Highspeed TCP

1
High-speed TCP

• FAST TCP motivation, architecture, algorithms,
  performance
• (by Cheng Jin, David X. Wei and Steven H.
  Low)
• Modifying TCP's Congestion Control for High
  Speeds
• (by S. Floyd, S. Ratnasamy, and S. Shenker)
• Scalable TCP Improving Performance in
  High-Speed WAN
• (by Tom Kelly)

2
Problem with TCP

• Sending rate T 1.2 / sqrt(p) packets per
  rtt
• where p is packet loss rate
• Example
• 1500bytes packet, 100ms rtt, 10Gbp pipe
• require window size W 83,333 packets
• at most 1 drop every 5,000,000,000 packets
• at most 1 drop every 6000 seconds
• W sqrt(1.5) / sqrt(p) N W2 / 1.5
• Real drop rate makes TCP a bottleneck which leads
  to poor network utilization

3
Problem with TCP

• AIMD
• ACK w w a / w (each
  rtt)
• Drop w w b w
• Slow StartACK w w c
• where a
  1, b 0.5, c 1
• TCP steady state response function

4
HSTCP Goals

• Performance
• Sustain high speeds without requiring
  unrealistically low loss rates
• Reach high speeds reasonably quickly in when
  slow start
• Recover from congestion without huge delays

5
HSTCP Goals

• Compatibility
• Deployment without router involvement
• Fair treatment of unmodified TCP (unrealistic)
• Fair treatment of unmodified TCP original TCP
  get as much bandwidth as if packet loss rate is
  very small

6
HSTCP Approach

• Leave slow start phase as it is
• Needs only 17 rtt make W 83333 packets
• Change response function by tweaking parameters
  a and b
• Same for p gt P 0.0015 (W 31)
• For smaller p treat a and b as functions of
  current window size

7
HSTCP Response function

• Suggestion of new RF to reach high speed
• w 10S(logp-lopP)logW for S (logW1 logW) /
  (logP1 logP) gives
• w pS (1 / P)S W
• For two points (P, W) (P1, W1)
• P 0.0015, W 31
• P1 10-7, W1 83000
• w 0.15
  / p0.82

8
HSTCP Response function
9
HSTCP Fairness
10
HSTCP Tweaking of a and b

• For w lt W a(w) 1 b(w) 0.5
• For w gt W need such a(w) and b(w) that
• w() gives p(W) P, p(W1) P1

11
HSTCP Testing

• Not available

12
STCP Scalable TCP

• Same goals
• More aggressive increase
• Less aggressive decrease
• Fair treatment of unmodified TCP
• Approach
• ACK W W 0.01 (each ACK)
• Drop W W 0.125 W
• Doubles sending rate in about 70rtt

13
STCP Scaling properties

• In original TCP scaling depends on sending rate
• Sending rate c lt
  Sending rate C

14
STCP Scaling properties

• In Scalable TCP there is no such dependence
• Sending rate c lt
  Sending rate C

15
STCP Response function
For P gt 0.00586 (W15) native TCP function is used
16
STCP Scaling properties
Rate TCP recovery time STCP recovery
time 1Mbps 1.7s
2.7s 10Mbps 17s
2.7s 100Mbps 2mins
2.7s 1Gbps
28mins 2.7s 10Gbps
4hrs 43mins 2.7s
Environment 1500 bytes packet, 200ms rtt
17
STCP Experiments

• Implemented in Linux kernel version 2.14.19
• Competitors TCP, TCPGB modifications, STCP
• Topology and environment

• 2.4Gz Xeon
• 2Gb RAM
• Gigabit Ethernet card
• x 12

18
STCP Experiments

• Experiment 1
• 4 pairs of 2Gb file exchangers

Number of 2Gb transfers completed in 1200 sec
19
STCP Experiments

• Experiment 2
• 3 pairs of Web-Traffic emulators (1400 users
  each)
• 2 pairs of 2Gb file exchangers

Concurrent run of 4200 web users and 8 bulk
transfers within 1200 sec
20
Problems with TCP

• Packet level AIMD provides slow increase and
  drastic decrease
• Flow level Maintaining large congestion windows
  requires small equilibrium loss probability
• Packet level Binary congestion measure leads to
  oscillation
• Flow level Dynamics is unstable. Resulting
  oscillations can be reduced only by accurate
  estimation of packet loss probability and stable
  design of flow dynamics

21
HSTCP and STCP vs TCP Reno

• HSTCP and STC increase more aggressively and
  decrease less drastically so they can tolerate
  larger
• loss probabilities than TCP Reno therefore
  achieve larger equilibrium windows and solve
  problems 1 and 2

22
TCP Oscillations

• Loss based approach
• Full utilization large delays and oscillations

Delay based approach Full utilization
stabilized window, predictable delays and no
oscillations
23
FAST TCP Strategy

• Window adjustment depends on distance from
  equilibrium
• Use queueing delay as congestion measure
• Multi-bit measure eliminates packet level
  oscillations
• Stabilize window near the point where buffer is
  large and delay is small
• Stabilizes flow dynamics since queueing delay
  dynamics scales with respect to network capacity

24
FAST TCP Window adjustment
Window adjustment is independent of where
equilibrium is
25
FAST TCP Design

• Feedback model
• Flow level design such u(wi, Ti) and k(wi, Ti)
  that feedback model above has an equilibrium
  fair, efficient, and stable in presence of
  feedback delay
• Packet level take care of issues ignored by flow
  level such as burstiness control, loss recovery
  and parameter estimation

26
FAST TCP Architecture

• Data control which packets to transmit
• Window control how many
• Burstiness control when
• Estimation provides information to above
  components

27
FAST TCP Window update

• Where
• gamma is in (0, 1
• baseRTT is a minimum RTT observed so far
• qdelay is the average end-to-end queueing delay
• alpha is a constant reflecting of packets each
  flow attempts to maintain in network buffer at
  equilibrium. Provides linear window increase.
  However it can be constant when qdelay is
  nonzero. When qdelay is zero increase is
  exponential
• Window is updated every 2RTT

28
FAST TCP Events and computations

• Acknowledgement
• Qdelay
• Decision about packet injection into the network
• After packet transmission
• Time stamp for each packet
• New window size
• End of RTT
• Target throughput
• Packet loss
• When to retransmit dropped packets

29
FAST TCP Performance

• Testbed and instrumentation
• 2.6 GHz Xeon, 2GB RAM
• Dual onboard gigabit Ethernet interface
• Network bottleneck capacity 800Mbps and 2000pkts
  buffer
• Environment
• Static and 2 types of dynamic

30
FAST TCP Static test
X - of flows, Y - propagation delay, Z
aggregate throughput
31
FAST TCP Dynamic test 1
Throughput and window trajectory
Queue size, packet losses, link utilization
32
FAST TCP Dynamic test 2
Throughput and window trajectory
Queue size, packet losses, link utilization
33
FAST TCP Overall evaluation
Throughput
Fairness
34
FAST TCP Overall evaluation
Responsiveness
Stability