![]() |
|
сделать стартовой | добавить в избранное |
![]() |
Компьютеры, Программирование
Компьютеры и периферийные устройства
Evaluating the GPRS Radio Interface for Different Quality of Service Profiles |
Abs rac . his paper prese s a discre e-eve simula or for he Ge eral Packe Radio Service (GPRS) o he IP level. GPRS is a s a dard o packe da a i GSM sys ems ha will become commercially available by he e d of his year. he simula or focuses o he commu ica io over he radio i erface, because i is o e of he ce ral aspec s of GPRS. We s udy he correla io of GSM a dGPRS users by a s a ic a d dy amic cha el alloca io scheme. I co ras o previous work, our approach represe s he mobili y of users hrough arrival ra es of ew GSM a d GPRS users as well as ha dover ra es of GSM a d GPRS users from eighbori g cells. Fur hermore, we co sider users wi h differe QoS profiles modeled by a weigh ed fair queuei g scheme. he simula or co siders a cell clus er comprisi g seve hexago al cells. We provide curves for average carried raffic a d packe loss probabili ies for differe cha el alloca io schemes a d packe priori ies as well as curves for average hroughpu per GPRS user. A de ailed compariso be wee s a ic a d dy amic cha el alloca io schemes is provided.1 I roduc io he Ge eral Packe Radio Service (GPRS) is a s a dard from he Europea elecommu ica io s S a dards I s i u e (E SI) o packe da a i GSM sys ems . By addi g GPRS fu c io ali y o he exis i g GSM e work, opera ors ca give heir subscribers resource-efficie wireless access o ex er al I er e pro ocol-bases e works, such as he I er e a d corpora e i ra e s. he basic idea of GPRS is o provide a packe -swi ched bearer service i a GSM e work. As impressively demo s ra ed by he I er e , packe -swi ched e works make more efficie use of he resources for burs y da a applica io s a d provide more flexibili y i ge eral. I previous work, several a aly ical models have bee developed o s udy da a services i a GSM e work. Ajmo e Marsa e al. s udied mul imedia services i a GSM e work by providi g more ha o e cha el for da a services . Boucherie a d Li je s developed a a aly ical model based o Markov chai a alysis o s udy he performa ce of GPRS u der a give GSM call charac eris ic . For a aly ical rac abili y, hey assumed expo e ially dis ribu ed arrival imes for packe s a d expo e ial packe ra sfer imes, respec ively. O he o her ha d, discre e-eve simula io based s udies of GPRS were co duc ed. Meyer e al. focused o he performa ce of CP over GPRS u der several carrier o i erfere ce co di io s a d codi g schemes of da a . Fur hermore, hey provided a de ailed impleme a io of he GPRS pro ocol s ack . Malomsoky e al. developed a simula io based GPRS e work dime sio i g ool . S uckma e al. s udied he correla io of GSM a d GPRS users wi h he simula or GPRSim . his paper describes a discre e-eve simula or for GPRS o he IP level. he simula or is developed usi g he simula io package CSIM a d co siders a cellclus er comprisi g of seve hexago al cells. he prese ed performa ce s udies were co duc ed for he i ermos cell of he seve cell clus er. he simula or focuses o he commu ica io over he radio i erface, because his is o e of he ce ral aspec s of GPRS. I fac , he air i erface mai ly de ermi es he performa ce of GPRS.
We s udied he correla io of GSM a d GPRS users by a s a ic a d dy amic cha el alloca io scheme. A firs approach of modeli g dy amic cha el alloca io was i roduced by Bia chi e al. a d is k ow as Dy amic Cha el S eali g (DCS) . he basic DCS co cep is o emporarily assig he raffic cha els dedica ed o circui -swi ched co ec io s bu u used because s a is ical raffic fluc ua io s. his ca be do e a o expe se i erms of radio resource, a d wi h o impac o circui swi ched services performa ce if he cha el alloca io o packe -swi ched services is permi ed o ly for idle raffic cha els, a d he s ole cha els are immedia ely released whe reques ed by he circui -swi ched service. I co ras o he models developed i , our approach addi io ally represe s he mobili y of users hrough arrival ra es of ew GSM a d GPRS users as well as ha dover ra es of GSM a d GPRS users from eighbori g cells. Fur hermore, we co sider users wi h differe QoS profiles modeled by a weigh ed fair queuei g scheme accordi g o . he remai der of he paper is orga ized as follows. Sec io 2 describes he basic GPRS e work archi ec ure, he radio i erface, a d differe QoS profiles, which will be co sidered i he simula or. I Sec io 3 we describe he sof ware archi ec ure of he GPRS simula or, de ails abou he mobili y of GSM a d GPRS users, he way we modeled quali y of service profiles, a d he workload model we used. Resul s of he simula io s udies are prese ed i Sec io 4. We provide curves for average carried raffic a d packe loss probabili ies for differe cha el alloca io schemes a d packe priori ies as well as curves for average hroughpu per GPRS user.3 he Simula io Model We co sider a clus er comprisi g of sever hexadiago al cells i a i egra ed GSM/GPRS e work, servi g circui -swi ched voice a d packe -swi ched da a calls. he performa ce s udies prese ed i Sec io 4 were co duc ed for he i ermos cell of he seve cell clus er. We assume ha GSM a d GPRS calls arrive i each cell accordi g o wo mu ually i depe de Poisso processes, wi h arrival ra es GSM a d GPRS, respec ively. GSM calls are ha dled circui -swi ched, so ha o e physical cha el is exclusively dedica ed o he correspo di g mobile s a io . Af er he arrival of a GPRS call, a GPRS sessio begi s. Duri g his ime a GPRS user alloca es o physical cha el exclusively. I s ead he radio i erface is scheduled amo g differe GPRS users by he Base S a io Co roller (BSC). Every GPRS user receives packe s accordi g o a specified workload model. he amou of ime ha a mobile s a io wi h a o goi g call remai s wi hi he area covered by he same BSC is called dwell ime. If he call is s ill ac ive af er he dwell ime, a ha dover oward a adjace cell akes place. he call dura io is defi ed as he amou of ime ha he call will be ac ive, assumi g i comple es wi hou bei g forced o ermi a e due o ha dover failure. We assume he dwell ime o be a expo e ially dis ribu ed ra dom variable wi h mea 1/h,GSM for GSM calls a d 1/h,GPRS for GPRS calls. he call dura io s are also expo e ially dis ribu ed wi h mea values 1/GSM a d 1/GPRS for GSM a d GPRS calls, respec ively.
o exac ly model he user behavior i he seve cell clus er, we have o co sider he ha dover flow of GSM a d GPRS users from adjace cells. A he bou dary cells of he seve cell clus er, he i e si y of he i comi g ha dover flow ca o be specified i adva ce. his is due o he ha dover ra e ou of a cell depe ds o he umber of ac ive cus omers wi hi he cell. O he o her ha d, he ha dover ra e i o he cell depe ds o he umber of cus omers i he eighbori g cells. hus, he i era ive procedure i roduced i is used o bala ce he i comi g a d ou goi g ha dover ra es, assumi g ha he i comi g ha dover ra e h GSM i i , ( ) ( ) &mi us;1 compu ed a s ep i-1. Si ce i he e d- o-e d pa h, he wireless li k is ypically he bo le eck, a d give he a icipa ed raffic asymme ry, he simula or focuses o resource co e io i he dow li k (i.e., he pa h BSC →B S →MS) of he radio i erface. Because of he a icipa ed raffic asymme ry he amou of upli k raffic, e.g. i duced by ack owledgme s, is assumed o be egligible. I he s udy we focus o he radio i erface. he fu c io ali y of he GPRS core e work is o i cluded. he arrival s ream of packe s is modeled a he IP layer. Le be he umber of physical cha els available i he cell. We evalua e he performa ce of wo ypes of radio resource shari g schemes, which specify how he cell capaci y is shared by GSM a d GPRS users: he s a ic scheme; ha is he cell capaci y of physical cha els is spli i o GPRS cha els reserved for GPRS da a ra sfer a d GSM = - GPRS cha els reserved for GSM circui -swi ched co ec io s. he dy amic scheme; ha is he physical cha els are shared by GSM a d GPRS services, wi h priori y for GSM calls; give voice calls, he remai i g - cha els are fairly shared by all packe s i ra sfer. I bo h schemes, he PDCHs are fairly shared by all packe s i ra sfer up o a maximum of 8 PDCHs per IP packe (&quo ;mul islo mode&quo ;) a d a maximum of 8 packe s per PDCH . he sof ware archi ec ure of he simula or follows he e work archi ec ure of he GPRS e work . o accura ely model he commu ica io over he radio i erface, we i clude he fu c io ali y of a BSC a d a B S. IP packe s ha arrive a he BSC are logically orga ized i wo dis i c queues. he ra sfer queue ca hold up o Q &sdo ;8 packe s ha are served accordi g o a processor shari g service discipli e, wi h he umber of physical cha els ha are po e ially available for da a ra sfer, i.e. = GPRS u der he s a ic scheme a d = u der he dy amic scheme. he processor shari g service discipli e fairly shares he available cha el capaci y over he packe s i he ra sfer queue. A arrivi g IP packe ha ca o e er he ra sfer queue immedia ely is held i a firs -come firs -served (i case of o e priori y) access queue ha ca s ore up o K packe s. he access queue models he BSC buffer i he GPRS e work. Upo ermi a io of a packe ra sfer, he IP packe a he head of he access queue is polled i o he ra sfer queue, where i immedia ely shares i he assig me of available PDCHs. For his s udy, we fix he modula io a d codi g scheme o CS-2 .
Where did you get that idea that given a problem to solve, the first thing you need is an interactive programming environment? Ingalls: ThatБЂ™s a good question. I think anything where you get immediate gratification feeds on itself by definition. Seibel: So what was that first place that you had that immediate gratification? Ingalls: There were a couple of experiences. I had a chance to work with a semi-interactive PL/I. And a friend of mine was working at an IBM where they had an interactive APL environment. I canБЂ™t remember which of those was first. I really remember the APL one. That affected me in a number of ways because of both the immediacy of interactionБЂ”seeing your results come backБЂ”and the expression evaluation, which is really different from FortranБЂ™s statement-oriented programming. You still see that now, itБЂ™s amazingБЂ”the whole C/C++/Java tradition, although it goes in an object-oriented direction, is still statement-oriented. But if itБЂ™s convenient to make expressions, it really makes the experience different
1. Find the right sports program for your kids
2. The people must fight for their laws as for their walls
3. Поиск культурных корней Американцев (Looking for cultural roots of Americans)
4. The manager as a teacher: selected aspects of stimulation of scientsfsc thinking
5. The face of every city is different. Washington D.C.
10. The Medieval Tower: A refuge and a base for royal power
11. Multiple Intelligences in the structure of a new English syllabus for secondary school
12. Would lower oil prices be good or bad news for the world economy
13. The effect of light intensity on the amount of chlorophyll in “Cicer arietinum”
14. Consequence of building the National Missile Defense
15. The Economy of Great Britain
16. Royal Dutch Shell: Evaluation of Oil Reserves
17. История Советсткого флота (History of the Soviet fleet)
18. Роль СМИ в современном мире (The mass media in the life of Society)
19. Comparison of the Renaissance and Enlightenment (Сравнение Ренессанса и Просвещения)
21. On the problem of crystal metallic lattice in the densest packings of chemical elements
25. Does the crime pay or not ?
26. Higher Education in The U.K.
27. Motivation: Reward system and the role of compensation
29. Russia and the international economy
31. The capturing
32. The cinema in Russia today
33. The fastest computers of the world
34. The Impact of the Afghan War on soviet soldiers
35. The Irish and South America
36. The People Trade
37. The Renaissance
41. Geometrization of the Fundamentals of Chemistry
42. The Economy of Great Britain
43. Особенности ведения бизнеса в Китае (The peculiarities of marketing strategy development in China)
44. The New-York City, Places of interest
45. Династия Тюдоров (essay the house of Tudor)
46. Raskolnikov and Svidrigailov: on the brink of suicide. Ф.М. Достоевский, Преступление и наказание
47. The influence of the Great Britain on our daily life
48. The history of Old English and its development
50. The Consequences of the Soviet-Afghan War
52. The Cinema World. Moden film festivals and film industry and stars
53. Entertainment outside the home: Pubs and clubs Leisure and sports
57. British painting in the 17-18th centuries (Британская живопись 17-18 вв.)
58. The Influence of English Mass Culture on Estonia
59. Raskolnikov and Svidrigailov: on the brink of suicide. Ф.М. Достоевский, Преступление и наказание
60. History of `The Beatles` and biographies of members in english
61. The Union Jack
62. The History of Alaska (история Аляски)
63. The fastest computers of the world
67. Alaska’s Wildlife: on the Verge of Extinction (Живая природа Штата Аляска на грани исчезновения)
68. Education in the Middle Ages
69. The Consequences of the Soviet-Afghan War
73. Russia and the international economy
75. Финансовые инструменты ("Financial Instruments. Teaching materials of the course")
76. Redesigning the Dragon Financial Reform in the Peoples Republic of China
77. The Impact the Civil War 1861-1865 on Economic, Politic and Industry Development in the USA
79. the Great Sailor
80. Мальтийский еврей (The Jew of Malta)
81. The JAZZ Story
83. Анализ деятельности the Coca-Cola company
84. The practice of modern medicine
85. The teaching of Hugo Gratius
89. The Value Based Leadership Theory
90. История английского языка по книге Bill Bryson The Mother Tongue
91. Душа (The Soul)
92. Прошлое экологического движения (The Environmental Movement As It Was)
93. Under what conditions will the oligopolists agree to co-operate in their decisions
95. Метафорический потенциал слова и его реализация в поэме Т.С. Элиота The Waste Land
96. The United States of America
97. Polysemy In The Semantic Field Of Movement In The English Language