Presentation is loading. Please wait.

Presentation is loading. Please wait.

A Blueprint for Constructing Peer-to-Peer Systems Robust to Dynamic Worst-Case Joins and Leaves Fabian Kuhn, Microsoft Research, Silicon Valley Stefan.

Similar presentations


Presentation on theme: "A Blueprint for Constructing Peer-to-Peer Systems Robust to Dynamic Worst-Case Joins and Leaves Fabian Kuhn, Microsoft Research, Silicon Valley Stefan."— Presentation transcript:

1 A Blueprint for Constructing Peer-to-Peer Systems Robust to Dynamic Worst-Case Joins and Leaves Fabian Kuhn, Microsoft Research, Silicon Valley Stefan Schmid, ETH Zurich Joest Smit, ETH Zurich Roger Wattenhofer, ETH Zurich 14th IEEE Int. Workshop on Quality of Service (IWQoS) Yale University, New Haven, CT, USA, June 2006

2 Stefan Schmid, ETH Zurich @ IWQoS 20062 Brief Intro to Peer-to-Peer Computing (1) P2P computing = power by accumulating distributed resources (CPU cycles, disk space, …) vs Client / Server -Centralized („one machine“) -Bottleneck -Single Point of Failure -… Peer-to-Peer -Decentralized („all machines“) -Scalable -Efficient -…

3 Stefan Schmid, ETH Zurich @ IWQoS 20063 Brief Intro to Peer-to-Peer Computing (2) Examples: - computing power (Folding@Home, …) - file sharing (eMule, Kangoo, …) - internet telephony (Skype, …) - media streaming (Swistry, …) file sharing distributed computations Swistry (live media streaming)

4 Stefan Schmid, ETH Zurich @ IWQoS 20064 Churn (1) But: unlike server, peers are transient! –Machines under control of individual users –E.g., just connecting to download one file –Membership changes are called churn Successful P2P systems have to cope with churn (i.e., guarantee correctness, efficiency, etc.)!

5 Stefan Schmid, ETH Zurich @ IWQoS 20065 Churn (2) Dynamic resources: A challenge in P2P computing! Churn characteristics: –Depends on application (Skype vs. eMule vs. …) –But: There may be dozens of membership changes per second! –Peers may crash without notice! How can peers collaborate in spite of churn?

6 Stefan Schmid, ETH Zurich @ IWQoS 20066 Churn (3) Churn is important, as it threatens “advantages of P2P computing”! a lot of churn We have to actively maintain P2P systems!

7 Stefan Schmid, ETH Zurich @ IWQoS 20067 Our Paper… Peer degree, network diameter, … „adversary“ non-stop attacks weakest part (system „never fully repaired, but always fully functional“) … presents techniques to: - … build and provably maintain P2P systems with desirable properties… - … in spite of ongoing worst-case membership changes. Unfortunately, only few P2P systems have been analyzed under churn! Our paper…

8 Stefan Schmid, ETH Zurich @ IWQoS 20068 Talk Outline A model for dynamics Overview of techniques Example: A robust system with degree and diameter O(log n / loglog n) Conclusion

9 Stefan Schmid, ETH Zurich @ IWQoS 20069 Talk Outline A model for dynamics Overview of techniques Example: A robust system with degree and diameter O(log n / loglog n) Conclusion

10 Stefan Schmid, ETH Zurich @ IWQoS 200610 Model for Dynamics Churn = possibly concurrent membership changes, at any time! We assume worst-case perspective: Adversary ADV(J,L) –i.e., joins and leaves may take place at the weakest spot of the network Synchronous model: time divided into rounds (e.g., max round trip time) time ADV(J,L): In each round, at most J peers may joins and at most L peers leave (crash).

11 Stefan Schmid, ETH Zurich @ IWQoS 200611 Talk Outline A model for dynamics Overview of techniques Example: A robust system with degree and diameter O(log n / loglog n) Conclusion

12 Stefan Schmid, ETH Zurich @ IWQoS 200612 Talk Outline A model for dynamics Overview of techniques Example: A robust system with degree and diameter O(log n / loglog n) Conclusion

13 Stefan Schmid, ETH Zurich @ IWQoS 200613 Topology Maintenance An efficient P2P topology under churn: π1π1 π2π2 Almost impossible to maintain the hypercube! –How does peer 1 know that it should replace peer 2? –How does it get there when there are concurrent joins and leaves? –… Is there a more robust topology but with same small degree and diameter?

14 Stefan Schmid, ETH Zurich @ IWQoS 200614 Our Approach Simple idea: Simulate the topology!

15 Stefan Schmid, ETH Zurich @ IWQoS 200615 General Recipe for Robust Topologies 1.Take a graph with desirable properties -Low diameter, low peer degree, etc. 2.Replace vertices by a set of peers 3. Maintain it: a. Permanently run a peer distribution algorithm which ensures that all vertices have roughly the same amount of peers (“token distribution algorithm”). b. Estimate the total number of peers in the system and change “dimension of topology” accordingly (“information aggregation algorithm” and “scaling algorithm”). Resulting structure has similar properties as original graph (e.g., connectivity, degree, …), but is also maintainable under churn! There is always at least one peer per node (but not too many either).

16 Stefan Schmid, ETH Zurich @ IWQoS 200616 Talk Outline A model for dynamics Overview of techniques Example: A robust system with degree and diameter O(log n / loglog n) Conclusion

17 Stefan Schmid, ETH Zurich @ IWQoS 200617 Talk Outline A model for dynamics Overview of techniques Example: A robust system with degree and diameter O(log n / loglog n) Conclusion

18 Stefan Schmid, ETH Zurich @ IWQoS 200618 The Pancake Graph (1) A robust system with degree and diameter O(log n / loglog n): the pancake graph –E.g., Papadimitriou & Gates! Pancake of dimension d: –d! nodes represented by unique permutation {l 1, …, l d } of set {1,…,d} –Two nodes u and v are adjacent iff u is a prefix-inversion of v 4-dimensional pancake: 12344321 2134 3214

19 Stefan Schmid, ETH Zurich @ IWQoS 200619 The Pancake Graph (2) Properties –Node degree Θ (log n / loglog n) –Diameter Θ (log n / loglog n) –… where n is the total number of nodes –A factor loglog n better than hypercube! –But: difficult graph (diameter unknown!) No other graph can have a smaller degree and a smaller diameter!

20 Stefan Schmid, ETH Zurich @ IWQoS 200620 Contribution Using peer distribution and information aggregation algorithms… … on the simulated pancake topology, we can construct: a peer-to-peer system (“distributed hash table”) with –Peer degree and lookup / network diameter in Θ (log n / loglog n) –Robustness to ADV(Θ (log n / loglog n), Θ (log n / loglog n)) –No data is ever lost! Asymptotically optimal!

21 Stefan Schmid, ETH Zurich @ IWQoS 200621 The Pancake System

22 Stefan Schmid, ETH Zurich @ IWQoS 200622

23 Stefan Schmid, ETH Zurich @ IWQoS 200623 Basic Components Peer Distribution Algorithm –Balance peers between neighboring nodes –One (pancake-) dimension after the other! Information Aggregation Algorithm –Exploit recursive structure of pancake –Aggregate „sub-pancakes“ with increasing order Both happens concurrently to ongoing churn! If fast enough, pancake is maintained! Always at least one peer per node!

24 Stefan Schmid, ETH Zurich @ IWQoS 200624 Internals (1) How are peers connected in the simulated topology? Idea: Clique Matching Problem: - There are up to Θ ((log n / loglog n) 2 ) many peers in each node - Clique would render peer degree too large! Inside node, peers have to form a grid!

25 Stefan Schmid, ETH Zurich @ IWQoS 200625 Internals (2) Solution: Grid Matching Grid Each peer is connected to all peers which are either in the same row or column Degree is OK now, and still robust enough to churn!

26 Stefan Schmid, ETH Zurich @ IWQoS 200626 Internals (3) “Distributed Hash Table”: - Stores data at nodes - But on which peers of node of given ID? - On just one is bad in dynamic enviroment! All? - Possible! - But much data movement during peer distribution. Better idea: - Peers of a node fall into two categories: Protons and Electrons - Protons = „core peers“, store data, are „seldom“ used during token distribution - Electrons = „peripheral peers“, do not store data, are used for balancing - Make sure that there are always enough protons (no data loss)!

27 Stefan Schmid, ETH Zurich @ IWQoS 200627 Talk Outline A model for dynamics Overview of techniques Example: A robust system with degree and diameter O(log n / loglog n) Conclusion

28 Stefan Schmid, ETH Zurich @ IWQoS 200628 Talk Outline A model for dynamics Overview of techniques Example: A robust system with degree and diameter O(log n / loglog n) Conclusion

29 Stefan Schmid, ETH Zurich @ IWQoS 200629 Conclusion Contribution: A scheme to maintain quality of a peer-to-peer system in spite of worst- case membership changes. –Ingredients: “base graph”, token distribution & information aggregation algorithm –Proofs possible! Simulated graph can have similar properties as base graph. –Degree, diameter, etc. –May require some additional thinking, though! (e.g., grid) A peer-to-peer system with degree and diameter in O(log n/loglog n) which tolerates O(log n/loglog n) joins and leaves per round. –Better than often-used hypercube graph! –But: difficult graph! (e.g., dimension change) Open questions –How to coordinate dynamic peers or resources: An exciting field of research! –E.g.: Self-stabilization, dirty leaves, etc.

30 Stefan Schmid, ETH Zurich @ IWQoS 200630 Questions and Feedback? Stefan Schmid Distributed Computing Group schmiste@ethz.ch http://dcg.ethz.ch/members/stefan.html Thank you for your attention!


Download ppt "A Blueprint for Constructing Peer-to-Peer Systems Robust to Dynamic Worst-Case Joins and Leaves Fabian Kuhn, Microsoft Research, Silicon Valley Stefan."

Similar presentations


Ads by Google