ITU-R S 1711-1-2010 Performance enhancements of transmission control protocol over satellite networks《卫星网络传输控制协议的性能提升》.pdf

1、 Recommendation ITU-R S.1711-1(01/2010)Performance enhancements of transmission control protocol over satellite networksS SeriesFixed-satellite serviceii Rec. ITU-R S.1711-1 Foreword The role of the Radiocommunication Sector is to ensure the rational, equitable, efficient and economical use of the r

2、adio-frequency spectrum by all radiocommunication services, including satellite services, and carry out studies without limit of frequency range on the basis of which Recommendations are adopted. The regulatory and policy functions of the Radiocommunication Sector are performed by World and Regional

3、 Radiocommunication Conferences and Radiocommunication Assemblies supported by Study Groups. Policy on Intellectual Property Right (IPR) ITU-R policy on IPR is described in the Common Patent Policy for ITU-T/ITU-R/ISO/IEC referenced in Annex 1 of Resolution ITU-R 1. Forms to be used for the submissi

4、on of patent statements and licensing declarations by patent holders are available from http:/www.itu.int/ITU-R/go/patents/en where the Guidelines for Implementation of the Common Patent Policy for ITU-T/ITU-R/ISO/IEC and the ITU-R patent information database can also be found. Series of ITU-R Recom

5、mendations (Also available online at http:/www.itu.int/publ/R-REC/en) Series Title BO Satellite delivery BR Recording for production, archival and play-out; film for television BS Broadcasting service (sound) BT Broadcasting service (television) F Fixed service M Mobile, radiodetermination, amateur

6、and related satellite services P Radiowave propagation RA Radio astronomy RS Remote sensing systems S Fixed-satellite service SA Space applications and meteorology SF Frequency sharing and coordination between fixed-satellite and fixed service systems SM Spectrum management SNG Satellite news gather

7、ing TF Time signals and frequency standards emissions V Vocabulary and related subjects Note: This ITU-R Recommendation was approved in English under the procedure detailed in Resolution ITU-R 1. Electronic Publication Geneva, 2010 ITU 2010 All rights reserved. No part of this publication may be rep

8、roduced, by any means whatsoever, without written permission of ITU. Rec. ITU-R S.1711-1 1RECOMMENDATION ITU-R S.1711-1 Performance enhancements of transmission control protocol over satellite networks (Question ITU-R 263-1/4) (2005-2010) Scope Most of the current IP transmissions use transmission c

9、ontrol protocol (TCP) as transport protocol. However the TCP protocol presents some shortcomings when used in satellite networks. Therefore various techniques, collectively referred to as “TCP performance enhancements”, were developed in order to overcome these limitations. This Recommendation provi

10、des test results and measurements of a number of such techniques in order to get a more accurate view of their effectiveness and appropriateness depending on the type of satellite network. The ITU Radiocommunication Assembly, considering a) that satellite systems are being used increasingly for Inte

11、rnet protocol (IP) packet transmissions, in particular providing broadband applications directly to users in addition to their role as backbone links; b) that transmission of IP packets on satellite links requires performance objectives different from those contained in Recommendation ITU-T G.826 an

12、d Recommendations ITU-R S.1062 and ITU-R S.1420; c) that the performance of transmission control protocol (TCP) may suffer from degradation due to long satellite transmission delay, which affects the quality of service of end-users applications; d) that the enhancement of TCP performance is therefor

13、e critical in designing satellite links to carry IP packets; e) that radio frequency resources are not efficiently utilized without implementing any performance enhancement of TCP over satellite in some network environments, noting a) that enhancing the performance of TCP may not be required for low

14、 throughput links (see 3.1 of Report ITU-R S.2148), recommends 1 that the reference models, set out in Annex 1 of this Recommendation, should be considered as a basis when developing methods to enhance TCP performance over satellite links; 2 that system designers implementing TCP connections in netw

15、orks including satellite links should assess the appropriateness of a specific TCP performance enhancement method for their system based on the results of simulations and measurements as contained in Annex 2; Rec. ITU-R S.1711-1 23 that, in order not to affect TCP throughput, satellite links intende

16、d to carry such TCP-based transmissions should be designed to ensure that the bit-error ratio is better than 107during the available time (see 2 of Annex 2); 4 that, in order to improve the throughput of TCP-based transmissions over networks including satellite links: 4.1 splitting techniques (wheth

17、er associated or not to caching) should be implemented when the network topology permits it (see 2, 3 and 5 of Annex 2); 4.2 the TCP window scale option should be used to set the TCP congestion window to the maximum feasible value (see Note 1 and 4 of Annex 2); 5 that the following Note 1 should be

18、considered as part of this Recommendation. NOTE 1 Large TCP windows generate burst traffic, which could result in packet losses in intermediate routers due to buffer saturation. Buffer saturation can be alleviated by implementing larger buffers in the intermediate routers or traffic control at the T

19、CP source (see 4 of Annex 2). In cases where such implementations are not practical, the maximum value of the TCP congestion window should be set by taking into account the potential buffer saturation at intermediate routers. NOTE 2 Report ITU-R S.2148 provides background material on some shortcomin

20、gs of the TCP when used in satellite networks as well as an overview of TCP performance enhancements, briefly describing them and indicating the areas where they improve the performance of TCP over satellite networks. NOTE 3 Section 6 of Annex 2 contains results of measurements performed to assess t

21、he effectiveness of enhancements of two typical TCP-based applications (file transfer protocol (FTP) and hypertext transfer protocol (HTTP). Rec. ITU-R S.1711-1 3TABLE OF CONTENTS Page Annex 1 Satellite system reference models 7 1 Scope 7 2 Reference models . 7 2.1 Point-to-point links . 7 2.2 VSAT

22、networks 7 2.2.1 Star topology 7 2.2.2 Mesh topology 8 2.3 Broadband access 8 Annex 2 Tests and measured performance of TCP enhancements . 9 1 Scope 9 2 TCP performance with splitting enhancement . 9 2.1 Single TCP connection performance tested without any improvement . 9 2.1.1 Single TCP connection

23、 performance testing 9 2.1.2 TCP without performance enhancement 9 2.1.3 Performance of single TCP connection without performance enhancement . 9 2.2 Two-segment splitting 10 2.2.1 Configuration of satellite networks 10 2.2.2 Tests results 12 2.2.3 Summary 18 2.3 Three-segment splitting 18 2.3.1 Con

24、figuration of satellite networks 18 2.3.2 Test procedures 19 2.3.3 Test results . 19 2.3.4 Summary 24 3 Tests and measurements of TCP over satellite using caching and spoofing 24 3.1 Introduction . 24 3.2 Tests and measurements . 25 Rec. ITU-R S.1711-1 4Page 3.2.1 Equipment specifications . 25 3.2.2

25、 Network configurations for tests and measurements . 26 3.2.3 Test and measurement results 27 3.3 Summary . 29 4 TCP performance over a satellite ATM network . 29 4.1 Network configuration 29 4.2 TCP performance in ATM-based satellite-only network . 31 4.3 TCP behaviours in heterogeneous networks in

26、cluding satellite link 32 4.3.1 Simulation of buffer size 33 4.3.2 TCP performance with traffic control in TCP sender 35 4.4 Summary . 37 5 TCP performance in satellite access networks . 37 5.1 Network architecture and configuration . 37 5.1.1 Network architecture 37 5.1.2 Configuration . 38 5.2 Res

27、ults of performance measurement . 40 5.2.1 Availability . 40 5.2.2 Throughput . 40 5.2.3 Traffic . 42 5.3 Summary . 42 6 Application protocol measurements (FTP and HTTP) . 42 6.1 Satellite ATM network configuration . 43 6.2 FTP throughput over OC-3 satellite link 44 6.3 HTTP throughput over OC-3 sat

28、ellite link . 46 6.3.1 HTTP 1.0 with non-persistent connections 46 6.3.2 HTTP 1.0 with “keep-alive” option . 46 6.3.3 HTTP 1.1 without pipelining . 46 6.3.4 HTTP 1.1 with pipelining 47 6.3.5 Test results . 47 6.4 Summary . 50 7 Conclusions 50 Rec. ITU-R S.1711-1 5List of acronyms AAL ATM adaptation

29、layer ACK Acknowledgement ATM Asynchronous transfer mode BDP Bandwidth delay product BER Bit-error ratio BW Bandwidth CE Congestion experience bit CPU Central processing unit cwnd Congestion window (variable in TCP) DA Dedicated access DACK Delayed acknowledgement DAMA Demand assignment multiple acc

30、ess DVB-S Digital video broadcast via satellite ECN Explicit congestion notification e.i.r.p. Equivalent isotropically radiated power FEC Forward error correction FIN Final segment (in a TCP connection) FTP File transfer protocol G/T Gain to equivalent system temperature ratio GSO Geostationary sate

31、llite orbit GW Gateway HPA Hub page accelerator HSP Hub satellite processor HTML Hypertext markup language HTTP Hypertext transfer protocol ICMP Internet control message protocol IETF Internet engineering task force I/O Input/output IP Internet protocol IPSEC IP security protocol ISP Internet servic

32、e provider LAN Local area network LFN Long fat network MF-TDMA Multifrequency time division multiple access MPEG Moving picture experts group MPLS Multiprotocol label switching MSS Maximum segment size MTU Maximum transmission unit NNTP Network news transport protocol Rec. ITU-R S.1711-1 6NTP Networ

33、k time protocol OS Operating system PAD Padding bytes PAWS Protect against wrapped sequence(s) PC Personal computer(s) PDU Protocol data unit PEP Performance enhancing proxy RA Random access RAM Random access memory RBP Rate-based pacing rcvwnd Receive window (variable in TCP) RFC Request for commen

34、ts (issued by the IETF) RPA Remote page accelerator RS Reed-Solomon RTT Round trip time RTTM RTT measurement Rx Receiver SACK Selective acknowledgment SCPC Single channel per carrier SSPA Solid state power amplifier ssthres Slow start threshold (variable in TCP) SYN Synchronous start segment (used t

35、o establish a TCP connection) T/TCP TCP for transactions TBF Token buffer filter TC Trellis coded TCP Transmission control protocol TDMA Time division multiple access TWTA Travelling wave tube amplifier Tx Transmitter UDP User datagram protocol URL Uniform/universal resource locator VSAT Very small

36、aperture terminal WAN Wide area network Rec. ITU-R S.1711-1 7Annex 1 Satellite system reference models 1 Scope This annex presents reference models of networks including a satellite link, to carry IP packets, followed by a description of the limitations of TCP over satellite links. 2 Reference model

37、s 2.1 Point-to-point links Figure 1 provides a reference model for a network carrying IP packet transmissions. The network consists of a satellite link and associated terrestrial networks between two end-users. The satellite link is bidirectional and consists of link AB (from earth station A to eart

38、h station B with an information bit rate, RAB) and of link BA (from earth station B to earth station A with an information bit rate, RBA). The terrestrial networks can employ various data link layer protocols (e.g. ATM, frame relay, MPLS). FIGURE 1 Reference model for a point-to-point link including

39、 a satellite link NOTE 1 The reference model above considers only one satellite hop. Throughout this Recommendation, the techniques that segment the TCP connection to improve TCP performance over satellite links are described for one satellite hop. However an end-to-end connection may include severa

40、l satellite hops. In this case, such techniques will have to be implemented over each individual satellite link. 2.2 VSAT networks 2.2.1 Star topology Figure 2 depicts the standard star configuration in which signals from various remote users connect to a gateway earth station which in turn connects

41、 to terrestrial network. 1711-01Satellite systemLink BA ( )RBALink AB ( )RABTerrestrialnetworkUser 1Earthstation AEarthstation BTerrestrialnetworkUser 2Satellite linkRec. ITU-R S.1711-1 8FIGURE 2 Star topology 2.2.2 Mesh topology Figure 3 illustrates a mesh configuration whereby any pair of earth st

42、ations can be connected directly via satellite. FIGURE 3 Mesh topology 2.3 Broadband access Even if not completely similar to VSAT networks, broadband access networks generally use the same topologies (i.e. star or mesh). 1711-02C NCCCCommunity NCommunity 1Internetbackbone1711-03Community NCommunity

43、 1Rec. ITU-R S.1711-1 9Annex 2 Tests and measured performance of TCP enhancements 1 Scope Annex 2 presents the results of independent tests and measurements conducted to experiment some of the TCP performance enhancement methods and provides valuable information for satellite system designers. For f

44、urther details on the various TCP performance enhancement methods, see Report ITU-R S.2148. 2 TCP performance with splitting enhancement INTELSAT and KDDI have carried out measurement of TCP performance using segment splitting techniques (two-segment splitting and three-segment splitting) to verify

45、the effectiveness of these techniques. This section provides the results of these measurements. Section 2.1 presents the performance test results of a single TCP connection without any gateway. Section 2.2 contains the test results of the two-segment splitting technique. Section 2.3 gives the test r

46、esults of the three-segment splitting technique. 2.1 Single TCP connection performance tested without any improvement 2.1.1 Single TCP connection performance testing The first set of tests examined the throughput of a single TCP connection. Tests were run with and without protocol gateway enhancemen

47、t for round trip delays of 200 ms, simulating a terrestrial WAN connection, and 700 ms, simulating a combined satellite link into the terrestrial backbone. 2.1.2 TCP without performance enhancement The first set of tests was for single TCP connections without performance enhancement. The client wind

48、ow size was set to 8 kbytes to match the default settings on Windows 95, Windows 98, Windows NT, and many other common operating systems. The terrestrial link RTT delay was set to 200 ms and the combined satellite/terrestrial link to 700 ms. 2.1.3 Performance of single TCP connection without perform

49、ance enhancement The maximum throughput without performance enhancement is 320 kbit/s for the terrestrial connections and 91 kbit/s for the satellite link (see Fig. 4). These results demonstrate that without performance enhancement, the maximum single-connection TCP throughput rate will be approximately equal to the window size, kbit/s320ms200bits8kbytes8=RTTeven if the link rate is increased. Rec. ITU-R S.1711-1 10FIGURE 4 Single TCP connection over terrestrial and