1 / 36

Peer-to-Peer Systems: Challenges and Applications in Collaboration

Explore the world of decentralized, self-organizing peer-to-peer systems with applications like large-scale computation, file sharing, and collaborative work. Discover the potential benefits and scientific challenges in this dynamic network architecture.

rswenson
Télécharger la présentation

Peer-to-Peer Systems: Challenges and Applications in Collaboration

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. In collaboration with: Matthias Bender, Debora Donato, Alessandro Linari, Julia Luxenburger, Sebastian Michel, Nikos Ntarmos, Josiane Parreira, Peter Triantafillou, Christian Zimmer

  2. Peer-to-Peer (P2P) Systems Any applications that are useful and legal ? And scientifically challenging? Decentralized, self-organizing, highly dynamic loose coupling of many autonomous computers • unstructured overlay networks with epidemic dissemination (flooding) • structured overlay networks based on distributed hash tables (DHTs) • Applications: • Large-scale computation (SETI@home, etc.) • File sharing (Napster, Gnutella, KaZaA, BitTorrent, etc.) • Publish-Subscribe (Blogs, Marketplaces, etc.) • Collaborative work (Games, etc.) • IP telephony (Skype) Gerhard Weikum August 3, 2006

  3. Peer-to-Peer Web Search Vision: Self-organizing P2P Web Search Engine with Google-or-better functionality • Scalable & Self-OrganizingData Structures and Algorithms • (DHTs, Semantic Overlay Networks, Epidemic Spreading, Distr. Link Analysis, etc.) • Better Search Result Quality(Precision, Recall, etc.) • Powerful Search Methods for Each Peer • (Concept-based Search, Query Expansion, XML IR, Personalization, etc.) • Leverage User/Community Input („Wisdom of Crowds“) • (Bookmarks, Feedback, Query Logs, Click Streams, Evolving Web, etc.) • Collaboration among Peers • (Query Routing, Incentives, Fairness, Anonymity, etc.) • Benefit of Large-scale Social Networks: • Small-World Phenomenon • Breaking Information Monopolies Gerhard Weikum August 3, 2006

  4. Solution without Problem? > 1 Mio. scholars and graduate students • produce valuable text & scientific data, knowledge, annotations, etc. • leverage NLP, statistical learning, data curation & enrichment, etc. • would benefit from P2P system with XML IR • ex.: recent conference papers by computer scientists on percolation theory with • application of phase transition models to the analysis of Web graph dynamics TopX engine for XPath Full-Text [M. Theobald et al.: VLDB 05] > 100 Bio. page versions in Web archive • produced at > 20 TB/month by millions of Web sites • P2P system with time-travel queries could replace & enhance IA • ex.: „search engine scoring models“ as of Aug 1998 • „Jacques Chirac EU constitution“ from Jan 2005 to Jan 2006 Time-aware rank synopses [K. Berberich et al.: WWW 06] no killer app with business value !? • but: • P2P potentially useful also for server farms & grids • interesting non-business applications Gerhard Weikum August 3, 2006

  5. Outline Motivation and Research Directions  P2P Query Routing • • Overlap Awareness Discriminative Posting • P2P Link Analysis • • JXP Authority Scoring Personalized and Community-aware Ranking • • QRank and QReward Conclusion • Gerhard Weikum August 3, 2006

  6. Computational Model • Peers connected by overlay network • (e.g. DHT, random graph) and IP • Each peer has a full-fledged local search engine • with crawler/importer, indexer, query processor • Each peer has autonomously compiled (e.g. crawled) • its own content according to the user‘s thematic interests •  peer-specific collections • When a query is issued by a peer, it is first executed locally • and then possibly routed to carefully selected other peers • Peers can post summaries / synopses / metadata / QoS info • to (distr.) network-wide directory (space O(#terms * #peers)) • with efficient per-key lookup Gerhard Weikum August 3, 2006

  7. Minerva System Architecture peer ranking & statistics peer ranking & statistics url x: 37, 44, 12, ... peer lists (directory) term g: 13, 11, 45, ... term c: 13, 92, 45, ... term a: 17, 11, 92, ... term f: 43, 65, 92, ... url z: 54, 128, 7, ... url y: 75, 43, 12, ... query peer P0 book- marks B0 local index X0 term g: 13, 11, 45, ... based on scalable, churn-resilient DHT Query routing (QR) aims to optimize benefit/cost driven by distributed statistics on peers‘ content quality, content overlap, freshness, authority, trust, performability etc. Dynamically precompute „good peers“ to maintain a Semantic Overlay Network (SON) Exploit community input (bookmarks, etc.) Gerhard Weikum August 3, 2006

  8. P2P Query Routing (Resource Selection) • Principle: • Select peers with highest benefit/cost ratio where • benefit(Pi) = quality(Pi, q) ~  sim(q, Xi) + (1- ) sim (X0, Xi) • cost(Pi) ~ estimated response time or communication costs e.g. for sim(q,Xi) use prob.-IR CORI [Callan 95] and for sim(X0, Xi) use rel. entropy Method: • Precompute per-term peer-quality scores & keep in directory • QR aggregates PeerLists for query terms & selects top-k peers • Caveat: • Peer-peer similarity overfitsto content quality and ignores overlap Gerhard Weikum August 3, 2006

  9. Overlap Awareness [Bender et al.: SIGIR’05, EDBT‘06] Estimate overlap(p0, pj) = |X0Xj| / |X0Xj| between query initiator peer p0 and QR candidate pj using min-wise independent permutations (MIPs) [Broder 97] on the URLs in the collections of p0 and pj (with precomputed per-term MIPs posted to directory) Consider candidates pj in desc. order of estimated quality for q and re-rank peers by: Better: estimate novelty of additional pj and with and rank peers by integrated quality-novelty (IQN) Gerhard Weikum August 3, 2006

  10. Min-Wise Independent Permutations [Broder 97] MIPs (set1) MIPs (set2) 8 8 8 9 9 24 N 33 45 … 24 24 9 36 48 9 13 MIPs vector: minima of perm. estimated overlap = 2/6 set of ids 17 21 3 12 24 8 h1(x) = 7x + 3 mod 51 20 48 24 36 18 8 h2(x) = 5x + 6 mod 51 40 9 21 15 24 46 … hN(x) = 3x + 9 mod 51 9 21 18 45 30 33 compute N random permutations with: P[min{(x)|xS}=(x)] =1/|S| MIPs are unbiased estimator of overlap: P [min {h(x) | xA} = min {h(y) | yB}] = |AB| / |AB| MIPs can be viewed as repeated sampling of x, y from A, B Gerhard Weikum August 3, 2006

  11. IQN Experimental Results • Experiment: • based on 100 .Gov partitions (1.25 Mio. docs), assigned to 50 peers, • with each peer holding 10 partitions and 80% overlap for peers Pi, Pi+1 • with 50 TREC-2003 Web queries, • e.g.: „pest safety control“ „juvenile delinquency“, „Marijuana legalization“, etc. relative recall # queried peers For more experiments see our papers (Bender et al.: SIGIR’05, EDBT’06, WIRI’06) Gerhard Weikum August 3, 2006

  12. Discriminative Posting • peer pj posts a term only if pj has term-specific content • above average (or above quantile) of quality measure • reduces load on P2P directory • may ease decision on good query routing • requires global statistics on quality measures ! e.g. peer posts only if local df > *(global df) with  < 1 Experiment: 250 000 Web pages on 40 peers popular Google queries (e.g. „national hurricane center“) Gerhard Weikum August 3, 2006

  13. Efficiently Capturing Global Statistics • gdf (global doc. freq.) of a term is interesting key measure, • for discriminative posting or • for P2P result merging, • but overlap among peers makes simple distr. counting infeasible • hash sketches[Flajolet/Martin 85]: • duplicate-sensitive cardinality estimator for multisets • hash each multiset element x onto m-bit bitvector • and remember ls 1 bit (h(x)) • maxxS (h(x))estimates  log2 0.77351 |S| • with std.dev. / |S| = • rough intuition: • average multiple iid sketches Gerhard Weikum August 3, 2006

  14. Efficient & Accurate gdf Estimation [Bender et al.: WebDB 06] post a: 10110000 a: 10110000 c: 00101000 c: 00101000 post t: t: t: t: t: t: 00110000 00110100 00100000 00110110 00010000 00010010 t: 00100000 lookup(t.gdf) post Hash sketches of different peers collected at directory peer distributivity is free:i{(h(x)) | x Si} = {(h(x)) | x  i Si} • gdf estimation algorithm: • each peer p posts hash sketch for each (discriminative) term t to directory • directory peer for term t forms union of incoming hash sketches • when a peer needs to know gdf(t), simply ask directory peer for t • sliding-window techniques for dynamic adjustment dir(t) dir(c) dir(f) dir(d) dir(a) dir(e) Gerhard Weikum August 3, 2006

  15. gdf Estimation Experiments Experiment with steady-state P2P system: 1000 peers, each with 1000 randomly chosen docs from 1 Mio. docs Experiment with churn: Peers joining and leaving according to Poisson processes Gerhard Weikum August 3, 2006

  16. Outline Motivation and Research Directions  P2P Query Routing  • Overlap Awareness Discriminative Posting • P2P Link Analysis • • JXP Authority Scoring Personalized and Community-aware Ranking • • QRank and QReward Conclusion • Gerhard Weikum August 3, 2006

  17. Distributed PageRank (PR) Page authority important for final result scoring Exploit locality in Web link graph: construct block structure (disjoint graph partitioning) based on sites or domains • Compute page PR within site/domain & site/domain weights, • combine page scores with site/domain scores • [Kamvar03, Lee03, Broder04, Wang04, Wu05] or • communicate PR mass propagation across sites • [Abiteboul00, Sankaralingam03, Shi03, Jelasity05] Gerhard Weikum August 3, 2006

  18. PageRank (PR) in a P2P Network Every peer crawls Web fragments at its discretion and has its own local & personalized search engine  overlaps between peers’ graphs may occur Gerhard Weikum August 3, 2006

  19. JXP (Juxtaposed Approximate PageRank) [J.X. Parreira et al.: WebDB 05, VLDB 06] based on Markov-chain aggregation (state lumping) [Courtois 1977, Meyer 1988; cf. Chien et al. 2004, Langville/Meyer 2005] each peer represents external, a priori unknown part of the global graph by one superstate, a „world node“ • peers meet randomly • exchange their local graph fragments and PR vectors • learn about incoming edges to nodes of local graph • compute local PR on merged graphs or enhanced local graph • keep only improved PR and own local graph • don‘t keep other peers‘ graph fragments converges to global PR (experiments + theoretical arguments) convergence sped up by biased p2pDating strategy: prefer peers whose nodeset of outgoing links has high overlaps with our nodeset (use MIPs as synopses) Gerhard Weikum August 3, 2006

  20. JXP Algorithm at Work (1) W • Input: • G: local graph • GOUT: {qG | q s  sW} • n: #pages in G; N: #pages in U = GW • WIN(G): {pW | p q  qG} • WIN*(G)  WIN(G): known part of WIN(G) • Output: • *(q) for qG: • est. stationary prob‘s (PR) • *(G) = qG*(q)=1- *(W) • est. total mass of G F G W H • At each meeting with another peer: • compute • for all qG: • world self-loop: • compute all * values for G{w}; remember WIN*(G) info Gerhard Weikum August 3, 2006

  21. JXP Algorithm at Work (2) • Input: • G: local graph • GOUT: {qG | q s  sW} • n: #pages in G; N: #pages in U = GW • WIN(G): {pW | p q  qG} • WIN*(G)  WIN(G): known part of WIN(G) • Output: • *(q) for qG: • est. stationary prob‘s (PR) • *(G) = qG*(q)=1- *(W) • est. total mass of G F G W H • At each meeting with another peer: • compute • for all qG: • world self-loop: • compute all * values for G{w}; remember WIN*(G) info Gerhard Weikum August 3, 2006

  22. JXP Algorithm at Work (3) W • Input: • G: local graph • GOUT: {qG | q s  sW} • n: #pages in G; N: #pages in U = GW • WIN(G): {pW | p q  qG} • WIN*(G)  WIN(G): known part of WIN(G) • Output: • *(q) for qG: • est. stationary prob‘s (PR) • *(G) = qG*(q)=1- *(W) • est. total mass of G F G W H • At each meeting with another peer: • compute • for all qG: • world self-loop: • compute all * values for G{w}; remember WIN*(G) info Gerhard Weikum August 3, 2006

  23. JXP Convergence Theorem: In a fair sequence of P2P meetings, the JXP scores of every peer converge to the global PR scores. • Proof • based on Markov-chain aggr./disaggr. theory • [C.D. Meyer 1988, G.E. Cho & C.D. Meyer 1999] • for world node w: • JXP(w) is non-increasing and JXP(w)  PR(w) • for nodes q in peer‘s graph fragment: • JXP(q) is non-decreasing and JXP(q)  PR(q) Gerhard Weikum August 3, 2006

  24. p2pDating • Each peer pj precomputes two MIPs synopses for • M(pj): URLs in the collection of pj (the nodes of G) and • O(pj): URLs of the out-neighbors of pages of pj (OUT(G)) • repeatforever • peer pj randomly picks „blind date“ candidate pd: • pj and pk exchange their O synopses; • they may also recommend to each other a set of friends pf • and pass on their O synopses • peer pj maintains a list of dating candidates pc • ordered by resemblance (M(pj), O(pc)) • peer pj chooses best candidate for next date • (exchange of graphs, local PR computation, etc.) Gerhard Weikum August 3, 2006

  25. JXP Experiments 100 peers with simulated crawls of Amazon products categories (with recommended similar products as links) Ongoing work: peer trust measures & robustness to cheating similar and more results for real Web data also improves precision of query-result ranking, and query routing by combining quality-novelty with JXP mass Gerhard Weikum August 3, 2006

  26. Outline Motivation and Research Directions  P2P Query Routing  • Overlap Awareness Discriminative Posting •  P2P Link Analysis • JXP Authority Scoring Personalized and Community-aware Ranking • • QRank and QReward Conclusion • Gerhard Weikum August 3, 2006

  27. Personalized PageRank [Haveliwala et al. 2002] Authority (page q) = stationary prob. of visiting q Idea: random jumps favor designated high-quality pages such as personal bookmarks, frequently visited pages, etc. with random walk: uniformly random choice of links + biased jumps to personal favorites (or trusted pages or ...) see also: Jeh 2003, Benczur 2004, Gyöngyi 2004, Guha 2004 Gerhard Weikum August 3, 2006

  28. Exploiting Query Logs and Click Streams [J. Luxenburger et al.: WISE 04] max planck gesellschaft max planck mpg budget J.K. MPII MPII J.K. from PageRank: uniformly random choice of links + random jumps to QRank: + query-doc transitions + query-query transitions + doc-doc transitions on implicit links (w/ thesaurus) with probabilities estimated from log statistics Gerhard Weikum August 3, 2006

  29. Small-Scale Experiments Setup: 70 000 Wikipedia docs, 18 volunteers posing Trivial-Pursuit queries ca. 500 queries, ca. 300 refinements, ca. 1000 positive clicks ca. 15 000 implicit links based on doc-doc similarity • Results (assessment by blind-test users): • QRank top-10 result preferred over PageRank in 81% of all cases • QRank has 50.3% precision@10, PageRank has 33.9% Untrained example query „philosophy“: PageRank QRank x 1. Philosophy Philosophy 2. GNU free doc. license GNU free doc. license 3. Free software foundation Early modern philosophy 4. Richard Stallman Mysticism 5. Debian Aristotle Gerhard Weikum August 3, 2006

  30. Negative Feedback & Assessment • Users give implicit or explicit negative assessments: • non-clickedquery results ranked higher than clicked ones • encountered spam pages or personally disliked pages • ratings of pages or other users in social taggingnetworks • very valuable human input, but typically sparse • Approaches and problems using biased random walks • for quality&trust propagation [Eiron 2004, Guha 2004, Luxenburger 2006]: • penalize neg. pages by reducing their random-jump prob. • source-specific random-jump prob‘s and self-loops • force backward step or random jump when reaching neg. page • but probabilities are non-negative and L1-normalized •  ranking models become technically convoluted Better approach: decouple random walk from trust propagation  Markov reward models Gerhard Weikum August 3, 2006

  31. Markov Reward Models • discrete-time or continuous-time Markov chain with • state-specific lump reward rj R whenever j is entered • transition-specific lump reward rij  R when ij is traversed • (plus reward rates in CTMC case) • penalties expressed as negative rewards analysis of transient and stationary properties (used in queueing and performability models  textbooks by H.C. Tijms, R.W. Wolff; surveys by Haverkort/Trivedi) gained reward until step n: long-run average reward: Gerhard Weikum August 3, 2006

  32. QReward Ranking [J. Luxenburger et al.: WebDB 06] + +  • Add queries and users as nodes to the state graph • and connect to clicked, non-clicked, rated pages  + +  + +  • Associate transition-specific lump rewards • +1 for each positive assessment • -1 for each negative assessment • 0 otherwise + • Perform random walk in standard way, • using links and random jumps, yielding stationary prob‘s j • Compute long-run average reward gj for each state j • Quality of page j :=  gj + (1 ) j Gerhard Weikum August 3, 2006

  33. Fast Computation of QReward Renewal-Reward Theorem (Wolff p. 60): Compute jvalues as usual by power iteration (using QRank) but we need sufficient accuracy for all i, not just for the high-ranked ones  iterate QRank for j, QReward, and quality score; stop when quality scores of top pages converge Gerhard Weikum August 3, 2006

  34. Preliminary Experiments Setup: 70 000 Wikipedia docs, 18 volunteers posing queries ca. 500 queries, ca. 300 refinements, ca. 1000 positive clicks, ca. 2000 implict negative assessments(cf. Joachims et al.: SIGIR 05) • Results (based on relevance majority votes of 3 users): • PR has MAP 0.45 for top-15 of 14 test queries, • QRank has MAP 0.51, QReward has MAP 0.56 Ongoing work: combine with personalized LMs; trust models; larger-scale experimentation Example query „political system China“: PageRank QReward x 1. ChinaOne country, two systems 2. People‘s Republic of ChinaChina 3. List of countries Party discipline 4. Country List of countries 5. Chinese languageCommunist state Gerhard Weikum August 3, 2006

  35. Outline Motivation and Research Directions  P2P Query Routing  • Overlap Awareness Discriminative Posting • P2P Link Analysis  • JXP Authority Scoring Personalized and Community-aware Ranking  • QRank and QReward Conclusion • Gerhard Weikum August 3, 2006

  36. Conclusion: Challenges Remain Open • Distributed Statistics Management • Key to Query Routing, Quality/Overlap Estimation, Ranking (PR etc.) • Capturing Global Statistics in Decentralized Manner • Efficiently Disseminating Statistical Synopses • Robustness to Churn and Cheating • „Statistically Semantic & Social“ Overlay Networks • Experimental Evaluation • Benchmarking Methodology • Large-scale P2P Testbed • Capturing User/Community Behavior Gerhard Weikum August 3, 2006

More Related