One More Bit Is Enough One More Bit Is Enough Yong Xia, RPI - - PowerPoint PPT Presentation

▶
one more bit is enough one more bit is enough
SMART_READER_LITE
LIVE PREVIEW

One More Bit Is Enough One More Bit Is Enough Yong Xia, RPI - - PowerPoint PPT Presentation

Nov 23, 2023 202 likes •548 views

One More Bit Is Enough One More Bit Is Enough Yong Xia, RPI Lakshmi Subramanian, UCB Ion Stoica, UCB Shiv Kalyanaraman, RPI SIGCOMM 05 , Philadelphia, PA 08 / 23 / 2005 Motivation #1: TCPdoesnt work well in high b/w or delay

slide-1
SLIDE 1

One More Bit Is Enough One More Bit Is Enough

Yong Xia, RPI Lakshmi Subramanian, UCB Ion Stoica, UCB Shiv Kalyanaraman, RPI

SIGCOMM’05, Philadelphia, PA

08 / 23 / 2005

slide-2
SLIDE 2

Motivation #1: TCPdoesn’t work well in high b/w or delay

bandwidth (Mbps) bottleneck utilization TCP 70% 100% bottleneck utilization round-trip delay (ms) TCP 50% 100%

VCP 2

slide-3
SLIDE 3

Motivation #1: Why TCP does not scale?

  • TCP uses binary congestion signals, such as loss or one-bit

Explicit Congestion Notification (ECN)

time congestion window

Multiplicative Decrease (MD) Additive Increase (AI) slow!

  • AI with a fixed step-size can be very slow for large bandwidth

VCP 3

slide-4
SLIDE 4

Motivation #2: XCP scales

  • XCP decouples efficiency control and fairness control

x sender receiver router

C

spare bandwidth

  • - MIMD

DATA

  • - AIMD

rate & rtt

ACK

∆rate

  • But, XCP needs multiple bits (128 bits in its current IETF draft) to

carry the congestion-related information from/to network

VCP 4

slide-5
SLIDE 5

Goal Design a TCP-like scheme that: requires a small amount of congestion information (e.g., 2 bits) scales across a wide range of network scenarios

VCP 5

slide-6
SLIDE 6

Key Observation Fairness is not critical in low-utilization region Use Multiplicative Increase (MI) for fast convergence onto efficiency in this region Handle fairness in high-utilization region

VCP 6

slide-7
SLIDE 7

Variable-structure congestion Control Protocol (VCP)

  • Routers signal the level of congestion
  • End-hosts adapt the control algorithm accordingly

VCP 7

sender receiver x router

traffic rate link capacity

(11) (10) (01) code load factor region

low-load high-load

  • verload

control

Multiplicative Decrease (MD) Additive Increase (AI) Multiplicative Increase (MI)

range of interest scale-free

ACK

2-bit ECN

1

2-bit ECN

slide-8
SLIDE 8

VCP vs. ECN

  • ECN doesn’t differentiate between low-load and high-load regions

TCP AQM/ECN

sender receiver x router

code load factor region

  • verload

control

Multiplicative Decrease (MD) 1

(1) (0)

underload Additive Increase (AI)

ACK

1-bit ECN

VCP 8

slide-9
SLIDE 9

An illustration example

  • MI tracks available bandwidth exponentially fast

time (sec)

  • After high utilization is attained, AIMD provides fairness

link utilization flow cwnd (pkt) 9 flows join at 100s and leave at 200s

MI AIMD

VCP 9

slide-10
SLIDE 10

VCP vs. ECN

VCP 10

TCP+RED/ECN

cwnd utilization time (sec)

VCP

utilization cwnd time (sec)

slide-11
SLIDE 11

VCP key ideas and properties

  • verload

high-load low-load router fairness control efficiency control MI AIMD end-host

  • Use network link load factor as the congestion signal
  • Decouple efficiencyandfairness controls in different load regions
  • Achieve high efficiency, low loss, and small queue
  • Fairness model is similar to TCP:
  • Long flows get lower bandwidth than in XCP (proportional vs.

max-min fairness)

  • Fairness convergence much slower than XCP (solvable with

even more, e.g., 8 bits)

VCP 11

slide-12
SLIDE 12

Major design issues

control

Multiplicative Decrease (MD) Additive Increase (AI) Multiplicative Increase (MI)

(11) (10) (01) code load region

low-load high-load

  • verload

At the router

  • How to measure and encode the load factor?

At the end-host

  • When to switch from MI to AI?
  • What MI / AI / MD parameters to use?
  • How to handle heterogeneous RTTs?

VCP 12

slide-13
SLIDE 13

Design issue #1: measuring and encoding load factor tρ = 200ms >≈ most rtt

  • Calculate the link load factor ρ

demand load_factor capacity = capacity demand link_bandwidth * tρ arrival_traffic + queue_size =

  • The load factor is quantized and encoded into the two ECN bits

information loss affects fairness

VCP 13

slide-14
SLIDE 14

Design issue #2: setting MI/AI/MD parameters (ξ, α, β)

  • verload

high-load low-load MI AIMD

VCP 14

slide-15
SLIDE 15

Design issue #2: setting MI/AI/MD parameters (ξ, α, β) Q: load factor transition point ρ* for MI AI?

load factor 100% 80%

safety margin

0% 0.875 TCP: ξ = 1.0 VCP: ξ = 0.06 STCP: ξ = 0.01 MI AI MD k (1 − ρ*) / ρ*

where k = 0.25 (for stability)

1.0 α = β = ξ = ρ* =

VCP 15

slide-16
SLIDE 16

Design issue #3: Handling RTT heterogeneity for MI/AI

  • Scale ξ to prevent MI from overshooting capacity when RTT is small

rtt < tρ

  • vershoot

ξ s

rtt = tρ

ξ

tρ tρ

time spare capacity

1

VCP 16

slide-17
SLIDE 17

VCP scales across b/w, rtt, num flows

  • Evaluation using extensive ns2 simulations
  • 150Mbps, 80ms, 50 forward flows and 50 reverse flows

VCP 17

slide-18
SLIDE 18

VCP achieves high efficiency

bottleneck utilization

TCP VCP XCP

bandwidth (Mbps) bottleneck utilization

TCP VCP XCP

round-trip delay (ms)

VCP 18

slide-19
SLIDE 19

VCP minimizes packet loss rate

TCP VCP XCP TCP VCP XCP bottleneck utilization bottleneck utilization TCP VCP XCP

round-trip delay (ms)

packet loss rate TCP VCP XCP

bandwidth (Mbps)

packet loss rate

VCP 19

slide-20
SLIDE 20

VCP comparisons

Compared to TCP+AQM/ECN

Same architecture (end-hosts control, routers signal) Router congestion detection: queue-based load-based Router congestion signaling: 1-bit 2-bit ECN End-host adapts (MI/AI/MD) according to the ECN feedback End-host scales its MI/AI parameters with its RTT

Compared to XCP

Decouple efficiency/fairness control across load regions Functionality primarily placed at end-hosts, not in routers

VCP 20

slide-21
SLIDE 21

Theoretical results Assumptions:

One bottleneck of infinite buffer space is shared by synchronous

flows that have identical RTTs;

The exact value of load factor is echoed back.

Theorem for the VCP fluid model:

It is globally stable with a unique and fair equilibrium, if k ≤ 0.5; The equilibrium is max-min fair for general topologies; The equilibrium is optimal by achieving all the design goals.

VCP protocol differs from the model in fairness.

VCP 21

slide-22
SLIDE 22

Conclusions With a few minor changes over TCP+AQM/ECN, VCP is able to approximate the performance of XCP

High efficiency Low persistent bottleneck queue Negligible congestion-caused packet loss Reasonable (i.e., TCP-like) fairness

VCP 22

slide-23
SLIDE 23

Future work How do we get there, incrementally?

End-to-end VCP TCP-friendliness Incentive

Extensions

Applications: short-lived data traffic, real-time traffic Environment: wireless channel

Security

Robust signaling, e.g., ECN nonce

VCP 23

slide-24
SLIDE 24

The end

Thanks!

VCP 24

slide-25
SLIDE 25

Design issue #3: Handling RTT heterogeneity for MI/AI

  • TCP throughput is biased against flows with large RTT

packet_size rate rtt ⋅ sqrt( loss_rate ) ≈

  • VCP scales α for fair rate sharing (regardless of RTT)

rate = cwnd / rtt

VCP 25

slide-26
SLIDE 26

VCP keeps small bottleneck queue

queue length in % buffer size

TCP VCP XCP

bandwidth (Mbps) queue length in % buffer size

TCP VCP XCP

round-trip delay (ms)

per-flow b/w-delay- product ≤ 1 packet

VCP 26

slide-27
SLIDE 27

Vary the number of flows

number of flows

VCP TCP bottleneck utilization queue length in % buffer size XCP VCP TCP XCP

VCP 27

slide-28
SLIDE 28

Influence of RTT on fairness To some extent, VCP distributes reasonably fairly

VCP 28

slide-29
SLIDE 29

Influence of RTT on fairness (cont’d)

VCP 29

slide-30
SLIDE 30

VCP converges onto fairness

VCP 30

slide-31
SLIDE 31

VCP converges onto fairness faster with 8 bits

VCP 31

slide-32
SLIDE 32

Responsiveness

VCP

50 flows + 150 flows – 150 flows

32