pathChirp Efficient Available Bandwidth Estimation

1
pathChirp Efficient Available Bandwidth
Estimation

• Vinay Ribeiro
• Rice University
• Rolf Riedi Jiri Navratil
• Rich Baraniuk Les Cottrell
• (Rice) (SLAC)

2
Network Model

• End-to-end paths
• Multi-hop
• No packet reordering
• Router queues
• FIFO
• Constant service rate

• Packet delay constant term
• (propagation,
• service time)
• variable term
• (queuing delay)

3
Available Bandwidth

• Unused capacity along path

Available bandwidth

• Goal use end-to-end probing to estimate
  available bandwidth

4
Applications

• Server selection

• Route selection (e.g. BGP)

• Network monitoring

• SLA verification

• Congestion control

5
Available Bandwidth Probing Tool Requirements

• Fast estimate within few RTTs
• Unobtrusive introduce light probing load
• Accurate

• No topology information (e.g. link speeds)
• Robust to multiple congested links

• No topology information (e.g. link speeds)
• Robust to multiple congested links

6
Principle of Self-Induced Congestion
Probing rate lt available bw ? no delay
increase Probing rate gt available bw ? delay
increases

• Advantages
• No topology information required
• Robust to multiple bottlenecks
• TCP-Vegas uses self-induced congestion principle

7
Trains of Packet-Pairs (TOPP) Melander et al
8
Pathload Jain Dovrolis

• CBR packet trains
• Vary rate of successive trains
• Converge to available bandwidth

• Shortcoming
• Efficiency only one data rate per train

9
Chirp Packet Trains

• Exponentially decrease packet spacing within
  packet train
• Wide range of probing rates
• Efficient few packets

10
Chirps vs. Packet-Pairs

• Each chirp train of N packets contains N-1 packet
  pairs at different spacings
• Reduces load by 50
• Chirps N-1 packet spacings, N packets
• Packet-pairs N-1 packet spacings, 2N-2 packets
• Captures temporal queuing behavior

11
Chirps vs. CBR Trains

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

12
CBR Cross-Traffic Scenario

• Point of onset of increase in queuing delay gives
  available bandwidth

13
Bursty Cross-Traffic Scenario

• Goal exploit information in queuing delay
  signature

14
PathChirp Methodology

• Per-packet pair available bandwidth, (kpacket
  number)
• Per-chirp available bandwidth
• Smooth per-chirp estimate over sliding time
  window of size

15
Self-Induced Congestion Heuristic

• Definitions delay of packet k
• inst rate at packet k

16
Excursions

• Must take care while using self-induced
  congestion principle
• Segment signature into excursions from x-axis
• Valid excursions are those consisting of at least
  L packets
• Apply only to valid excursions

17
Setting Per-Packet Pair Available Bandwidth
18
pathChirp Tool

• UDP probe packets
• No clock synchronization required, only uses
  relative queuing delay within a chirp duration
• Computation at receiver
• Context switching detection
• User specified average probing rate
• open source distribution at spin.rice.edu

19
Performance with Varying Parameters

• Vary probe size, spread factor
• Probing load const.
• Mean squared error (MSE) of estimates

Result MSE decreases with increasing probe size,
decreasing spread factor
20
Multi-hop Experiments

• First queue is bottleneck
• Compare
• No cross-traffic at queue 2
• With cross-traffic at queue 2
• Result MSE close in both scenarios

21
Internet Experiments

• 3 common hops between SLAC?Rice and Chicago?Rice
  paths

• Estimates fall in proportion to introduced
  Poisson traffic

22
Comparison with TOPP

• Equal avg. probing rates for pathChirp and TOPP
• Result pathChirp outperforms TOPP

• 30 utilization

70 utilization
23
Comparison with Pathload

• 100Mbps links
• pathChirp uses 10 times fewer bytes for
  comparable accuracy

Available bandwidth Efficiency Efficiency Accuracy Accuracy
Available bandwidth pathchirp pathload pathChirp 10-90 pathload Avg.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
24
Conclusions

• Chirp trains
• Probe at multiple rates simultaneously
• Efficient estimates
• pathChirp
• Self-induced congestion
• Excursion detection
• Experiments
• Internet experiments promising
• Large probe packet size, small spread factor
  better
• Outperforms existing tools
• open-source code is available at
• spin.rice.edu
• Demo during 1030a.m. break