Dynamic Search Algorithm in Unstructured Peer-to-Peer Networks

By:

I. V. Sai Manoj07K91A0557P. Rakesh07K91A0532P. Sowmya07K91A0548

Batch - 10

ABSTRACT

• AIM : To design efficient search algorithms as a key challenge in unstructured peer-to-peer networks.

• Previous designs: Flooding, Random Walk, etc.

• DSA takes the advantage of various contexts under which each search algorithm performs well and efficient.

What are Unstructured Peer-to-Peer Networks ?

Internet

Example: Gnutella Network

Previous Works

BFS DFS

inefficient

Generates Overloading Not meeting the userrequirements

Flooding - BFS

Not Scalable

Quick & AggressiveSearch

Random Walk - DFS

Long Search Time

Reduced Search Cost

Flooding Random Walk

Not Scalable

Quick & AggressiveSearch

Long Search Time

Reduced Search Cost

KBS Knowledge Learnt fromPrevious Search Results

DSA

Performance Evaluation• Success Rate• Search Time• Number of Query Hits• Number of Query Messages• Query Efficiency• Search Efficiency

Dynamic Search Algorithm – Operation

Phase 1 Phase 2

‘h’ ‘n’

When h <= n When h > n• Query messages sent to ( d . p ) neighbors

• Acts as Flooding or MBFS

• Strategy switches to RW

• Each Node that receives the query message would send the query message to one of its neighbors

Requirements• HARDWARE : Any processor above 500 MHz

128 MB RAM10 GB HDD650 MB Compact DiskStandard Keyboard and MouseVGA and High Resolution Monitor

• SOFTWARE : Windows OSJDK 1.5Microsoft SQL Server 2005Java Swings

Thank you…

QUERIES?????