Bandwidth estimation in computer networks:

measurement techniques &applications

Constantine Dovrolis

College of ComputingGeorgia Institute of Technology

The Internet as a “black box”

Several network properties are important for applications and transport protocols: Delay, loss rate, capacity, congestion, load, etc

But routers and switches do not provide direct feedback to end-hosts (except ICMP, also of limited use) Mostly due to scalability, policy, and simplicity reasons

Can we guess what is in the black box?

End-systems can infer network state through end-to-end (e2e) measurements Without any feedback from routers Objectives: accuracy, speed, non-intrusiveness

probe packets

Bandwidth estimation in packet networks

Bandwidth estimation (bwest) area: Inference of various throughput-related metrics

with end-to-end network measurements Early works:

Keshav’s packet pair method for congestion control (‘89) Bolot’s capacity measurements (’93) Carter-Crovella’s bprobe and cprobe tools (’96) Jacobson’s pathchar - per-hop capacity estimation (‘97) Melander’s TOPP method for avail-bw estimation (’00)

In last 3-5 years: Several fundamentally new estimation techniques Many research papers at prestigious conferences More than a dozen of new measurement tools Several applications of bwest methods Significant commercial interest in bwest technology

Overview

This talk is an overview of the most important developments in bwest area over the last 10 years A personal bias could not be avoided..

Overview Bandwidth-related metrics Capacity estimation

Packet pairs and CapProbe technique Available bandwidth estimation

Iterative probing: Pathload, Pathvar and PathChirp Comparison of tools

Applications SOcket Buffer Auto-Sizing (SOBAS)

Network bandwidth metrics

Ravi S.Prasad, Marg Murray, K.C. Claffy, Constantine Dovrolis,

“Bandwidth Estimation: Metrics, Measurement Techniques, and Tools,”

IEEE Network, November/December 2003.

Capacity definition Maximum possible end-to-end throughput at IP layer

In the absence of any cross traffic Achievable with maximum-sized packets

If Ci is capacity of link i, end-to-end capacity C defined as:

Capacity determined by narrow link

iHiCC

,...,1min

Available bandwidth definition Per-hop average avail-bw:

Ai = Ci (1-ui)

ui: average utilization A.k.a. residual capacity

End-to-end avg avail-bw A:

Determined by tight link ISPs measure per-hop avail-

bw passively (router counters, MRTG graphs)

iHiAA

,...,1min

The avail-bw as a random process

Instantaneous utilization ui(t): either 0 or 1 Link utilization in (t, t+)

Averaging timescale: Available bandwidth in (t, t+)

End-to-end available bandwidth in (t, t+)

t

t

ii dttuttu )(1

),(

)],(1[),( ttuCttA iii

),(min),(,...,1

ttAttA iHi

Available bandwidth distribution Avail-bw has significant variability

Need to estimate second-order moments, or even better, the avail-bw distribution

Variability depends on averaging timescale Larger timescale, lower variance

Distribution is Gaussian-like, if >100-200 msec and with sufficient flow multiplexing

E2E Capacity estimation (a’):

Packet pair technique

Constantine Dovrolis, Parmesh Ramanathan, David Moore,

“Packet Dispersion Techniques and Capacity Estimation,”

In IEEE/ACM Transactions on Networking, Dec 2004.

Packet pair dispersion Packet Pair (P-P) technique

Originally, due to Jacobson & Keshav

Send two equal-sized packets back-to-back Packet size: L Packet trx time at link i:

L/Ci

P-P dispersion: time interval between last bit of two packets

Without any cross traffic, the dispersion at receiver is determined by narrow link:

iinout C

L,max

C

L

C

L

iHi

R

,...,1max

Cross traffic interference Cross traffic packets can affect P-P dispersion

P-P expansion: capacity underestimation P-P compression: capacity overestimation

Noise in P-P distribution depends on cross traffic load Example: Internet path with 1Mbps capacity

Multimodal packet pair distribution Typically, P-P distribution includes several local modes

One of these modes (not always the strongest) is located at L/C Sub-Capacity Dispersion Range (SCDR) modes:

P-P expansion due to common cross traffic packet sizes (e.g., 40B, 1500B)

Post-Narrow Capacity Modes (PNCMs): P-P compression at links that follow narrow link

E2E Capacity estimation (b’):

CapProbe

Rohit Kapoor, Ling-Jyh Chen, Li Lao, Mario Gerla, M. Y. Sanadidi,

"CapProbe: A Simple and Accurate Capacity Estimation Technique,"

ACM SIGCOMM 2004

Compression of p-p dispersion First packet queueing => compressed dispersion

Capacity overestimation

Expansion of p-p dispersion Second packet queueing => expansion of dispersion

Capacity underestimation

CapProbe Both expansion and compression are result queuing delays Key insight: packet pair that sees zero queueing delay would yield exact estimate

measure RTT for each probing packet of packet pair estimate capacity from pair with minimum RTT-sum

CapacityCapacity

Measurements - Internet, Internet2

To

UCLA-2 UCLA-3 UA NTNU

Time Capacity Time Capacity Time Capacity Time Capacity

CapProbe

0’03 5.5 0’01 96 0’02 98 0’07 97

0’03 5.6 0’01 97 0’04 79 0’07 97

0’03 5.5 0’02 97 0’17 83 0’22 97

Pathrate

6’10 5.6 0’16 98 5’19 86 0’29 97

6’14 5.4 0’16 98 5’20 88 0’25 97

6’5 5.7 0’16 98 5’18 133 0’25 97

Pathchar

21’12 4.0 22’49 18 3 hr 34 3 hr 34

20’43 3.9 27’41 18 3 hr 34 3 hr 35

21.18 4.0 29’47 18 3 hr 30 3 hr 35

• CapProbe implemented using PING packets, sent in pairs

Avail-bw estimation (a’):

Pathload

Manish Jain, Constantine Dovrolis,“End-to-End Available Bandwidth:

Measurement Methodology, Dynamics, and Relation with TCP Throughput,”

IEEE/ACM Transactions on Networking, Aug 2003.

Probing methodology

Sender transmits periodic packet stream of rate R K packets, packet size L, interarrival T = L/R

Receiver measures One-Way Delay (OWD) for each packet D(k) = tarv(k) - tsnd(k) OWD variations: Δ(k) = D(k+1) – D(k)

Independent of clock offset between sender/receiver

With stationary & fluid-modeled cross traffic: If R > A, then Δ(k) > 0 for all k Else, Δ(k) = 0 for all k

Self-loading periodic streams

Increasing OWDs means R>A Almost constant OWDs means R<A

Example of OWD variations 12-hop path from U-Delaware to U-Oregon

K=100 packets, A=74Mbps, T=100μsec Rleft = 97Mbps, Rright=34Mbps

Ri

Iterative probing and Pathload

Iterative probing in Pathload: Send multiple probing streams (duration τ) at various rates Each stream samples path at different time interval Outcome of each stream is either Ri > A or Ri < A Estimate upper and lower bound for avail-bw variation range

Avail-bw time series from NLANR trace

Avail-bw estimation (b’): percentile estimation

Manish Jain, Constantine Dovrolis,“End-to-End Estimation of the Available

Bandwidth Variation Range,”ACM SIGMETRICS 2005

Avail-bw distribution and percentiles

Avail-bw random process, in timescale A(t) Assume stationarity

Marginal distribution of A:

F(R) = Prob [A ≤ R]

Ap :pth percentile of A, such that p = F(Ap)

Objective: Estimate variation range [AL, AH] for given averaging timescale ALand AH are pL and pH percentiles of A

Typically, pL =0.10 and pH =0.90

Percentile sampling

Question : Which percentile of A corresponds to rate R?

Given R and estimate F(R)

Assume that F(R) is inversible Sender transmits periodic packet stream of rate R

Length of stream = measurement timescale Receiver classifies stream as

Type-G if A ≤ R: I(R)= 1 with probability F(R)

Type-L otherwise: I(R)= 0 with probability 1-F(R) Collect N samples of I(R) by sending N streams of rate R

Number of type-G streams: I(R,N)= Σi Ii(R)

E[ I(R,N) ] = F(R) * N Percentile rank of probing rate R: F(R) = E[I(R,N)] / N

Use I(R,N)/N as estimator of F(R)

Non-parametric estimation

Does not assume specific avail-bw distribution Iterative algorithm

Stationarity requirement across iterations N-th iteration: probing rate Rn

Use percentile sampling to estimate percentile rank of Rn

To estimate the upper percentile AH with pH = F(AH): fn = I(Rn,N)/N If fn is between pH±report AH = Rn

Otherwise, If fn > pH + , set Rn+1 < Rn

If fn < pH - , set Rn+1 > Rn

Similarly, estimate the lower percentile AL

Validation example Verification using real Internet traffic traces

b=0.05 b=0.15

Non-parametric estimator tracks variation range within 10-20% Selection of b depends on traffic

Traffic spikes/dips may not be detected if b is too small But larger b causes larger MSRE

Parametric estimation Assume Gaussian avail-bw distribution

Justified assumption for large degree of traffic multiplexing And/or for long averaging timescale (>200msec)

Gaussian distribution completely specified by Mean and standard deviation

pth percentile of Gaussian distribution Ap = + -1(p)

Sender transmits N probing streams of rates R1 and R2 Receiver determines percentiles ranks corresponding to R1 and R2 and can be then estimated by solving

R1 = + -1(p1) R2 = + -1(p2)

Variation range is then calculated from: AH = + -1(pH) AL = + -1(pL)

Validation example

Gaussian traffic non-Gaussian traffic

Verification using real Internet traffic traces Evaluated with both Gaussian and non-Gaussian traffic

Parametric algorithm is more accurate than non-parametric algorithm, when

traffic is good match to Gaussian model in non-stationary conditions

Avail-bw estimation (c’):

PathChirp

Vinay Ribeiro, Rolf Riedi, Jiri Navratil, Rich Baraniuk, Les Cottrell,

“PathChirp: Efficient Available Bandwidth Estimation,”PAM 2003

Chirp Packet Trains

Exponentially decrease packet spacing within packet train

Wide range of probing rates

Efficient: few packets 100Mbps-1 packets, 134.1

Chirps vs. CBR Trains

Multiple rates in each chirping train Allows one estimate per-chirp Potentially more efficient estimation

CBR Cross-Traffic Scenario

Point of onset of increase in queuing delay gives available bandwidth

Comparison with Pathload

• 100Mbps links• pathChirp uses 10

times fewer bytes for comparable accuracy

Available bandwidth

Efficiency Accuracy

pathchirp pathload pathChirp10-90%

pathloadAvg.min-max

30Mbps 0.35MB 3.9MB 19-29Mbps 16-31Mbps

50Mbps 0.75MB 5.6MB 39-48Mbps 39-52Mbps

70Mbps 0.6MB 8.6MB 54-63Mbps 63-74Mbps

Experimental comparison of avail-bw estimation tools

Alok Shriram, Marg Murray, Young Hyun, Nevil Brownlee, Andre Broido, k claffy,

”Comparison of Public End-to-End Bandwidth Estimation Tools on High-Speed

Links,”PAM 2005

Testbed topology

Accuracy comparison - SmartBitsDirection 1, Measured AB

Direction 2, Measured AB

Actual AB

Accuracy comparison - TCPreplayActual Available Bandwidth

Measured Available Bandwidth

What’s wrong with Spruce? Spruce was proposed as fast & accurate estimation

method Strauss et al., IMC’03: example of direct probing

With a single-link model: Sender sends periodic stream with input rate Ri Receiver measures output rate Ro Avail-bw A given by:

Assumptions: Tight link capacity Ct is known Input rate Ri can be controlled by source But what would happen in a multi-hop path?

Ri can be less than source rate (due to previous link with lower capacity) and, even worse, it can be unknown!

)1( o

tit R

CRCA

Measurement latency comparison

• Abing: 1.3 to 1.4 s

• Spruce: 10.9 to 11.2 s

• Pathload: 7.2 to 22.3 s

• Patchchirp: 5.4 s

• Iperf: 10.0 to 10.2 s

Probing traffic comparison

Applications of bandwidth estimation

Applications of bandwidth estimation

Large TCP transfers and congestion control Bandwidth-delay product estimation Socket buffer sizing

Streaming multimedia Adjust encoding rate based on avail-bw

Intelligent routing systems Overlay networks and multihoming Select best path based on capacity or avail-bw

Content Distribution Networks (CDNs) Choose server based on least-loaded path

SLA and QoS verification Monitor path load and allocated capacity

End-to-end admission control Network “spectroscopy” Several more..

Application-1:Improved TCP

Throughput with BwEst

Ravi S. Prasad, Manish Jain, Constantine Dovrolis,

“Socket Buffer Auto-Sizing for High-Performance Data Transfers,”

Journal of Grid Computing, 2003

TCP performs poorly in fat-long paths Basic idea in TCP congestion control:

Increase congestion window until packet loss occurs Then reduce window by a factor of two (multiplicative decrease)

Consequences Increased loss-rate Avg. throughput less than

available bandwidth Large delay-variation

Example: 1Gbps path, 100msec RTT, 1500B pkts Single loss ≈ 4000 pkts

window reduction Recovery from single loss

≈ 400 sec Need very small loss

probability (ρ<2*10-8) to saturate path

TCP background

Socket-layer buffers Send buffer: Bs Receive buffer: Br

Socket BufferSender Receiver

Socket Layer

TCP Layer WrWs

TCP windows Send: Ws <= Bs Congestion: Wc Receiver (advertised): Wr<=

Br Ws = min {Bs, Wc, Wr}

Ideally, TCP send window Ws should be set to connection’s bandwidth-delay product “Bandwidth”: connection’s fair share “Delay”: connection’s RTT Achieve efficiency, fairness, no congestion

Can we avoid congestive losses w/o changing TCP?

Ws = min {Wc, Wr} But Wr<= S

S: rcv socket buffer size We can limit S so that connectionis limited by receive window: Ws = S Non-congested path:

Throughput R(S) = S/T(S) R(S) increases with S until it reaches MFT MFT: Maximum Feasible Throughput (onset of congestion) Objective: set S close to MFT, to avoid any losses and

stabilize TCP send window to a safe value Congested path:

Path with losses before start of target transfer Limiting S can only reduce throughput

Socket buffer auto-sizing (SOBAS) Apply SOBAS only in non-congested paths Receiving application measures rcv-throughput Rrcv

When receive throughput becomes constant: Connection has reached its maximum throughput Rmax

If the send window becomes any larger, losses will occur Receiving application limits rcv socket buffer size S:

S = RTT * Rmax

With congestion unresponsive cross traffic: Rmax = A With congestion responsive cross traffic: Rmax = MFT

SOBAS does not require any changes in TCP But estimation of RTT and congestion-status requires

“ping-like” probing

Measurements at WAN path

Path: GaTech to LBL,CA SOBAS flow get higher average throughput than non-

SOBAS flow because former avoids all congestion losses SOBAS does not significantly increase path’s RTT

Non-SOBAS SOBAS

To conclude..

Things I have not talked about Other estimation tools & techniques

Abing, netest, pipechar, STAB, pathneck, IGI/PTR, … Per-hop capacity estimation (pathchar, clink, pchar)

Other applications Bwest in new TCPs (e.g., TCP Westwood, TCP FAST) Bwest in wireless nets Bwest in overlay routing Bwest in multihoming Bwest in bottleneck detection

Scalable bandwidth measurements in networks with many monitoring nodes

Practical issues Interrupt coalescence Traffic shapers Non-FIFO queues

Conclusions

We know how to do the following: Estimate e2e capacity & avail-bw Apply bwest in several applications

We know that we cannot do the following: Estimate per-hop capacity with VPS techniques Avoid avail-bw estimation bias when traffic is

bursty and/or path has multiple bottlenecks We do not yet know how to do the following:

Scalable bandwidth measurements Integrate probing traffic in application data

Many research questions are still open Help wanted!