL When the HTTP traffic is secured using SSL only, and there is no IPSec tunnel in use, several approaches are possible to ensure acceptable performance. If the security requirement of the client and the Web server allow for trusted intermediate nodes, then the SSL accelerator concept of [14] can be a viable solution. This would require no change to the protocols at the expense of higher overhead in order to set up multiple SSL connections between the client, proxy, and Web server. When the security policy does not allow for trusted third parties, a different approach is needed. We propose the use of a modified SSL protocol, which we term dual-mode SSL (DSSL). As shown in Fig. 7, the secure connection in DSSL has two modes — an end-to-end main mode connection between the client and the Web server, and a secondary mode connection that has the hub HTTP proxy as an intermediate node. When secure HTTP traffic is requested, the DSSL main mode connection is first negotiated between the client and the server. As part of the handshake for the main mode, the client and the Web server also negotiate the parameters for the secondary mode. Let K1 be the encryption key for the main mode, and K2 be the encryption key for the secondary mode. The client transfers the parameters of the secondary mode to the client and hub HTTP proxy servers only after the proxy servers authenticate themselves to the client. When the client makes an HTTP request, the client proxy sends local replies to the client browser, as discussed previously. The Web server, on receiving the request, parses the requested HTML page to obtain the embedded object links, which are collated into a new HTML page. The object links HTML page is then encrypted by DSSL using K2 to create the proxy SSL record. DSSL encrypts the base HTML page using K1 to create the primary SSL record. The two records are appended together and sent to the client in an IP packet (Fig. 7). The hub proxy intercepts the IP packet, extracts the object links from the proxy SSL record using K2, and prefetches the embedded objects. The Web server always encrypts the actual objects using K1, so that the hub proxy cannot read the base HTML page data. The hub proxy transfers all the embedded objects together to the client at one time. Therefore, the HTTP proxy functionality is preserved in DSSL while maintaining the end-to-end security of the HTML page contents. However, the security is less than in the end-to-end SSL connection case, since the HTTP proxy can read the object links. In standard SSL, the proxy servers can read no part of the base HTML page, not even the object links. We believe this slight reduction in security is acceptable, given the considerable improvement in performance using this method. The DSSL design is more complex in comparison to SSL since it requires the creation of an additional connection, and therefore involves a higher overhead. There is also the added overhead of multiple encryptions and decryptions with two different keys, and the complexity of parsing the HTML page for the object links. All these require changes to the base SSL protocol. The DSSL concept is similar to the multiplechannel SSL concept proposed in [20]. However, the authors do not differentiate encryption in primary and secondary SSL records but instead suggest that HTTP traffic with lower security requirements be encrypted entirely with keys known to intermediate nodes. For our security requirements, that approach would not be acceptable. Differential Encryption in Single SSL Record — The use of a proxy SSL record is not necessary if various parts of the HTML page can be encrypted with The DSSL design is more complex in comparison to SSL since it requires the creation of an additional connection, and therefore involves a higher overhead. There is also the added overhead of multiple encryptions and decryptions with two different keys. IEEE Wireless Communications †¢ December 2005 57 Secondary SSL encryption Primary SSL encryption New IP header ESP Original TCP Proxy SSL header IP header header record Secondary IPSec encryption Primary SSL record ESP trailer Primary IPSec encryption n Figure 8. Packet format for dual-mode SSL with IPSec. different keys. In that case, the Web server can encrypt the object links in the HTML page with key K2 and the rest of the HTML page contents with key K1, thus creating a single SSL record with different encryption. The hub proxy server can parse the SSL record and decrypt only the object links with key K2, before forwarding the IP packet to the client proxy. We assume that the primary and secondary encryption keys K1 and K2 have been set up and distributed as described in the previous sections, with K1 known to the client and the Web server only, while K2 is known to the client, the Web server, and the intermediate proxy servers. A similar technique can be applied when IPSec encryption is used instead of SSL encryption. The advantage here is that the size of the packet does not increase, although there is the overhead of distributing key K2 to the proxy servers to be considered. HTTPS OVER IPSEC For the sake of completeness, we consider the situation where a secure Web page is requested over an IPSec tunnel. This method involves redundancy of resources, since use of SSL when IPSec is being used does not provide any substantially added security. However, our approach can take care of the performance in this scenario as well. In this situation, we propose integrating DSSL with layered IPSec. Then the secondary keys for both the layered IPSec connection and the DSSL connection are shared with the proxy servers. The secondary key for layered IPSec is shared with both the TCP proxy and the HTTP proxy. When layered IPSec encrypts the packet, the secondary key encryption extends up to the proxy SSL record. The TCP proxy servers can therefore decrypt the TCP header of the ESP packet, and the HTTP proxy server can decrypt the proxy SSL record. Consequently, performance optimizations for both TCP and HTTP are allowed without letting the intermediate servers read the HTML page. A schematic of the IPSec packet in this setting is shown in Fig. 8. A HIERARCHICAL APPROACH TO KEY MANAGEMENT FOR DATA SECURITY IN HYBRID NETWORKS In [21], we have proposed a key-management framework for distributing cryptographic keys securely and in a scalable manner to users taking part in group communication in a hybrid satellite network. The objective is to ensure data confidentiality, by encrypting the data traffic with group keys known to all the group members. The key-management framework is built on top of the multicast routing architecture. We have considered the hybrid network topology shown in Fig. 1b and designed a multicast routing architecture to allow users to communicate seamlessly between multiple terrestrial LANs (also referred to as subnetworks) [22]. Our routing design makes specific use of asynchronous transfer mode (ATM) point-to-multipoint routing [23] over the satellite links, and Protocol-Independent Multicast Sparse-Mode (PIM-SM) multicast routing [24] in terrestrial LANs. We have extended PIM-SM to allow multiple rendezvous points (RPs) in each multicast group. The satellite gateway in each LAN acts as the local RP for the LAN and creates the local multicast trees for group members within the LAN. The local multicast trees are connected together over the satellite links by using the ATM point-to-multipoint virtual connection, thereby creating one end-to-end multicast tree for each group, encompassing all the LANs with group members in them. The multicast routing architecture is thus adapted closely to the hierarchical network topology, and allows for building efficient multicast trees with low control and data overhead. The design of the key-management protocol is independent of the routing algorithm, although it is based on the same underlying principle, that is, a hierarchical breakup of the network based on the topology. We divide the network into two levels — the lower level, comprised of terrestrial LANs where the users are located, and a higher level consisting of the satellite, the NOC, and the satellite gateways or RPs in each LAN, which together form an overlay (Fig. 9a) interconnecting terrestrial LANs. The RPs act as the â€Å"bridge† between the two levels. Key management is done separately in the two levels. In each LAN we introduce a local group controller (called the â€Å"subnetwork key controller† or SKC) to manage the keys for all groups active in the LAN. The SKC is responsible for access control of all members of all groups that are active in its LAN, generating the group keys for all local groups, and updating the keys on group-member joins and leaves when a group is active. The keys managed by an SKC are entirely local to its LAN, and do not affect the key management in any other LAN in the network. The SKC uses the LKH algorithm to manage keys in its LAN, creating a logical key tree that we term the SN Tree. Each group active in a LAN has its own SN Tree. The leaves of the SN Tree for a group correspond to the longterm shared secrets between the SKC and the local users in the LAN who are active as sources and/or receivers in the group. The root of the SN Tree corresponds to the session key that is used for encrypting the group traffic within the LAN at any particular instant. On member joins and leaves, the session key, and all the keys on the path from the root to the leaf node corresponding to the member joining/leaving, are updated, while all other keys in the SN Tree remain unchanged. The overlay has its own key management, 58 IEEE Wireless Communications †¢ December 2005 Overlay RP tree root key Level-1 Satellite Overlay network NOC Gateway Gateway Gateway SN tree SKC RP (root) RP (leaf) SKC RP tree RP (leaf) Gateway Level-0 SKC Subnet n Subnet 1 Subnet 2 (a) Subnet 3 Subnetwork Subnetwork (b) Subnetwork n Figure 9. A hierarchical approach to key management in hybrid networks: a) hierarchy in the hybrid network; b) tiered tree key management. also based on the LKH algorithm. At the overlay level, the key management for a particular group is controlled by the satellite gateway/RP (known as the root RP for that group) of the LAN that has group sources active for the longest continuous period in the group. The logical key tree for any group thus formed at the overlay is termed the RP Tree. The root RP is responsible for generating keys for the RPs of the LANs who subscribe to the particular group, that is, have sources and/or receivers active in the LAN. Each group has its own RP Tree. The design ensures that the NOC cannot receive/transmit data to any active group, unless it explicitly subscribes to the group as a member node. However, LANs joining any particular group initially register with the NOC, which maintains a group membership table for all active groups, so that at all times the NOC is aware of the LANs which are participating in all active groups. The NOC is also responsible for selecting the root RP of the RP Tree for each group, which it does based on the earliest-to-join policy. The root RP also might be different for different groups, since the LAN with the longest continuously active sources might be different for different groups. Our algorithm has the provision to allow the root RP for any group to change — this happens if the currently active root RP leaves the group, when all sources/receivers within its local LAN cease to participate in the group. Our algorithm therefore builds a hierarchy of logical key trees that closely follow the hierarchy in the network topology, as shown in Fig. 9b. We term this framework Tiered Tree-based Key Management. In this hierarchy of key trees, the gateway RPs are responsible for performing key translation on all the multicast group traffic as it transmits the data from local sources to receivers in remote LANs, or when it receives group traffic from remote sources for local receivers. This translation is necessary since the data traffic is encrypted with the RP Tree session key in the overlay, and with the SN Tree session ey within the local LAN, with the two session keys being independent of one another. The detailed design of Tiered Tree-based Key Management, analysis of its security, and experimental results can be found in [25]. The primary objective in our design is to minimize the amount of key-management control traffic n Figure 10. Tiered tree framework: total key management traffic vs. RP tree traffic for three groups (Y-axis shows the traffic in bytes per second; X-axis is the simulation duration in minutes). hat flows over the satellite links, due to the long delay involved as well as susceptibility to channel errors. We have attempted to ensure that the security of the data traffic does not add any overhead in terms of delay other than that absolutely unavoidable, and that the security protocol does not contribute to deadlocks in group-data dissemination where some group members in certain LANs cannot read the data due to having wrong keys. From the simulation results, Fig. 10 shows the reduction in key-control traffic over the satellite links using our tiered-tree approach. The graph compares the total key-management IEEE Wireless Communications †¢ December 2005 59 Our solution is a generic solution aimed specifically at multicast key management and does not deal with an end-to-end security solution for secure communication or give any implementation specifics. information sent in the network for three simultaneous groups (i. e. , sent over the RP trees, sent over the satellite links, and all SN trees limited to local LANs), to the total key information sent on the RP trees (satellite links) only. As the graph shows, the resource savings on the satellite links is substantial when the tiered-tree scheme is used. Even though the group dynamics are high, the amount of message exchanges are very few in the RP tree, that is, over the satellite links. If a flat key-management hierarchy had been used instead, the total key-management traffic would have been sent over the satellite links, thus leading to increased delay and increasing the possibility that the correct keys do not reach all the members at the same time. Our solution is therefore very scalable. It also acknowledges the fact that the group members might be located in different security domains and, therefore, a single network-wide security management might not be possible. This is a more realistic scenario, since terrestrial LANs might be individual company domains, while the satellite overlay infrastructure is usually owned by a separate entity that provides network connectivity to the LANs, and is not responsible for generating the network traffic. This framework addresses the problem that all users might not be visible to a single, centralized security authority, and the dynamics of user joins or leaves in one LAN should not create an overhead to users in other LANs. Also, in widearea satellite networks we consider that the satellite channel conditions at a given point in time might be different in different sections of the network. There might be loss in information due to bad channel conditions in some network segments; however, this should not disrupt communication in network segments where the channel conditions are better. Solutions which treat all users in a single tree will not be able to perform as robustly under such conditions. Our solution is also similar to the ML-IPSec concept in that the satellite terminals are only partially trusted; they are allowed to do partial decryption/encryption of the IP packets for efficient routing. However, it is a generic solution aimed specifically at multicast key management and does not deal with an end-to-end security solution for secure communication or give any implementation specifics. approaches for typical topologies and validating the proposed designs by simulation. Lastly, we have described our hierarchical approach of key management for providing data security in hybrid networks. We are continuing our research in this area and examining designs to integrate our keymanagement protocol with the unicast case. A considerable amount of work needs to be done with regard to secure protocols for hybrid networks, specifically for the case where users are mobile. Here we have touched upon only a small subset of the problems. None of the proposed solutions, including our own, address the question of user authentication or message integrity for group communication. However, we believe the security problems discussed here will receive further treatment from the research community, and this work will be a useful contribution to the field. ACKNOWLEDGMENT The authors would like to thank the anonymous reviewers for their valuable comments and suggestions. The research reported here is supported by the National Aeronautics and Space Administration (NASA) Marshall Space Flight Center under award no. NCC8-235. The views expressed in this article are solely the responsibility of the authors and do not reflect the views or position of NASA or any of its components. REFERENCES [1] J. Border et al. , â€Å"Performance Enhancing Proxies Intended to Mitigate Link-Related Degradations,† IETF RFC 3135, June 2001. [2] V. Arora et al. , â€Å"Effective Extensions of Internet in Hybrid Satellite-Terrestrial Networks,† University of Maryland, College Park, Tech. Rep. CSHCN TR 96-2, 1996. [3] V. Bharadwaj, â€Å"Improving TCP Performance over HighBandwidth Geostationary Satellite Links,† University of Maryland, College Park, Tech. Rep. ISR TR MS-99-12, 1999. [4] N. Ehsan, M. Liu, and R. Ragland, â€Å"Evaluation of Performance Enhancing Proxies in Internet over Satellite,† Wiley Int’l. J. Commun. Sys. , vol. 16, Aug. 2003, pp. 513–34. [5] NIST, â€Å"Digital Signature Standard (DSS),† May 19, 1994. [6] H. Krawczyk, M. Bellare, and R. Canetti, â€Å"HMAC: KeyedHashing for Message Authentication,† IETF RFC 2104, Feb. 1997. [7] R. Atkinson and S. Kent, â€Å"Security Architecture for the Internet Protocol,† IETF RFC 2401, Nov. 1998. [8] IETF Transport Layer Security Working Group, â€Å"The SSL Protocol Version 3. 0,† Nov. 1996, available at http://wp. netscape. com/eng/ssl3/draft302. txt [9] R. Atkinson and S. Kent, â€Å"IP Encapsulating Security Payload (ESP),† IETF RFC 2406, Nov. 998. [10] Y. Zhang, â€Å"A Multilayer IP Security Protocol for TCP Performance Enhancement in Wireless Networks,† IEEE JSAC, vol. 22, no. 4, 2004, pp. 767–76. [11] M. Karir and J. Baras, â€Å"LES: Layered Encryption Security,† Proc. ICN’04, Guadeloupe (French Caribbean), Mar. 2004. [12] E. Olechna, P. Feighery, and S. Hryckiewicz, â€Å"Virtual Private Network Issues Using Satellite Based Networks,† MILCOM 2001, vol. 2, 2001, pp. 785–89. [13] P. Chitre, M. Karir, and M. Hadjitheodosiou, â€Å"TCP in the IPSec Environment,† AIAA ICSSC 2004, Monterey, CA, May 2004. 14] SSL Accelerator, Spacenet Inc. , available at http://www. spacenet. com/technology/advantages/ssl. ht ml [15] M. P. Howarth et al. , â€Å"Dynamics of Key Management in Secure Satellite Multicast,† IEEE JSAC, vol. 22, no. 2, 2004, pp. 308–19. [16] C. Wong, M. Gouda, and S. S. Lam, â€Å"Secure Group Communications Using Key Graphs,† IEEE/ACM Trans. Net. , vol. 8, 2000, pp. 16–30. CONCLUSION Security is a critical component in hybrid IPbased satellite networks. In this article we have focused on some of the challenges that lie ahead. We have discussed the unique characteristics of hybrid satellite networks that make the problem of ensuring secure communication different from that of purely terrestrial networks. We have presented a survey of the various security solutions that have been proposed, and discussed their advantages and disadvantages. We have proposed several approaches to solve the performance problems of TCP and HTTP in satellite networks arising from secure communication. However, a lot of further work needs to be done to validate our approaches, and we are in the process of developing specific detailed security 0 IEEE Wireless Communications †¢ December 2005 [17] D. Wallner, E. Harder, and R. Agee, â€Å"Key Management for Multicast: Issues and Architectures,† IETF RFC 2627, June 1999, available at http://www. apps. ietf. org/rfc/ rfc2627. html [18] G. Noubir and L. von Allmen, â€Å"Security Issues in Internet Protocols over Satellite Links,† Proc. IEEE VTC ‘99, Amsterdam, The Netherlands, 1999. [19] L. Duquerroy et al. , â€Å"SatIPSec: An Optimized Solution for Securing Multicast and Unicast Satellite Transmissions,† 22nd AIAA Int’l. Commun. Sat. Sys. Conf. and Exhibit, Monterey, CA, May 2004. [20] Y. Song, V. Leung, and K. Beznosov, â€Å"Supporting Endto-End Security across Proxies with Multiple-Channel SSL,† Proc. 19th IFIP Info. Sec. Conf. , Toulouse, France, Aug. 2004, pp. 323–37. [21] A. Roy-Chowdhury and J. Baras, â€Å"Key Management for Secure Multicast in Hybrid Satellite Networks,† 19th IFIP Info. Sec. Conf. , Toulouse, France, Aug. 2004. [22] A. Roy-Chowdhury and J. Baras, â€Å"Framework for IP Multicast in Satellite ATM Networks,† AIAA ICSSC 2004, Monterey, CA, May 2004. [23] G. Armitage, â€Å"Support for Multicast over UNI 3. 0/3. 1 Based ATM Networks,† Internet RFC 2022, Nov. 1996. [24] S. Deering et al. , â€Å"The PIM Architecture for Wide-Area Multicast Routing,† IEEE/ACM Trans. Net. , vol. 4, no. 2, 1996, pp. 153–62. [25] A. Roy-Chowdhury, â€Å"IP Routing and Key Management for Secure Multicast in Satellite ATM Networks,† Master’s thesis, University of Maryland, College Park, 2003, available at http://techreports. isr. umd. edu/reports/2004/ MS2004-1. pdf Paper Award, 2004 WiSe Conference. He holds three patents. His research interests include wireless networks and MANET, wireless network security and information assurance, integration of logic programming and nonlinear programming for trade-off analysis, multicriteria optimization, noncooperative and cooperative dynamic games, robust control of nonlinear systems and hybrid automata, mathematical and statistical physics algorithms for control and communication systems, distributed asynchronous control and communication systems, object-oriented modeling of complex engineering systems, satellite and hybrid communication networks, network management, fast Internet services over hybrid wireless networks, stochastic systems, planning and optimization, intelligent control and learning, biologically inspired algorithms for signal processing, and sensor networks. MICHAEL HADJITHEODOSIOU [M] received an M. A. (honours) in electrical and information sciences from the University of Cambridge, United Kingdom, in 1989, an M. S. in electrical and computer engineering from the University of California, Irvine in 1992, and a Ph. D. n engineering (specializing in satellite communications) from the Centre for Satellite Engineering Research (CSER) at the University of Surrey, United Kingdom, in 1995. Among his awards are a scholarship award for studies at the University of Cambridge from the Cambridge Commonwealth Trust (1984–1986); a Fulbright Scholarship for post-graduate work in the United States (1989–1991); a Research Fellowship from the U. K. Engineering and Physical Sciences Research Council (EPSRC) (1992); and the Canadian National Science and Engineering Research Council (NSERC) post-doctoral fellowship award (1995). He worked as a research fellow in the Communication Systems group of CSER (1991–1995) and spent a year as a visiting fellow at the Canadian Government Communications Research Center (CRC) (1995–1996). In November 1996 he joined the Center for Satellite and Hybrid Communication Networks (CSHCN) at the Institute for Systems Research, University of Maryland, College Park, where he is currently an assistant research scientist. He is an expert on space communications and satellite networks. His research interests include performance optimization of wireless and hybrid networks, security and protocol support issues for satellite systems, and design optimization of next-generation broadband satellite networks and applications. He is currently working on supporting the communication needs of NASA enterprises and the communication architecture enabling space exploration. He is currently serving as secretary of the IEEE Satellite and Space Communications Technical Committee. SPYRO PAPADEMETRIOU received his B. S. in computer science from George Mason University, Fairfax, Virginia. Since then he has been actively involved in Internet research and development within both industry and academia. He was the principal Internet researcher at Synectics Corp. , where he developed network and database software. He worked as a researcher at the University of Maryland’s Institute for Systems Research, where he designed and developed their first networking laboratory, which is part of the CSHCN. At Inktomi Corp. he spearheaded client acceleration research and was a member the content-distribution network design team. These resulted in several patent filings, of which he holds one. The latter also resulted in American Online’s Web client accelerator product. Currently he is with Orbital Data Corp. working on network and application optimization. His research interests include network optimization, application optimization, satellite and terrestrial wireless networking, delay-tolerant networks, sensor networks, distributed systems, and network software architecture. We have touched upon only a small subset of the problems. None of the proposed solutions, including our own, address the question of user authentication or message integrity for group communication. BIOGRAPHIES AYAN ROY-CHOWDHURY ([email  protected] umd. edu) received his B. E. in electronics and telecommunications engineering in 1998 from Jadavapur University, India, and his M. S. in electrical engineering in 2003 from the University of Maryland, College Park, where he is currently a Ph. D. student. Between 1998 and 2000 he worked as a senior software engineer at Wipro Technologies, India. His research focuses on the design of protocols and frameworks for secure communication in hybrid networks. He is working on secure protocols for unicast and multicast routing in networks that have wired and wireless terrestrial components interconnected by satellite links. He is also looking into key management techniques for secure data transmission for these network architectures, and efficient user-authentication mechanisms for the same. As part of these topics, he is also investigating performance problems for network communication in satellite networks when security is involved. J OHN S. B ARAS [F] received a B. S. in electrical engineering from National Technical University of Athens, Greece, in 1970, and M. S. and Ph. D. degrees in applied mathematics from Harvard University in 1971 and 1973, respectively. He was founding director of the Institute for Systems Research (one of the first six NSF Engineering Research Centers) from 1985 to 1991. Since August 1973 he has been with the Electrical and Computer Engineering Department and Applied Mathematics Faculty at the University of Maryland, College Park. In 1990 he was appointed to the Lockheed Martin Chair in Systems Engineering. Since 1991 he has been director of the Center for Hybrid and Satellite Communication Networks (a NASA Research Partnership Center). Among his awards are the 1980 Outstanding Paper A