Presentation is loading. Please wait.

Presentation is loading. Please wait.

Considerations of SCTP Retransmission Delays for Thin Streams Jon Pedersen 1, Carsten Griwodz 1,2 & Pål Halvorsen 1,2 1 Department of Informatics, University.

Similar presentations


Presentation on theme: "Considerations of SCTP Retransmission Delays for Thin Streams Jon Pedersen 1, Carsten Griwodz 1,2 & Pål Halvorsen 1,2 1 Department of Informatics, University."— Presentation transcript:

1 Considerations of SCTP Retransmission Delays for Thin Streams Jon Pedersen 1, Carsten Griwodz 1,2 & Pål Halvorsen 1,2 1 Department of Informatics, University of Oslo, Norway 2 Simula Research Laboratory, Norway {jonped, griff, paalh}@ifi.uio.no LCN 2006: 31st IEEE Conference on Local Computer Networks, Tampa, FL, USA, November 2006

2 2006 Jon Pedersen, Carsten Griwodz & Pål Halvorsen LCN 2006, Tampa, FL, USA, November 2006 Overview Latency problems for thin streams SCTP as an alternative to TCP Experiments New experiments Conclusions

3 2006 Jon Pedersen, Carsten Griwodz & Pål Halvorsen LCN 2006, Tampa, FL, USA, November 2006 Thins Streams Transport protocols being developed for throughput-bound applications  BUT, there exist several low-rate, time-dependent applications Anarchy Online MMORPG Case Study  average delay:~250 ms  max delay: 67 seconds (6 retransmissions)  packets per second: < 4 (less then one per RTT)  average packet size: ~120 bytes  average bandwidth requirement: ~4 Kbps All TCP variations available in Linux (2.6.15) fail to properly support time - dependent “thin streams”  targeted for high rate streams only [nossdav 2006]

4 2006 Jon Pedersen, Carsten Griwodz & Pål Halvorsen LCN 2006, Tampa, FL, USA, November 2006 Stream Control Transmission Protocol Network senderreceiver (re)transmission queue SACK SCTP should support signaling  acknowledged error-free transfers  data fragmentation according to MTU  packet boundary maintenance  sequenced delivery within multiple streams  bundling  partial reliability  …  suppose to address low latencies “require response between 500 – 1200 ms” … or “initiation of error procedures” [rfc 2719]

5 2006 Jon Pedersen, Carsten Griwodz & Pål Halvorsen LCN 2006, Tampa, FL, USA, November 2006 Test Set Up Linux 2.6.15 with lksctp 100 bytes packets 4 packets per second  3.2 Kbps SCTP Network emulated using netem dropp delays (RTTs: 0, 100, 200, 400 ms)

6 2006 Jon Pedersen, Carsten Griwodz & Pål Halvorsen LCN 2006, Tampa, FL, USA, November 2006 Results: lksctp for Thins Streams Even worse than TCP!!! Why these high delays? Two ways of triggering retransmissions of a lost chunk…

7 2006 Jon Pedersen, Carsten Griwodz & Pål Halvorsen LCN 2006, Tampa, FL, USA, November 2006 Retransmission by Time-Out Network senderreceiver (re)transmission queue retransmission of packet with green chunks due to timeout Timeout is dependent on  minRTO = 1000 ms  estimated RTT based on SACKs BUT SACKs are delayed o one ACK for two packets or o 200 ms timer  influences estimated RTT, especially for thin streams  RTO value grows SACK

8 2006 Jon Pedersen, Carsten Griwodz & Pål Halvorsen LCN 2006, Tampa, FL, USA, November 2006 Retransmission by Fast Retransmit Network senderreceiver SACK no SACK no SACK no SACK no 4 SACKs needed for fast retransmit +thin streams = “all” retransmissions due to timeouts

9 2006 Jon Pedersen, Carsten Griwodz & Pål Halvorsen LCN 2006, Tampa, FL, USA, November 2006 Enhancement: Removal of Exponential Backoff Network senderreceiver (re)transmission queue retransmission of green packet due to timeout ENHANCEMENT: remove exponential backoff retransmission number time in RTTS 2 4 6 8 1234

10 2006 Jon Pedersen, Carsten Griwodz & Pål Halvorsen LCN 2006, Tampa, FL, USA, November 2006 Enhancement: Fast Retransmit Bundling Network senderreceiver retransmission queue retransmission of green packet (chunks) due to dupACKs blue packet is NOT piggybacked when dupACKs (but would be if due to timeout) ENHANCEMENT: piggyback all chunks in retransmission queue SACK no SACK no SACK no SACK no

11 2006 Jon Pedersen, Carsten Griwodz & Pål Halvorsen LCN 2006, Tampa, FL, USA, November 2006 Enhancements Modified retransmission timer  removal of exponential backoff  minRTO = 200 ms (as in TCP) Modified retransmission bundling  always allow aggressive bundling for fast retransmit Modified fast retransmit  tested fast retransmit after 1 SACK Thin stream detection  fewer packets in flight to trigger a fast retransmit  added tracking of outstanding packets  less than 4 in flight = thin stream

12 2006 Jon Pedersen, Carsten Griwodz & Pål Halvorsen LCN 2006, Tampa, FL, USA, November 2006 Enhancement: Results (200 ms) Considerable reduction in average and maximum latencies Increase in number of fast retransmissions compared to timeouts  Increase in number of retransmissions original SCTPreduced minRTO & fast retransmit modified timer restartno SACK delay Timeout266197331633 Fast retransmit352842881 Total301481619634

13 2006 Jon Pedersen, Carsten Griwodz & Pål Halvorsen LCN 2006, Tampa, FL, USA, November 2006 New lksctp versions & New Test Set Up New lksctp versions has been developed  lksctp in 2.6.16 (2.5.72-0.7.1) only one retransmission due to fast retransmit, next timeout only 3 SACKs required for fast retransmits  lksctp in 2.6.17 has no major changes for our scenario New tests  100 B packets  RTTs: 0, 50, 100, 150, 200, 250 ms  Packet inter-arrival times: 50, 100, 150, 200, 250 ms  Dynamic thin stream detection  Many web-connections generating cross traffic (and thus losses) SCTP WEB

14 2006 Jon Pedersen, Carsten Griwodz & Pål Halvorsen LCN 2006, Tampa, FL, USA, November 2006 Results: New lksctp  Still high average and worst case latencies

15 2006 Jon Pedersen, Carsten Griwodz & Pål Halvorsen LCN 2006, Tampa, FL, USA, November 2006 Results: Fast Retransmission Modification Reduction in maximum and average latency As expected a large increase in fast retransmit  An increase in spurious retransmissions Fast retransmit modification – 1 SACK

16 2006 Jon Pedersen, Carsten Griwodz & Pål Halvorsen LCN 2006, Tampa, FL, USA, November 2006 Results: Removed Exponential Backoff Reduction in maximum and average latency  An increase in spurious retransmissions Retransmission aggressiveness does not really pose a congestion threat since the amount of data waiting to be sent is always less than the minimum transmission window Removed exponential backoff

17 2006 Jon Pedersen, Carsten Griwodz & Pål Halvorsen LCN 2006, Tampa, FL, USA, November 2006 Results: Reduced Minimum Time-out Faster timeouts Reduction in maximum and average latency  An increase in spurious retransmissions Reduced minRTO

18 2006 Jon Pedersen, Carsten Griwodz & Pål Halvorsen LCN 2006, Tampa, FL, USA, November 2006 Results: All Modifications Combined A further reduction in maximum and average latency As expected an increase in fast retransmit  An increase in spurious retransmissions All thin stream modifications

19 2006 Jon Pedersen, Carsten Griwodz & Pål Halvorsen LCN 2006, Tampa, FL, USA, November 2006 Conclusions Based on SCTP description we expected (hoped for) reduced latencies compared to TCP Enhancements like  reduced minRTO  removal of exponential backoff  removal of delayed SACKs  … reduce latencies for thin streams The enhancements increase the number of spurious retransmissions, but maybe not important for thin streams!!??


Download ppt "Considerations of SCTP Retransmission Delays for Thin Streams Jon Pedersen 1, Carsten Griwodz 1,2 & Pål Halvorsen 1,2 1 Department of Informatics, University."

Similar presentations


Ads by Google