Routing Indices For Peer-to-Peer Systems Svetlana Strunjas University of Cincinnati May,2002.

Slides:



Advertisements
Similar presentations
Connectivity-Aware Routing (CAR) in Vehicular Ad Hoc Networks Valery Naumov & Thomas R. Gross ETH Zurich, Switzerland IEEE INFOCOM 2007.
Advertisements

Ion Stoica, Robert Morris, David Karger, M. Frans Kaashoek, Hari Balakrishnan MIT and Berkeley presented by Daniel Figueiredo Chord: A Scalable Peer-to-peer.
Scalable Content-Addressable Network Lintao Liu
Multicast in Wireless Mesh Network Xuan (William) Zhang Xun Shi.
Rumor Routing in Sensor Networks David Braginsky and Deborah Estrin Presented By Tu Tran 1.
University of Cincinnati1 Towards A Content-Based Aggregation Network By Shagun Kakkar May 29, 2002.
Routing Indices For Peer-to-Peer Systems Arturo Crespo, Hector Garcia-Molina Stanford ICDCS 2002.
P2p, Spring 05 1 Topics in Database Systems: Data Management in Peer-to-Peer Systems Routing indexes A. Crespo & H. Garcia-Molina ICDCS 02.
Denial-of-Service Resilience in Peer-to-Peer Systems D. Dumitriu, E. Knightly, A. Kuzmanovic, I. Stoica and W. Zwaenepoel Presenter: Yan Gao.
Mobile and Wireless Computing Institute for Computer Science, University of Freiburg Western Australian Interactive Virtual Environments Centre (IVEC)
P2p, Spring 05 1 Topics in Database Systems: Data Management in Peer-to-Peer Systems March 29, 2005.
More routing protocols Alec Woo June 18 th, 2002.
CMPE 150- Introduction to Computer Networks 1 CMPE 150 Fall 2005 Lecture 22 Introduction to Computer Networks.
1 Efficient Retrieval of User Contents in MANETs Marco Fiore, Claudio Casetti, Carla-Fabiana Chiasserini Dipartimento di Elettronica, Politecnico di Torino,
Efficient Content Location Using Interest-based Locality in Peer-to-Peer Systems Presented by: Lin Wing Kai.
A Scalable Content-Addressable Network Authors: S. Ratnasamy, P. Francis, M. Handley, R. Karp, S. Shenker University of California, Berkeley Presenter:
presented by Hasan SÖZER1 Scalable P2P Search Daniel A. Menascé George Mason University.
A Distributed Search Service for Peer-to-Peer File Sharing in Mobile Application Presented by Tony Sung On Loy, MC Lab, CUHK IE 1 A Distributed Search.
Vassilios V. Dimakopoulos and Evaggelia Pitoura Distributed Data Management Lab Dept. of Computer Science, Univ. of Ioannina, Greece
LPT for Data Aggregation in Wireless Sensor networks Marc Lee and Vincent W.S Wong Department of Electrical and Computer Engineering, University of British.
Chord-over-Chord Overlay Sudhindra Rao Ph.D Qualifier Exam Department of ECECS.
Freenet A Distributed Anonymous Information Storage and Retrieval System I Clarke O Sandberg I Clarke O Sandberg B WileyT W Hong.
Routing.
Efficient Search in Peer to Peer Networks By: Beverly Yang Hector Garcia-Molina Presented By: Anshumaan Rajshiva Date: May 20,2002.
Spring Routing & Switching Umar Kalim Dept. of Communication Systems Engineering 06/04/2007.
Searching in Unstructured Networks Joining Theory with P-P2P.
MULTICASTING Network Security.
Peer-to-peer file-sharing over mobile ad hoc networks Gang Ding and Bharat Bhargava Department of Computer Sciences Purdue University Pervasive Computing.
1CS 6401 Peer-to-Peer Networks Outline Overview Gnutella Structured Overlays BitTorrent.
Ad Hoc Wireless Routing COS 461: Computer Networks
ROUTING ON THE INTERNET COSC Aug-15. Routing Protocols  routers receive and forward packets  make decisions based on knowledge of topology.
INTRODUCTION TO PEER TO PEER NETWORKS Z.M. Joseph CSE 6392 – DB Exploration Spring 2006 CSE, UT Arlington.
1 Napster & Gnutella An Overview. 2 About Napster Distributed application allowing users to search and exchange MP3 files. Written by Shawn Fanning in.
IR Techniques For P2P Networks1 Information Retrieval Techniques For Peer-To-Peer Networks Demetrios Zeinalipour-Yazti, Vana Kalogeraki and Dimitrios Gunopulos.
09/07/2004Peer-to-Peer Systems in Mobile Ad-hoc Networks 1 Lookup Service for Peer-to-Peer Systems in Mobile Ad-hoc Networks M. Tech Project Presentation.
Searching In Peer-To-Peer Networks Chunlin Yang. What’s P2P - Unofficial Definition All of the computers in the network are equal Each computer functions.
Distributed Asynchronous Bellman-Ford Algorithm
Lyon, June 26th 2006 ICPS'06: IEEE International Conference on Pervasive Services 2006 Routing and Localization Services in Self-Organizing Wireless Ad-Hoc.
Peer to Peer Research survey TingYang Chang. Intro. Of P2P Computers of the system was known as peers which sharing data files with each other. Build.
Ad-hoc On-Demand Distance Vector Routing (AODV) and simulation in network simulator.
P2p, Fall 06 1 Topics in Database Systems: Data Management in Peer-to-Peer Systems Routing indexes A. Crespo & H. Garcia-Molina ICDCS 02.
Using the Small-World Model to Improve Freenet Performance Hui Zhang Ashish Goel Ramesh Govindan USC.
Chord: A Scalable Peer-to-peer Lookup Protocol for Internet Applications Xiaozhou Li COS 461: Computer Networks (precept 04/06/12) Princeton University.
Routing Indices For P-to-P Systems ICDCS Introduction Search in a P2P system –Mechanisms without an index –Mechanisms with specialized index nodes.
TRICKLE: A Self-Regulating Algorithm for Code Propagation and Maintenance in Wireless Sensor Networks Philip Levis, Neil Patel, Scott Shenker and David.
Benjamin AraiUniversity of California, Riverside Reliable Hierarchical Data Storage in Sensor Networks Song Lin – Benjamin.
Salah A. Aly,Moustafa Youssef, Hager S. Darwish,Mahmoud Zidan Distributed Flooding-based Storage Algorithms for Large-Scale Wireless Sensor Networks Communications,
Connectivity-Aware Routing (CAR) in Vehicular Ad Hoc Networks Valery Naumov & Thomas R. Gross ETH Zurich, Switzerland IEEE INFOCOM 2007.
Rushing Attacks and Defense in Wireless Ad Hoc Network Routing Protocols ► Acts as denial of service by disrupting the flow of data between a source and.
03/19/02Scalab Seminar Series1 Routing in Peer-to-Peer Systems Ramaswamy N.Vadivelu Scalab, ASU.
Rendezvous Regions: A Scalable Architecture for Service Location and Data-Centric Storage in Large-Scale Wireless Sensor Networks Karim Seada, Ahmed Helmy.
Network Coding Data Collecting Mechanism based on Prioritized Degree Distribution in Wireless Sensor Network Wei Zhang, Xianghua Xu, Qinchao Zhang, Jian.
1 Improve search in unstructured P2P overlay. 2 Peer-to-peer Networks Peers are connected by an overlay network. Users cooperate to share files (e.g.,
Graph Algorithms Why graph algorithms ? It is not a “graph theory” course! Many problems in networks can be modeled as graph problems. Note that -The topology.
Ad Hoc On-Demand Distance Vector Routing (AODV) ietf
Large Scale Sharing Marco F. Duarte COMP 520: Distributed Systems September 19, 2004.
P2P Networking: Freenet Adriane Lau November 9, 2004 MIE456F.
Distance Vector Routing
CS 6401 Intra-domain Routing Outline Introduction to Routing Distance Vector Algorithm.
Incrementally Improving Lookup Latency in Distributed Hash Table Systems Hui Zhang 1, Ashish Goel 2, Ramesh Govindan 1 1 University of Southern California.
Performance Comparison of Ad Hoc Network Routing Protocols Presented by Venkata Suresh Tamminiedi Computer Science Department Georgia State University.
Distributed Caching and Adaptive Search in Multilayer P2P Networks Chen Wang, Li Xiao, Yunhao Liu, Pei Zheng The 24th International Conference on Distributed.
Network Topologies for Scalable Multi-User Virtual Environments Lingrui Liang.
Chord: A Scalable Peer-to-Peer Lookup Service for Internet Applications * CS587x Lecture Department of Computer Science Iowa State University *I. Stoica,
Intra-Domain Routing Jacob Strauss September 14, 2006.
Routing in Packet Networks Shortest Path Routing
Paraskevi Raftopoulou, Euripides G.M. Petrakis
DATA RETRIEVAL IN ADHOC NETWORKS
DSDV Destination-Sequenced Distance-Vector Routing Protocol
Routing in Mobile Wireless Networks Neil Tang 11/14/2008
Presentation transcript:

Routing Indices For Peer-to-Peer Systems Svetlana Strunjas University of Cincinnati May,2002

Outline  Motivation and background  P2P systems  Routing Indices  Experimental results  Conclusions

Motivation  Create efficient and feasible distributed – index search mechanism for P2P systems

Background  Napster – centralized indices (vulnerable to attack and difficult to scale)  Gnutella – queries flood significant part of the network(simple but very costly approach)

Background(cont’d 1)  Freenet – each node builds an index with the location of recently requested documents (queries are restricted to the specific documents and it takes time for node to build an effective index)

Background(cont’d 2)  Oceanstore – distributed – index approach similar to Routing Indices(special case of Compound RIs)  Key difference is in static network and queries on document identifier (Oceanstore) versus dynamic network and queries on content of document(Routing Indices)

Background(cont’d 3)  Selecting a neighbor for forwarding query using Routing Indices is generalization of the Bellman-Ford algorithm  Bellman – Ford algorithm -transmits packet between two nodes through the shortest route -destination of packet is pre-defined(IP routing)  Routing Indices -transmit packet from node to node in order to find best answer to the query -destination of packet is not pre-defined, but instead depends on query contained in packet

P2P Systems (Introduction)  Formed by a large number of nodes that can join or leave the system at any time and have equal capabilities  Each node has local documents database which can be accesed through local index  The local index receives content queries and returns pointer to the document with requested content

Query Processing in Distributed P2P Systems  Users submit queries to any node along with the stop condition  Node receives query, evaluates the query against it’s own database,returns pointer to any results, and, if the stop condition is not reached, forwards the query to neighbor(s)(parallel or sequentially)

Routing Indices Introduction  RI allows a node to select “best“ neighbors to send a query to  It is data structure(and associated algorithm) which,for given query, returns a list of neighbors, rank according to their goodness for query  Goodness reflects the number of documents in “nearby” nodes

Routing Indices Types  Compound Routing Indices(CRI)  Hop-count Routing Indices(HRI)  Exponential Routing Indices(ERI)

Compound Routing Indices (example)

CRI(cont’d1)  Documents are on zero or more topics,and queries request documents on particular topic  Each node has a local index for quickly finding local documents  CRI contain the number of documents along each path and the number of documents of each topic of interest

CRI (“goodness” estimators)  Measure of “goodness” of the neighbor, for given query, is number of documents that can be found in that path  Estimated number of results along each path is given by formula:

CRI (drawbacks)  CRI do not take into account the difference in cost due to the number of hops necessary to reach the document  Are not applicable if we have cycles in network

Using RI RIs with four topics of interests : database,network,theory and languages

Storage space required  It can be adjusted by increasing or decreasing the level of summarization of index  For centralized systems it is C*(T+1)*N bytes  For distributed systems it is C*(T+1)*B*N C – number of categories T – counter size in bytes B – branching factor N – number of nodes

Creating (maintaining)RIs

Disconnection from Network

Algorithm for Creating/Updating RIs  There are two phases : 1.Newly connected node sends a summary of its local index to its new neighbors (creation phase) 2.The node waits for update messages or for changes on its local index ; after updating its RI, the node sends a new aggregate RIs to all its neighbors

Algorithm for Answering Queries  It runs every time query is received  Query is given together with stop condition  There are three phases: 1.Node attempts to answer the query using its local db 2.If not enough results are obtained to reach stop condition, the algorithm ranks all the neighbors using estimating function 3.Node sends the query to each neighbors sequentially,checking if the stop condition was reached whenever each query returns

Hop-count RI  Store aggregated RI for each “hop” up to a maximum number of hops(horizon of the RI)

Hop-count RI(cont’d1)  Goodness of neighbor is defined as ratio between the number of documents available through that neighbor and the number of messages required to get those documents  Simple model that allows us to compute this ratio is regular-tree model

Hop-count RI(cont’d2)  Model assumes that all nodes are connected to exactly F+1 nodes, except ones at the leaves which are connected to only one  Goodness of the neighbor is given by formula: goodness() – estimating function for CRI N i [j] –RI entry for j hops through i-th neighbor F-fanout H-horizon

Hop-count RI (drawbacks)  HRI performances can be negatively affected by the lack of information beyond the horizon  Higher storage and transmission cost than CRI

Exponentially Aggregated RI  Stores the result of applying the regular-tree cost formula to hop-count RI

Exponentially Aggregated RI(cont’d)  Each entry of the ERI for node N contains a value computed as : th – height of the assumed regular tree goodness() – estimating function for CRIs N[j] – summary of the local index of neighbor j of N T – topic of inerest for entry  With ERIs we can keep information for all nodes accessible from each neighbor in the RI,while HRI’s do not have any information beyond horizon

Cycles in P2P Network There are three approaches for dealing with cycles:  No-op solution  Cycle avoidance solution  Cycle detection and recovery

No-op solution  Works only with the HRI and ERI  If there are cycles CRI algorithm can be trapped in an infinite loop

NO-op solution (cont’d1) G1=21.67 G2=23.58

NO-op solution (cont’d2)  In HRI cycles longer than horizon will not affect the RI  If we use regular-tree cost model we can limited effect of short cycles  In HRI and ERI difference between two values of goodness introduced by cycle is very small(algorithm will not propagate updates which are not much different than old value)

Cycle avoidance solution  Do not allow nodes to create an update connection to other node if such connection creates the cycle  Drawback : Absence of global information causes suboptimal update network

Cycle detection and recovery  Detects cycles sometime after they are formed and takes recovery actions to eliminate the effect of the cycles  Cycles are detected using unique message identifier  Recovery procedure can decrease accuracy of the RI

Experimental Results (Introduction)  Results are obtained using different models for elements of search mechanism  Steps of experiment: 1.Setting P2P system as network of nodes T, where each node contains set of documents 2.Users send requests consisting of query Q and stopCondition to a node of P2P system 3. search mechanism answers requests by obtaining a set of documents of size stopCondition that matches Q 4.search mechanism allows for update such as addition of nodes and new documents

Experimental Results (Techniques and measure of cost)  Search techniques : CRI, HRI, ERI, No-RI(for comparison purposes)  Measure of cost:numbers of messages generated

Experimental Results (Network topologies)  Topology of network defines number of nodes and how they are connected  In this model, three kinds of topologies are considered: 1.Tree 2.Tree with added cycles 3.Power-law graph

Experimental Results (Distribution of documents)  Distributions for modeling the location of document results : 1.Uniform distribution( all nodes have the same probability of having each document result) 2.80/20 biased distribution ( assigns uniformly 80% of the document results to the 20% of nodes, and the remaining 20% of the documents to the remaining 80% of nodes)

Experimental Results (Simulation parameters)

Experimental Results (Comparison of CRI,HRI and ERI)

Experimental Results (Number of results)

Experimental Results (Effects of overcounts)

Experimental Results (Effects of cycles)

Experimental Results (Network topology)

Experimental Results (Updates and network topology)

Experimental Results (Updates and cycle policy)

Experimental Results (Updates per minute)

Conclusions  ERI and HRI offer significant improvements versus not using an RI,while keeping update costs low  Routing indices (in particular ERI and HRI can help improve the search performance of current and future P2P systems

References: [1] A.Crespo, H. Garcia-Molina.Routing Indices For Peer-to-Peer Systems(2001).The 22 nd International Conference on Distributed Computing Systems,Vienna, Austria. [2] J.F.Kurose, K.W.Ross.(1999).”Routing Principles”.Computer Networking – A Top Down Approach Featuring Internet, 2 nd ed., Addison Wesley,4.