pathChirp Spatio-Temporal Available Bandwidth Estimation

1. pathChirpSpatio-Temporal Available Bandwidth Estimation Vinay Ribeiro Rolf Riedi, Richard Baraniuk Rice University

2. A Bird’s Eye View of the Internet • Data transmitted as packets • Multiple routers on end-to-end path • Routers queue bursts of packets

3. Edge-based Probing • Internet provides connectivity • Lack of optimization • Difficult to obtain information from routers Solution: inject probe packets to measure internal properties

4. Questions awaiting Answers • What does the Internet topology look like? • Where does congestion occur and why? • Given several mirror sites to download data which one to choose? • Is my ISP honoring the service-level agreement?

5. Available Bandwidth Estimation

6. Network Model Packet delay = constant term (propagation, service time) + variable term (queuing delay) • End-to-end paths • Multi-hop • No packet reordering • Router queues • FIFO • Constant service rate

7. Available Bandwidth • Unused capacity along path Available bandwidth: • Goal: use end-to-end probing to estimate available bandwidth

8. Applications • Server selection • Route selection (e.g. BGP) • Network monitoring • SLA verification • Congestion control

9. Available Bandwidth Probing Tool Requirements • Fast estimate within few RTTs • Unobtrusive introduce light probing load • Accurate • No topology information(e.g. link speeds) • Robustto multiple congested links • No topology information(e.g. link speeds) • Robustto multiple congested links

10. Principle of Self-Induced Congestion • Advantages • No topology information required • Robust to multiple bottlenecks • TCP-Vegas uses self-induced congestion principle Probing rate < available bw  no delay increase Probing rate > available bw  delay increases

11. Vary sender packet-pair spacing • Compute avg. receiver packet-pair spacing • Constrained regression based estimate • Shortcoming: packet-pairs • do not capture temporal • queuing behavior useful for • available bandwidth • estimation Packet-pairs Packet train Trains of Packet-Pairs (TOPP)[Melander et al]

12. Pathload [Jain & Dovrolis] • CBR packet trains • Vary rate of successive trains • Converge to available bandwidth • Shortcoming • Efficiency: only one data rate per train

13. Chirp Packet Trains • Exponentiallydecrease packet spacing withinpacket train • Wide range of probing rates • Efficient:few packets

14. 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

15. Chirps vs. CBR Trains • Multiple rates in each chirping train • Allows one estimate per-chirp • Potentially more efficient estimation

16. CBR Cross-Traffic Scenario • Point of onset of increase in queuing delay gives available bandwidth

17. Bursty Cross-Traffic Scenario • Goal: exploit information in queuing delay signature

18. PathChirp Methodology • Per-packet pair available bandwidth, (k=packet number) • Per-chirp available bandwidth • Smooth per-chirp estimate over sliding time window of size

19. Self-Induced Congestion Heuristic • Definitions: delay of packet k inst rate at packet k

20. 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 validexcursions

21. Valid excursion increasing queuing delay • Valid excursion decreasing queuing delay • Invalid excursions • Last excursion Setting Per-Packet Pair Available Bandwidth

22. pathChirp Tool • UDPprobe 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

23. 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

24. 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

25. Internet Experiments • 3 common hops between SLACRice and ChicagoRice paths • Estimates fall in proportion to introduced Poisson traffic

26. Comparison with TOPP • Equal avg. probing rates for pathChirp and TOPP • Result: pathChirp outperforms TOPP 30% utilization 70% utilization

27. Comparison with Pathload • 100Mbps links • pathChirp uses 10 times fewer bytes for comparable accuracy

28. Tight Link Localization

29. Key Definitions Path available bandwidth Sub-path available bandwidth Tight link: link with least available bandwidth • Goal: use end-to-end probing to locate tight link in space and over time

30. Applications • Science: where do Internet tight links occur and why? • Network aware applications • - server selection • Network monitoring • - locating hot spots

31. Methodology • Estimate A[1,m] • For m>tight link, A[1,m] remains constant

32. Principle of Self-Induced Congestion • Probing rate = R, path available bandwidth = A R < A no delay increase R > A delay increases • Advantages • No topology information required • Robust to multiple bottlenecks

33. Packet Tailgating • Large packets of size P (TTL=m)small packets of size p • Large packets exit at hop m • Small packets reach receiver with timing information • Previously employed in capacity estimation

34. Estimating A[1,m] • Key: Probing rate decreases by p/(p+P) at link m • Assumption:r<A[m+1,N],no delay change after link m R < A[1,m] no delay increase R > A[1,m] delay increases

35. Tight Link Localization • Tight link: link after which A[1,m] remains constant • Applicable to any self-induced congestion tool: pathload, pathChirp, IGI, netest etc.

36. ns-2 Simulation tight link • Heterogeneous sources • Tight link location changes over time • pathChirp tracks tight link location change accurately estimate

37. Internet Experiment SLACRice tight link • Two paths:UIUC Rice and SLACRice • Paths share 4 common links • Same tight link estimate for both paths UIUCRice tight link

38. Comparison with MRTG Data SLACRice UIUCRice • A[1,m] decreases as expected • Tight link location differs from MRTG data by 1 hop

39. High Speed Probing • System I/O limits probing rate • On high speed networks:  cannot estimate A using self-induced congestion

40. Receiver System I/O Limitation • Treat receiver I/O bus as an extra link • Use packet tailgating • If then we can estimate A[1,N-1]

41. Sender System I/O Limitations • Combine sources to increase net probing rate • Issue: machine synchronization

42. Conclusions • Towards spatio-temporal available bandwidth estimation • Combine self-induced congestion and packet tailgating • Available bandwidth and tight link localization in space and over time • ns-2 and Internet experiments encouraging • Solutions to system I/O bandwidth limitations spin.rice.edu