1 / 19

Seed Scheduling for Peer-to-Peer Networks

Seed Scheduling for Peer-to-Peer Networks. Flavio Esposito Ibrahim Matta Pietro Michiardi Nobuyuki Mitsutake Damiano Carra. Seed. Seed. [-,-]. [-,-]. Seed. Seed. Seed. Seed. 2. LRF. LRF. [-,-]. INTERESTED. UNCHOKE. REQUEST. PIECE. HAVE(2). INTERESTED.

arlo
Download Presentation

Seed Scheduling for Peer-to-Peer Networks

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Seed Scheduling for Peer-to-Peer Networks Flavio Esposito Ibrahim Matta Pietro Michiardi Nobuyuki Mitsutake Damiano Carra

  2. Seed Seed [-,-] [-,-] Seed Seed Seed Seed 2 LRF LRF [-,-] INTERESTED UNCHOKE REQUEST PIECE HAVE(2) INTERESTED • The Algorithms Peer Selection Piece Selection Seed Round Robin FCFS Leecher Local Rarest First (LRF) Tit-for-Tat + Optimistic BitTorrent Protocol • The Messages of the Protocol

  3. Outline Problem and Motivation Seed Scheduling Our file sharing client BUtorrent Analytical results Experimental Results Conclusions

  4. A = [1,2,-,-,-,-] B = [4,5,-,-,-,-] C = [3,6,-,-,-,-] 2 6 5 Pieces uploaded [1,2,3,4,5,6] Second Scheduling round A = [1,-,-,-,-,-] B = [4,-,-,-,-,-] C = [3,-,-,-,-,-] 3 1 4 A = [1,-,-,-,-,-] B = [4,-,-,-,-,-]C = [3,-,-,-,-,-] Already uploaded in previous rounds Collision at this round Pieces uploaded [1,-,3,4,-,-] 1 First Scheduling round 3 A = [1,2,-,-,-,-] B = [4,2,-,-,-,-]C = [3,4,-,-,-,-] 4 2 4 2 Pieces uploaded [1,-,3,4,-,-] First Scheduling round Pieces uploaded [1,2,3,4,-,-] Second Scheduling round Problems and Motivation Def (Initial Phase): time between first scheduling to instant all pieces have been uploaded once. • Example of Sub-optimal Seed Scheduling: Example of Optimal Seed Scheduling:

  5. Theorem (Bound on Collision):Expected collision with L leechers in the first p rounds is: • Proof on the Tech Report based on: • Bonferroni inequality • And on the following inequality: Do we really have “Collisions”? - Leecher unaware of each other - LRF is not modeled Collisions in initial phase:

  6. leechers disconnected (peers) • slow HAVE messages propagation (pieces) Round 2 Round 3 Round 1 Round 4 Seed Seed Seed Seed Seed Seed 2 1 1 1 2 2 [1,2] [1,2] [1,-] [1,2] [1,-] [1,-] [-,-] [-,-] [1,-] [1,-] [1,2] [-,2] Round 2 Round 1 Round 3 Seed 1 2 Seed [1,2] [1,2] Problem and Motivation Main problem : Unawareness

  7. Waste uploads during Initial Phases • Increase in downloading time Leechers unaware (topology and propagation) of each other waste seed’s upload capacity 4. Burst Arrivals produce same effect of aninitial phase A = [1,2,3,4,5,6] B = [4,2,1,3,6,5]C = [3,4,2,5,1,6] 2 4 2 Problem and Motivation Summarizing:

  8. Outline Problems and Motivation Seed Scheduling Analytical results Experimental Results Conclusions

  9. Upload the pieces i* that are • - Requested the most in this round (max ri (t) ) • - Uploaded the least in the past rounds (min ti (t-1) ) Seed Scheduling • Seeds need to help LRF considering the past and the current requests Algorithm IDEA based on Proportional Fair Scheduling (PFS) Unchoke every possible leecher Details of the algorithm in the paper 2.Collect all possible requests

  10. Seed forget previous scheduling decisions with: Exponential Weighted Moving Average where: • After N rounds seed needs to completely forget • N = max seed’s TCP connections (80-200) • If a seed gets the (N+1)th request for piece i, • It must be a new leecher asking Seed Scheduling History should be forgotten

  11. Outline Problems and Motivation Seed Scheduling Analytical results Experimental Results Conclusions

  12. Here LRF and PFS are modeled • We show that the distribution follows a Zipf Analytical Results • Correct effectiveness definition in initial phase(s) We consider Qiu-Srikant model for seed and leechers evolution

  13. Analytical Results • Take Home messages: • Even SMALL improvements of effectiveness translates in faster downloading time 2. Higher effectiveness reduces downloading time • 3.The improvement is higher • for higher arrival rate • (peerset more dynamic) Solving Qiu-Srikant differential equations numerically with new effectiveness:

  14. Outline Problems and Motivation Seed Scheduling Analytical results Experimental Results Simulations PlanetLab Conclusions

  15. Simulation Results Take home 1. If Burstiness B too low for  = 0.02 forgetting too fast has same effect as not remembering (as BitTorrent) 2. If B too high, peersets are too small for effective piece exchange 3. Under PFS up to 25% improvement over BitTorrent. We compare average downloading time ofPFS withFCFSandGRF Burstiness *B := peak rate /average rate of arrival of leechers

  16. Simulation Results Take home • Even for less dynamic peerset, download time does not improve much, Seed utilization improves under PFS • 2. Improvement increases for • bigger files. Seed Utilization* vs File Size *Seed Utilization := # seed uploads / avg # leechers uploads

  17. BUtorrent on PlanetLab Average downloading time of BUtorrent vs BitTorrent • We implemented a new file-sharing client with our seed scheduling algorithm (BUtorrent) • Simulation result confirmed on PlanetLab Take home

  18. Conclusions Showed the need for smarter scheduling to reduce initial phases and so Downloading Time Our idea: to give a more global view to seeds “Supporting and not Substituting LRF” New seed scheduling algorithm implemented in a real client Boston University BUtorrent BUtorrent improves BitTorrent’s average downloading time by up to 25%.

  19. Thank you !!! Download BUtorrent from http://csr.bu.edu/butorrent

More Related