スループット監視に基づく複製サーバ選択方法のスケーラビリティ向上
6
0
0
全文
(2) 1 32ÿ 1 {{Ç{Ç«Òb6ÇôÜ·HÌQèé]8 (¼{ÇêÇÄHÌÌèÇ· 2. fhiromori,h-yamagu,higashino,[email protected] [email protected]. â¶IP ÑØÉÂÔÇËÌÊËÚô¡° 6µÅÍ£ª¾«¾Ýã À 6µÜ³½ ø± òß¨ï µÑØÉÂÔÇËÇË×Ó²¨° ©½3ê±äü4ú£ö5±ýþà³2 ¦±ö(´ÅÍç· øÙÐØ ÇË×ÓÀ쩽8!÷ÀÛ¨°¤°½§µ!÷´¡°¶ü4ú¶9ÇË×Ó´¯°ªµ ë* ´Þ®°æõÎÄÊËÀ 4¨2 ýþÀñ¬¹6µü4ú£ ô©½åé´¡°¶§µËÖÏÁÊõ
(3) £$ ²³½#ð±¶ïµÇË×Ó´¯°½Ý6µü4úµ¸£æõÎÄÊËÀ 4©½§²±ü4ú6£
(4) ¬ 1ó±ºæõÎÄÊË´»½ËÖÏÁÊÃÀ¥' ½!"ÀÛ©½æõÎÄÊ˵6À' ½²ª¾¦¼¾½2 £ÿ³¥³½£Û!÷±¶äü4ú£û¨¬2 À9°µü4ú±è%©½§²±ÌÊËÚÃÆÒÕÙÈÀ& ¬ùîíᢼÌÊËÚÃ0´)¾½æõÎÄÊ˵6À' ³£¼è%¨¬æõ2 À&°7´ÇË×Ó8 £ñ ½§²£¿¢®¬ Improving Scalability in Monitoring-based Multicast Server Selection. 1. 1. 2. Akihito Hiromori , Hirozumi Yamaguchi , Keiichi Yasumoto ,. 1 Teruo Higashino. 1. 1 and Kenichi Taniguchi. Graduate School of Engineering Science, Osaka University. 2. Faculty of Economics, Shiga University. In our previous work, we have proposed a selection technique for replicated multicast video streams from di erent servers. In the technique, we have assumed that there exist replicated video servers, each of which encodes the same video into the ones of di erent quality levels and continuously transmits them as di erent multicast streams. Using an IGMP facility like mtrace, each receiver monitors the packet count information of all those streams on routers and periodically selects the one which is expected to provide low loss rate and to be suitable in the current available bandwidth of the receiver. However, under the exsistense of a large number of receivers, the amount of tracing trac may cause a signi
(5) cant problem. In this paper, we improve the monitoring technique used in the selection technique to keep the amount of tracing trac low even if the number of recever is large. In order to reduce the amount of tracing trac, only some part of receivers send mtrace queries for each stream. Also, each receiver selects a stream using the shared infomaion which is obtained by the other receivers. Our experimental results using a network simulator have shown that the amount of tracing trac is reasonable and receivers could select appropriate streams.. R Î+MS+Ã2J07 *êQy.f 6K ^} M8c`3+à 2J0SaLRy.Qv? å¦+ .pÒµRU¹®Q¯2+å2> f /
(6) Ac,<RF]+ 2J0Q7jµAcN å+j ë_mm PO+s9R u.kR~oy.|Q {ñR pl [3]. 1. »is 7 .pFMªh>dc<N7 ßd>dc,?6?+ª³Q»is S·Å {àñM1cF]+»iªJu.QsQRhp {y74Ac`3PGQS .pw©f¡ ¬Ac,ÍhçihªhMSw©f÷¸?P2 ~oyJ [1] 7ÝâM1c7+ªRu.M »iRªhMS+¢?L»if2JM8P2< N^ 4adc,<dQv?+åMSÅW~o yu.fÆ ª?F^NM2J07YPu. f^}AcÎ+QJ2L+29J6ôÜ7 ]ad L2c [2, 3], º-S<d[MQ+IP .pQ5;cÅW ~oy »iu.R^}Î+f£?L8L 2c [3],<RÎ+MS+ÅQR~oyu. 7@ »is f¦PcÅQR¿.Q ò?+¡ì?FÉ-R~oyy. N?LªhAc,DR^NM+Ã2J0S~o y.|FMRy.RrQHÌ (h> dFrRQ) f N23.fâ2L ÓQË"?+DRHÌf^NQYP¿.7 ad cy.fQ^}Ac, mtrace. mtrace. .{.x ( Q`cË" {.x) 7 4 Acâd71c,y.^}QSr QH Ì7Ý3M1c<NS [3] M)>dL5b+ R /
(7) f;QäRAcNYPy.^}7
(8) 4 P2¶§S71c7+{ÖÖP .pÊáR ¯S (yr.i) f5F>CcF]+ Q ¯3Ë"ipfíú>Ccºã71c, Ñ MS+Çó [3] RÎ+Q52L+Ë"i pfíú?P7a+6JYQy.7
(9) edc `3Ã2J0R R/
(10) vE+ÛU/
(11) ÀQJ 2LRñf
(12) 3,£Î+MS+Ã2J0R /
(13) vES2J?L2cy.R\N?+2J0 S /
(14) ï¸fa2J0Q~oyhJA c<NQ`IL+Ã2J07ay.QJ2LR rQHÌfË%M8c`3QAc,<dQ`b+ y.^}QºãPr QHÌf1 ?JJ Ë"ipfíúAc<N7M8c,Q^Ðea B+2J0Q7>aPcjµQ`b+u.kQå2 ~oy.|Q52L+ pl. {.xRQ7jµAc<N^ 4adc,6IL+ý-. 1 −31−. mtrace. mtrace. mtrace. mtrace. mtrace. mtrace. mtrace. mtrace.
(15) Ry.Qv?L+ f/
(16) M8c2J0R 3H+Ë"Ac&ÏÏÆÒ+DR&ÏQ/Q f/
(17) Ac2J0Qf1cªQ¤±QRü>dc` 3QAc,<R`3QÃ2J0R /
(18) Àf ±>Cc<NM+s9R2J07pAc`3PÊ áQ52L^+Ë"ipf9ä4c<N7M 8c, º-S .pw.|. [4] fâ2L+ Çó [3] RÎ+N£Î+f·ÅAcw.w f
(19) IF,/÷ï¸6a+[3] RÎ+N·Å?L Q`cip75Æ2^RNPIL5b+. Q 100 R&QSu.kR~oy.|F MSÈä 2 Kbps N/âQÚ|P2NPIL2F,[ F+Ã2J07 /Q2JM8Fy.R¿.S ¤_RÎ+NYZ RNPIL5b+y. R^}7YQ
(20) edL2c<N7Ä¥M8F, mtrace. server S1 (video transmission). server S2 (video transmission). mtrace. gr(2,2),1. a. mtrace. c. b. d. e. gr(1,2),1. ns. f. g. h. k. mtrace. 2. GB" )CD 2!1%£Î+MS+Ãu.SªR»is f. ¿./»iujz+¾i+. QPO0MÃAc,¤±MS<R¿.f* `QM¼?+1 7¿., L S ¿.f¼A,[F+ u. Si 6aªh>dc¿. l R»i. |y.fQy.N`U sti;l M¼A,à .|SÇó [5] M)>dL2cË"Õ§f/g?L 2c^RNAc,<RÕ§f IF IP ~oy pË"âRz¼P.Q [5] 7 1c, S~oyq.hyNh Jyhyf!Ac<NM+hJy6a f/
(21) Ac.vy[MR~oy ªhêöÛUDRêöFRÃ~oy.|M h>dF¤8òrQf1 Ac<N7M8c, Ã2J0Sª¾1FbQ.|Mh>dcr Qf cF]Q+ pl.fÓQhc,p l.Rï¸SÃ.|MRr QNpl.R &NPIL5b+KRpl.ï¸ 2 J6a Ã.|Mª¾1FbQh>dFr Qfì L R¦Pc. mtrace. mtrace. mtrace. mtrace. Ac<N7M8c, 2.1. dk{ozrwfej . 2J0 Rj S¤±RHÌf_^ILIL2c+[ FSË"Q`IL adc^RN³Ac, bwi;l; y. sti;l RhJ.,Ãy. S1a6@]e6IL2cÿ.Mªh> dc^RNAc, pathi;j ; u. Si 6a2J0 Rj [MRªhêö, pl.Rï¸6a c<N7M8c, gr i;l ;j ; 2J0 Rj Q vAcy. sti;l R qi q.|,qi q.|S+ ªhêö pathi;j FR.|R3Hy. sti;l f2J?L5b+2J0 Rj Q^å2. |f¼A, pl.Rï¸6aªhêö pathi;j FMy. sti;l RrQ7 0 M P20R.|N?L®õM8c, ratio i;l @r; y. sti;l R.| r MR r ë,¤± count i;l @r M.| r M ª¾1FbQh>dFrQf¼A,<d mtrace. (. ). mtrace. (. ). (. ). m. n. receiver R1. receiver R2. st1,1. st1,2. p. q. receiver R3. t. u. receiver R4. receiver R5. receiver R6. st2,2. st2,1. st2,1. st1,2. low bandwdith video stream (level1) from S1 high bandwidth video stream (level2) from S1. low bandwdith video stream (level1) from S2 high bandwidth video stream (level2) from S2. P 1: y.^}Rô S pl.fÓQhc<NM 1 Mr count 8c,<df â2L+ratioi:l @r S count S N?Lì M8c,P5+count i;l @Si Su. Si My. sti;l Rr7ª¾1FbO dG;hJ>dF6)A^RM+ª³QS<R rQfìc<NSM8P2,?6?+Si Q nAc@ LAN R.|M¨Ô>dcr SYNgOP2NAdT+<RS<R.|M h>dFQNYZ@GN 4adc,ª³ Q+üû7DR`3P.|MØ<c<NS[d M1b+<R³SP^RGN 4adc, ratio i;l @Rj ; 2J0 Rj MRy. sti;l R R rë,count count S N?Lì M8c, qtzcwk{ozrwh P 1 MS 2 JRu. S , S 7DdEdªR» is f 1, 2, 3 R 3 JR¿ ./<RôMSDdEd¿.+¿.+ ¿.0 M~oy hJ?L5b+2J0 Rj (j = 1::6) SDdEd1cu.6aR1c¿.Ry. f2J?L2c,ô4T R Su. S 6aR 1 Ry. st ; f+R S st ; f2J?L 2c, /Q adcy.R¿.S2J?P;dT e6aP27+£Î+MS+qi q.| MRr ë+qi q.|6a2J 0[MR Qf2J0Q adcy.R¿ .R!»NAc,<R^NM+Ã2J0SÒÑQ+2 J0yMRû2JRy.Rr2 Jë75Æ¢?L2cN°M8cGS`b¿. R 2y.f+¢?L2P2GS` b2y.fQ^}Ac,DR+/R 2J0Rrë7PcX9 2y.f ^}AcF]+.|RË"ï¸QÒK8+qi q.|MRr ë7Ò:¤FM+[F 2J06aqi q.|[MR Q7 ^?>2y.f^}Ac,<RÒ:S»i RVQYNgO Ú|7P2N°M8cYO 5Æ 2r ëf¼A^RNAc,<R`3QAc mtrace. (i;l) @. (i;l) @. (. (. ). ). j i. (i;l) @. (i;l) @. 2.2. 2 −31−. 1. 2. 1. 11. 1. 5. 23. i.
(22) èÞS+qi q.|6a2J0[MR ipSÎlM8P2F]+qi q.|M Rr ë7DR[[2J0MRr ëNPaP2¶§S71cF]M1c,£Î+MS+ qi q.|6a2J0[MRÞé7^? >2y.f^}Ac<NM+DRy.f /Q2J?FN8QÓxAc¿.7 adP2¶§ Sfú@>CL2c, ô4TP 1 Q52L+2J0 R MR st ; Rr ë75Æ¢?L2cN°M8FGQS+R S`b 2¿.M»if2JAcF]Q+ 2 R y.f2J?`3N#\c,[B+R S 2 R aLRy.QJ2Lqi q.| gr ; ; (path ; = a-c-g-m FR g)+gr ; ; (path ; = b-d-f -m F R d) f ]c,ratio ; @gr ; ; N ratio ; @gr ; ; (qi q.|MRr ë) ROHa^Ò:¤FNPIL2c^RN Ac,<RN8+R S st ; N st ; R3H+q i q.|WR Q7?>2y.f^ }Ac,<RôMS+R 6a g [MRÞé7 1 +d [MRÞé7 2 NPIL2cRM+R S st ; f^}Ac, áQ+2J0MRrë7Ò:f±¿ cGS+¿.7@6`b2y.R 3H¢?L2JM8cNßd>dc^Rf^}Ac, ô4T+ratio ; @R 7Ò:f±¿IF^RNAc N+R S st ; +st ; +st ; R3HRªJf2J?` 3NAc,ORy.QJ2Lqi q. |MRrë7Ò:f4L2dT+R S ûR¿.N¨%AcF]Q st ; f^}Ac, 1. server S1. server S2. a. b. c. 11. e. d. f. g. h. 1. 1. (1 2) 1. 11. (1 2). (2 2). (1 2) 1. (2 2) 1. 1. 12. 22. 1. 1. 12. (1 2). 3. 22. 3. 11. 21. 3. 22. qtzcwk{ozrw. Ã2J0 Rj S2J?L2cy. sti;l Rr ëQ¯@LÉRy.f^}Ac6+û 2J?L2cy.fëo?L2JAc6R° fªÓÏ=NQ
(23) 3,£Î+MSr ë QJ2L£JRÒ: P , P fâ2L+Rj RIyf (1) \¿.jµIy"+(2) \ëo2JIy"+(3) \ ¿.¨%/ú@Iy" R 3 IyQÆóAc,<< M+P Sy.R2J7YZrq-f¯e BQ
(24) edL5b+`b 2¿.Ry. f2JAcRQ5ÆGN°M8cN?+P Ss@ ry7®V?L2c^RR+»iRVQ SYNgO¬ã7P2N°M8cNAc, 1. 2Jy. sti;l Rr ë7 P (0 < P 1) ¤FM1c2J0 R S \ jµ Iy" QnAc,<RG+Rjj S+¿. l + 1 RÃy. sti ;l Rqi q .| gr i ;l ;j Q5;cr ë70 9Ac P ¤FM1cy.R3H+q i q.|6a Rj [MR Q7?M 1cy.f^}Ac,qi q. |Q52Lª¤FR¿./APeHr ë07/û>dL5b+6JDRqi q.|6aRÞé72y.f^}Ac, DR`3Py.7p?P2GQS Rj S sti;l f2J? o;c, 2. 2Jy. sti;l Rr ë7 P (0 < P < P ) ¤F P ÔÓM1c2J0 Rj S \ë 2.3. 1. 2. 1. 2. 1. 1. 0. (. 0. +1). 2. 1. 1. n. receiver R1. p. q. u. receiver R2. receiver R3. receiver R4. receiver R5. P 2: £JR~oyªh× o2JIy" QnAc,<RG+Rj S sti;l f ëo?L2JAc, 3. 2Jy. sti;l Rr ë7 P Ô ÓM1c2J0 Rj S \¿.¨%1c2S ú@Iy" QnAc,<RG+Rj S¿. l ^?9S l 0 1 Ry.6a+DRq i q.|MRrë7 P ¤ FM1cy.R3H^¿.R 2 y.f^}Ac,DR`3Py.7 ª¥Qî[aP2GS+Rj S qi q .|6a Rj [MRÞé7^2^Rf^} Ac, 3 $ 0*1' <L 7> IJ 2 AM )?F`3Q+£Î+SÃ2J0S fâ2Lr QHÌf1 ?+DRHÌf^NQ y.f^}Ac,ACK/NACK Q`ci. pføQ?F JæS~oyQ52L+ ACK/NACK ®Ac Ú| [6] 7 !>dL2c7+ £Î+Q52L^u.Qå2~oy. |M pl.Rr Q7 j{Ac¶§S 71c,<<M pl.RrRñfÍ QõXIL\c,N NR2J07 TM ¾=NÓQ pl.Rr (F6G6 0.2kb) fhJA c,DdaRrSªhêöFfu.RY3W Nh>du.Q4Ac,L f¿.RQG 2. 2. mtrace. mtrace. mtrace. mtrace. NAcN+u.å9R~oy.|Q5; c RipS¤±R`3QPc, mtrace. 0:2(KB ) 3 8(bit =B ) 3 2N (packet ) 3 L (bps) TM (second ). +1. 2. 2. m. 21. (2 2) 1. ô4T N = 10; 000, TM := 2 and L = 3 NAcN+ ; = 24 NP RipS b+ªh×RQ1Fc.|MSª¾1FbQ 5000 Rpl.f<èAc<NQPc,<RS/âP ^RMSP9+sNQR2J0Qv¯AcF]QS< Rpl.RQfä4cºã71c, |ej~ £Î+MS<RÚ|Qv<AcF]Q+ªJRy .Qv?29J6R2J0R\7 pl. mtrace. 0 238310 00033 2. Mbps. 3.1. 3 −33−. mtrace.
(25) fhJ?L2c,[F+ pl.fhJ?P2 y.QJ2LRHÌf2J0ÏMßÝM8c` 3Q?L2c,¤±M£Î+RDf9Xc, pl.Rï¸fßÝAcF]Q+aLR2J07n ?L2c~oyq. C fs .q .N?Lâ2c,Ã2J0S+y.^}R Qâ2c+r QHÌfË%?L2c,<RH ÌSaLRy.Ru.6aÃ2J0WRªh êöQJ2LR^RM+*M7 /
(26) ?F Rï ¸+CÃ6a adFï¸M9 K >dc, 2J0 Rj Sy.2JÀ _Q+aLR y.QJ2LªhêöfXcF]Q+a LRy.QJ2L pl.fhJ Ac, Ã2J0 Rj Sû2J?L2cy. sti;l QJ2LR\ pl.fhJAcòçf ^J, Ë"ÓÏ TM ¾=NQ+s .q. C QÓRF]R{.x7hJ>dc (u. R3HRªJQ`c), Rj S R!QÆÂQÒK2L7[c } P|j.f^IL2c^RNAc,2J0S DdEdÓ{.xf2JAcN|j.f y|.>Cc, |j.7Ydc[MQ Rj S2J?L2cy . sti;l QJ2L pl.Rï¸f adP2N8QS+HQ pl. {.xfhJAc, Rj 7 pl.Rï¸f F1N+H Qs .q. C Qv?LDdfhJ ?+*M7Ë%AcrQHÌf9KAc, Rj S C Q`IL adFpl.Rï¸M+y .^}RQ â2c+*M7 Ë%Ac rQHÌf9KAc, 2J07hc pl.ReQfRüAc F]Q+Rj SQ!QÆÂR.| f ʲ>Cc,<RØRF]Q Rj S+TM = N C Q`IL adF Rï¸fnk Ac. <df N NAc,N SªÓÏ¢ Q pl.fhIF2J0RQN@M 1c,Rj S'R+M Q`c{R ip (u.kN2J0k) fõX^c, u.k: N mtrace. mtrace. mtrace. mtrace. mtrace. mtrace. mtrace. mtrace. mtrace. mtrace. mtrace. mtrace. 0:2(KB ) 3 8(bit =B ) 3 2M (packet ) (bps) TM (second ). 2J0k : 0:2(. KB. ) 3 8(bit =B ) 3 N (packet ) (bps) TM (second ). M fu.RQNAcN,Ã2J0SF×Rß dipQ`IL'RÓÏR Rfî Ac, ßdipR3H2Bd67s .ipN?LÝ>dcf4 cN+2J0S = 2 NAc,D3MP;dT 1. = ( < 1) NAc,. 1 session delay 9IMFAJDHEGCKN7%)=[`Q8b Y]O5.=5( 9T; 1= = 3 3 session delay 5.=!9 -;7VXU' mtrace 7<=SP?^=0;7 mtrace B@L" ?ZX.=' SP?$=:47 2 3 session delay &&=!/, 3 C 4_Z+>=' /8_Z7 session delay &&=!<13 ?*8<#6\5,0!*8WR\?2 a 63 % 8VX U' C 4 mtrace 8SP?ZX.=*576=!. P 2 fâ2L<Rhjhf\ÕAc,£JRu. S +S 71b+DdaSDdEd st ; +st ; fh J?L2c,2J0 R , R , R S st ; f2J?L 5b+st QR\ pl.fhJM8c (ª Î+R +;R S st ; R\Q pl.fhJM 8c),TM =NQ R , R , R R29J67 pl.fhJM8c,<<MS R +R 7<RÓÏ Qpl.fhc^RNAc,DRN8¶aSªhê öFRÃ.|RKRrQfc<N7M8 (a-c-g -p N a-c-g -q )+DdEdDRï¸f 2J?F0 Q C Q hJAc,ªÎ+R S C 6a adFHÌ 6aDRªhêöF (a-c-e-m) R a, c RHÌS c< N7M8c7+e, m RHÌf c<N7M8P2,< RGQS+R S e, m QJ2L*M7Ë%?L2F ¤_RHÌfâ2c,£Î+MS!QÆÂQÒK2 L7[c }P|j.QQ62Ã2J07 pl.fhJ?L5b+DR3HÃ2J0R ªhêöFRrQHÌ79K>dc<NQPc, FG+u.Qå2.|RHÌS2J0kR^RN ·XLÀ²Q9K>dc<NQPc, 'Q Rj 72J?L2P2y.RHÌQJ2 L 4c,Rj SDR`3Py.Qv? pl.fhJ?P2RM+s .q. C ¤Â6aS<R`3PHÌf c<N7M8B+[F+ ªh×QÑ[dL2P2.|RHÌfc<NSM 8P2,ô4T+R , R S pl.Rï¸f s .q. C QhIL2c7+<dQS R QJ2LRªhêö path ; FR m RHÌSÑ[dL 2P2,?6?P7a+R S b, d, f RHÌfc< N7M8+R Sy.^}QºãPHÌS c< N7M8L2c<NQPc,(s .q. C 6a hadF pl.Rï¸QÑ[dL2c ^RR3H+2J0Q^å2.|7qi q.|NPc,) 1. 2. 11. 1. 4. 21. 11. mtrace. 11. 2. 3. 5. mtrace. 21. 1. 3. mtrace. 4. 3. 4. 1. 1. mtrace. mtrace. 2. mtrace. 5. 1. 21 1. 1. mtrace. . 0Q+>aPcyr.i5FN?L¤± R`3P<N7 4adc, DdEdRy.QJ2Ls .q .fÈbLc<N^ 4adc,<R`3 QAcN+Ã2J0S2J0QNILÐæR1c y.Rs .q.R\Q¤cG ;M`2<NQPb+ Ripe ñS>aQúcN 4adc,4 ARw. w /÷MSÍí²RF]s .q. SªJN?L2c, ÝàM8cs .q.Rip eñN?LS+RTCP [7] R.R1Q1c` 3Q+ª´¿.R2y.RhJ.R 5% 6a 10 % 7 tM1cN 4adc,£ Î+R/÷ï¸MS<RRü¤±M1c<N7) >dF,D?9S 4 AM9Xc, Rï¸fhc+TTL fR1Ac<NM ï¸fhc2J0NS®9édF2J0QDRï ¸fhc<NfÐ:<N7M8c,ª³Q+ èQédL2cDR`3P2J0ÏMSªhê öfßÝ?P2N 4ad+<R`3PR1^[ F3¸M1c, 3.2. −34− 4. mtrace. mtrace.
(27) 4. #,.3#/4?: AE F5 £Î+Rw.w /÷Q`c½´f
(28) 3. F]+[B2J07~oyy.Rªh êöFRr Qf1 AcF]R . pcRsf IP ~oyR. RùèQÒK2LZì?+Ddf .pw.| ns-2 [4] FQ/û?F,Ã2J 0S*M72J?L2cy.Rrë QÒK2Ly.R^}f
(29) 37+^}ht z (2.3 [) MçâAcý P , P SDdEd 95%, 85% N?F, Ã2J07+_R^}0+K?2y.f^} ¶§NPc&ÏS+£JR.| Tup, Tdown fâ 2L¤±R`3QUAc,¿. l Ry. f2J?L5b+¿. l + 1 Ry.f2 J?`3NAcGQS+_R^}6a Tup (=Ó 2 ¾) G;xFP;dTPaP2,<R¿.h 7-©?L l Ry.QÙIFGQS+DR 0RR¿.hS-©Ac¶§S7 2, D<M+¿.hR#\f?Ta9fäRA cF]Q+Tup f 2 «QAc,¿.h7T 2?FGQS Tup f 2 ¾Q=Ó²Ac,ªÎ+û R¿.6`b2Ry.Q§
(30) A cN8S+Tdown xJ^RNAc,w.w /÷MS Tdown f 0.5 ¾Qÿ?L2c,<R`3P 2J.R1ÎÏS+Ác²~oyQ5; cÁc¯ÎÏN?LÇó [8] POMB½>dL2c, mtrace. 1. 2. xcp}. £Î+Q52L2J0auN?L/QOR R¿.f/ûM8c6fXcF]Rw.w /÷f
(31) IF,/÷MS+y. sti;l f2J? F&Ï (ta, tb fDdEd2JÀ &ÏN3ð&ÏNA c) NDRN8Rrë pt 6aDRy. f2J?FN8Q adFauR¿.f l N pt R XÆ (APeH, Rtt l 3 ptdt) NÚ?+Óm NþV, /÷MS+2J0auQv? Ómeùfìl?+Ç ó [3] RÎ+N·Å?F,[3] RÎ+MS+2J0Sa LRy.Qv?L f/
(32) ?L5b+À² Qr QHÌf9K?L2c,ªÎ+£Î+M S+2J0S2J?L2cy.QR\ f /
(33) ?L5b+DRrRy.QJ2LSs .q. C 6arHÌf L2c,6I L+£Î+MS [3] RÎ+Q·X+Ã2J07 Ë% Acr QHÌSÀ²Q9K>dL5aB+DR F]Qy.^}Q¬ã7Mc<N7 4adc F]+£Î+Q52LYQy.^}7
(34) e dL2c6fXF, <R/÷MS+. Q 20 6a 220 R . pf }Q VT?+ì 35 R .pQ52 LDdEd 120 ¾Rw.w f
(35) IF,à pRw©S, 280kbps, 560kbps, 1.1Mbps R2Bd6 N?F,Ãw.w MSu.RQS 3 N?+ .vRQSe. QR±ÆN?F,Ãu.S¿ . 1, 2, 3 R 3 JRy.fhJ?DdaR hJ.SDdEd 256kbps, 512kbps, 1Mbps N? 4.1. b a. mtrace. mtrace. P 3: ÓmeùR. QQ`cʲ. F,[F+~oy.i qsS DVMRP N?F, P 3 MS+ÃÎÏM adFÓmReùf. Q=NQ)?L5b+7,7ÓmReù+°,7. Qf¼A,P 3 `b+2BdRÎ+^. Q 50 [ MSYZ @ NPIL2c<N7e6c,. Q 7j4cN£Î+RS [3] `b^{82Ómeù f~T?L2cN24c,<dS [3] RÎ+Q52L+ 2J0RQ7jµAcNDRÆ Rip 7 jµ?+y.RªhQ¬ãfÛZ?FF]M 1c, [F+Ómeù7 &Ïê¹NN^QMOR`3Q ʲAc6fXF,. Q 50, 100, 200 MRï¸ fP 4 Q)A,. Q 50(P 4 (a)) Rï¸Q52L S+£Î+MS+/÷À 6a5Q¾ÏS [3] RÎ+ N·X+ÓmeùSõILS2c7+/÷3ð&Q S`b 2NPIL2c,. Q 100(P 4 (b)) R ï¸MS+£Î+RÓmeùS/ÉÀ &6a [3] RfF¿IL5b+/÷3ð&QSÜ 30 % F¿IL 2F,. Q 220(P 4 (c)) Rï¸MS+£Î+N [3] NR7`böQ¼dL2c,OR. QQ5 2L^+&ÏRê¹Q¯22J0ÏMßÝAcHÌ7 jµAcN+y.^}7YQ
(36) edc`3Q Pc<Ne6c, mtrace. ~unstgijtyvlsmh. N·Å?LOR Q`cip 7íúM8L2c6lAcw.w /÷f
(37) IF,P5+<Rw.w /÷MS+Ë"Î +MçâAc (3.1) fÿN?+Qip RñfR1?L2P2,u.kMR Q`c Ë"ipRÃ$ôMRÈäîñfP 5 Q)A, P 5 MS°,S. Qf+7,S Q¯3 ipRÈäf)?L2c,[3] RÎ+N·ÅA cN+ipRñS 30 ÆR 1 ¤±NPIL5 b+6PbRipfíúM8L2c<N7e6 c,. Q7jµAcN+ipñ^jµ?L 2c7+DRñS. Q 220 R .pQ52 L^ 6 kbps N5Æ9ä4adL2c, 4.2. [3]. mtrace. mtrace. mtrace. Ñ MS+{ÖÖPÊáQ5;cyr.i f ï?F+IP ~oy .pFMRÅ Wu.fâ2F»iªJQ5;cy.^} Î+Q5;cyr.iR5FQJ2Lñ5 5. 5 −35−.
(38) 1.1e+06 proposed method previous method 1e+06 900000 800000. Quality Values. 700000 600000 500000 400000 300000 200000 100000. P 5:. 0 0. 20. 40. 60 second. 80. 100. 120. =<H9. (a) 50 Nodes 1.6e+06 proposed method previous method. 1.2e+06. Quality Values. 1e+06. 800000. 600000. úLfO 47; ¢Ñèª ÛÙÁ4𪠻i~oyªJQ5;cÅWu.^}h tzR£NS§½´. ~ih+Æ +àNj (DICOMO 2001) w xk øÇ4, pp. 331{336, 2001. [4] Berkeley MASH Research Group University of. 400000. 200000. 0 20. 40. 60 second. 80. 100. 120. (b) 100 Nodes 3e+06 proposed method previous method 2.5e+06. [5]. Quality Values. 2e+06. [6]. 1.5e+06. [7]. 1e+06. 500000. [8]. 0 0. 20. 40. 60 second. 80. 100. Q`cip. [1] C. Diot, J. Crowcroft, and W. Dabbous. Multipoint communication: A survey of protocols, functions, and mechanisms. IEEE Journal on Selected Areas in Communications, Vol. 15, No. 3, pp. 277{ 290, 1997. [2] Z. Fei, M. H. Ammar, and E. W. Zegura. Optimal allocation of clients to replicated multicast servers. In Proc. of 1999 Int. Conf. on Network Protocols (ICNP'99), pp. 69{76, 1999. [3] .. 1.4e+06. 0. mtrace. 120. (c) 220 Nodes. P 4: ÓmeùR&ÏQ`cʲ `U/Cf
(39) IF,Î+Q52LS+1cªÓÏQ Ë"{.xfhJAc2J0fªQQä4+Ë "{.xR4fä4L2c, ª³Qu.^}Qâ2cHÌRKSNyr. iS.mRÐæQ1cN 4adc,Ñ. MS+<R.mRÛf
(40) IL5b+3ë `9 .pRHÌf1J]+6J2J0R 2 Ómf~TM8L2cN24c,0RôÜN?L Shr.w WR¯_+MBONE PO /R .px.MRw.w /÷f 4L2c,. 6 −36−. California. The network simulator ns-2, 2000. http://www-mash.cs.berkeley.edu/ns/. W. Fenner and S. Casner. A \traceroute" facility for IP multicast. In Internet Draft, 2000. Brian Neil Levine and J.J. Garcia-Luna-Aceves. A comparison of reliable multicast protocols. Multimedia Systems, Vol. 6, pp. 334{348, 1998. Henning Schulzrinne, Stephen L. Casner, Ron Frederick, and Van Jacobson. RTP: A transport protocol for real-time application. In Request For Comments 1889, 1996. V. Jacobson, S. McCanne, and M. Vetterli. Receiver-driven layered multicast. In Proc. of ACM SIGCOMM'96, pp. 117{130, 1996..
(41)
関連したドキュメント
Maria Cecilia Zanardi, São Paulo State University (UNESP), Guaratinguetá, 12516-410 São Paulo,
Applications of msets in Logic Programming languages is found to over- come “computational inefficiency” inherent in otherwise situation, especially in solving a sweep of
Shi, “The essential norm of a composition operator on the Bloch space in polydiscs,” Chinese Journal of Contemporary Mathematics, vol. Chen, “Weighted composition operators from Fp,
A Darboux type problem for a model hyperbolic equation of the third order with multiple characteristics is considered in the case of two independent variables.. In the class
[2])) and will not be repeated here. As had been mentioned there, the only feasible way in which the problem of a system of charged particles and, in particular, of ionic solutions
This paper presents an investigation into the mechanics of this specific problem and develops an analytical approach that accounts for the effects of geometrical and material data on
Using variational techniques we prove an eigenvalue theorem for a stationary p(x)-Kirchhoff problem, and provide an estimate for the range of such eigenvalues1. We employ a
One of the goals of this paper was to examine the extent to which the analysis of Carleson measures and interpolating sequences for space of all functions on the tree with
