Thesis Abstract and Dissertation Abstracts
Design Of Efficient Searching Algorithms For Unstructured Peer-to-peer Networks
Abstract Category: Engineering
Course / Degree: Ph.D
Institution / University: Bharath University, India
Published in: 2012
Nowadays Peer-to-peer networks are very popular and common on small local area networks. Both wired and wireless home networks can be configured as Peer-to-peer networks. Much research is being done on peer-to-peer networks for various purposes. This research work aims to find some efficient searching algorithms or techniques to search resources in unstructured peer-to-peer network. In this research work, an efficient algorithm the Search Result Path Caching (SRPC) is developed to solve the search time issues of the existing algorithms. In this algorithm, the dynamic searching method is followed. As per this new algorithm, both breadth first search and depth first search are followed to find a resource. The path results obtained upon searching are stored in a path hash table. While searching, the search peer first evaluates its path table to find whether the path of the resource is available or not. If it is exists, the source node retrieves it from the table. This path table concept saves more search time and so this algorithm is better than the previous algorithm as shown by the results. If there is no path, it will send a query message to all the nodes to find the resource. In a P2P network, nodes may join or leave the network. If a node leaves, that particular path becomes invalid. So, as per this algorithm, the path table is periodically checked to remove invalid paths. This periodical checking of the invalid paths is a new technique of this algorithm.
The another algorithm proposed in this dissertation is the Friend and Path Result Sharing (FRPS) algorithm. This algorithm is developed to solve the issues like the generation of a lot of query messages inside the network, more search cost, more network traffic and poor scalability of the existing algorithms. In this algorithm, the friend set and HOF node (head of the friend set) concepts are involved. The HOF node is the head of all nodes in its friend set or group. The source node sends query messages only to the HOF nodes. This HOF node can answer a query message on behalf of all the other nodes in the set. Since the source sends the query only to the HOF nodes which are all within a 7 hop count distance, the number of query messages inside the network is limited, and thus reduce the traffic. Similarly, the HOF node bundles the path information of the resources in to a single array and sends it back to source node. So, both forward and backward traffic is reduced. This technique uses flooding and random walk techniques with the new Time to live (TTL) variable to search the nodes which are not in the friend sets. This algorithm reduces 60% of search cost.
Thesis Keywords/Search Tags:
invalid path validation, friend set, conditional deepening, odd and even depths
This Thesis Abstract may be cited as follows:
No user preference. Please use the standard reference methodology.
Submission Details: Thesis Abstract submitted by Anbu Subbiah from India on 21-Mar-2012 03:28.
Abstract has been viewed 2804 times (since 7 Mar 2010).
Anbu Subbiah Contact Details: Email: anbusubbaiah@yahoo.com Phone: 9884559506
Disclaimer
Great care has been taken to ensure that this information is correct, however ThesisAbstracts.com cannot accept responsibility for the contents of this Thesis abstract titled "Design Of Efficient Searching Algorithms For Unstructured Peer-to-peer Networks". This abstract has been submitted by Anbu Subbiah on 21-Mar-2012 03:28. You may report a problem using the contact form.
© Copyright 2003 - 2024 of ThesisAbstracts.com and respective owners.