Implementing High Speed TCP (aka Sally Floyd’s) Yee-Ting Li & Gareth Fairey 1 st October 2002 DataTAG CERN (Kinda!)

What is High Speed TCP? Changes the way TCP behaves at high speed (ie large cwnd) Changes the way TCP behaves at high speed (ie large cwnd) Standard TCP has two modes Standard TCP has two modes –Slow start (not very slow…) –Congestion Avoidance Focuses on Congestion Avoidance Mode Focuses on Congestion Avoidance Mode –ie when TCP knows (thinks it knows…) how well the network behaves… BUT only when we are at high speeds, else do what normal Standard TCP does… BUT only when we are at high speeds, else do what normal Standard TCP does… Readily deployable 1 st step towards Equation Based Congestion Control Readily deployable 1 st step towards Equation Based Congestion Control

What does it do? TCP uses two parameters during Congestion Avoidance - AIMD( a,b ) TCP uses two parameters during Congestion Avoidance - AIMD( a,b ) –Increase parameter, a –Decrease parameter, b StandardTCP uses StandardTCP uses –a=1 –b=0.5 High Speed TCP introduces a dependance of a and b depending on the current cwnd High Speed TCP introduces a dependance of a and b depending on the current cwnd –a->a(cwnd) –b->b(cwnd) If we increase a more with larger cwnd we can get back up to our ‘optimal’ cwnd size for the network path If we increase a more with larger cwnd we can get back up to our ‘optimal’ cwnd size for the network path If we decrease b less we don’t lose as much bandwidth due to a small congestion window If we decrease b less we don’t lose as much bandwidth due to a small congestion window

What exactly does it do? Based on the TCP response function Based on the TCP response function –Relates loss and throughput Uses the TCP response function to investigate certain parameters Uses the TCP response function to investigate certain parameters – High_Window, High_Loss ; largest cwnd needed for x throughput and the required loss for that throughput – Low_Window, Low_Loss ; smallest cwnd when we actually switch from Standard TCP and the required loss rate for that cwnd size – High_B ; the smallest decrease in b when we are at a large cwnd Equations to transform this information into a table for a(cwnd) and b(cwnd) Equations to transform this information into a table for a(cwnd) and b(cwnd)

a(cwnd) & b(cwnd)

Kernel Modifications Focus on investigating phase space of HSTCP variables Focus on investigating phase space of HSTCP variables 3 Phase approach 3 Phase approach –Hard code HSTCP as compile-time option – ‘recommended’ HSTCP by Sally Floyd –Kernel hooks to enable alteration of HSTCP variables – from user space Allows easier testing/exploration Allows easier testing/exploration Only a few changes necessary: Only a few changes necessary: –Code for calculating the a(cwnd) and b(cwnd) values –Modification of existing code for changing cwnd (during the Congestion Avoidance phase only) Implementation on kernel with web100 (currently alpha1.1) Implementation on kernel with web100 (currently alpha1.1)

a(cwnd) & b(cwnd) a(cwnd) and b(cwnd) implement via a static look-up table a(cwnd) and b(cwnd) implement via a static look-up table Using access function get_hstcp_val Using access function get_hstcp_val Changes to congestion avoidance algorithm required only slight modification in tcp_output.c Changes to congestion avoidance algorithm required only slight modification in tcp_output.c

UCL->CERN 100mbit link from UCL->CERN 100mbit link from UCL->CERN Saw tooth increase is greater! Saw tooth increase is greater! Ie a is larger Ie a is larger Limited by 100mbit link… (bandwidth delay ~ 180kbytes) Limited by 100mbit link… (bandwidth delay ~ 180kbytes)

UCL->CERN RTT: 20ms RTT: 20ms Not much difference! Not much difference! No clear advantage to using HSTCP for this link at this speed (100mbit)! No clear advantage to using HSTCP for this link at this speed (100mbit)!

UCL->SLAC Graph shows min and max of about 8 separate flows Graph shows min and max of about 8 separate flows RTT: 135ms RTT: 135ms Standard TCP performance poor Standard TCP performance poor HighSpeedTCP gives performance boost of about 2.5! HighSpeedTCP gives performance boost of about 2.5!

UCL->SLAC Cwnd grows a lot bigger Cwnd grows a lot bigger HSTCP is not linear cwnd growth HSTCP is not linear cwnd growth Grows faster as result (increased bandwidth!) Grows faster as result (increased bandwidth!) AIMD(7,0.34) AIMD(8,0.33) AIMD(9,0.32) AIMD(6,0.35) AIMD(5,0.36) AIMD(4,0.38) AIMD(3,0.41) AIMD(2,0.44) AIMD(1,0.50)

Performance Issues Slow Start Slow Start –High drop rates when we send too much out during Slow Start –Means that we have to linearly increment cwnd –As HSTCP has a much steeper increase, we get up to a large cwnd faster

Controlling Loss Rates Difficult to control what is happening on real life networks Difficult to control what is happening on real life networks Need to be able to control packet drops to quantify effects of TCP Need to be able to control packet drops to quantify effects of TCP Need facility for Sender side to ‘unacknowledged’ acknowledgements Need facility for Sender side to ‘unacknowledged’ acknowledgements –Artificial packet drops Easier to test as only requires Sender side (already modified with HSTCP) to be changed Easier to test as only requires Sender side (already modified with HSTCP) to be changed Investigate into Investigate into –Constant drop rates – unrealistic but easy to implement –Statistically predictable drop rates Can analyse implementation of b parameter easier! Can analyse implementation of b parameter easier!

Test Plan Need to verify a(cwnd) and b(cwnd) are functioning correctly Need to verify a(cwnd) and b(cwnd) are functioning correctly Comparison with Theoretical Results with Network Simulator by Evandro de Berkely Lab Comparison with Theoretical Results with Network Simulator by Evandro de Berkely Lab Investigation of ‘recommended’ HSTCP Investigation of ‘recommended’ HSTCP –Quantify advantages over StandardTCP –Investigate into Fairness Investigation of RTT independence of HSTCP Investigation of RTT independence of HSTCP –Proposed by Tom Kelly & Glenn Vinnicombe Comparison against Multiple Stream TCP Comparison against Multiple Stream TCP –If we can do it with Multiple Stream TCP, why bother with HSTCP? Investigation into Limited Slow Start Investigation into Limited Slow Start –Important if we’re using large cwnds (UCL->SLAC) to better guess the network state at start up

Conclusions Not much use for low latency, low throughput networks Not much use for low latency, low throughput networks Much improved for high latency, high throughput networks Much improved for high latency, high throughput networks Important! Important! –Fairness (and hence internet stability) needs much investigation –Comparisons to existing methods (ie Multistream TCP)