17-03-2014, 07:15 PM
Abstract
In this paper, we propose a query-feedback based
routing approach for efficient routing in unstructured
peer-to-peer network. In our approach, each node
maintains a table that stores information about past
queries that have been forward and answered. Each
past query represents a number of peers. Each
incoming query will be compared with the past queries
in the table and similar and reliable peers will be
selected to be routed to have an efficient routing
approach. Experimental results show that our
proposed approach outperforms other feedback based
query routing technique in unstructured peer-to-peer
networks.
1. Introduction
Peer-to-peer networks have become an active
research topic in computer science in recent years. It
has gained popularity since the existence of peer to
peer application such as Napster [1] and Gnutella [2,
3]. The advancement of high-speed internet access has
even transform peer to peer application to be very easy
for the user to use, search and share resources.
Unstructured networks do not enforce any
structure in the network and it is popular due to its
robustness and scalability. Unstructured network has
its own share of problems when flooding is used for
query routing which is computationally expensive and
demanding large communication overhead. Therefore,
it demands a very efficient search routing mechanism
to overcome this problem that uses smaller overhead
and messages, reduced query time and network traffic.
In order to achieve this, selective query routing is
needed. The challenge of selecting peers for query
routing in unstructured peer-to-peer network is that,
the network is dynamic and no administration
imposed.
The number of the retrieved results is also need to
be considered to justify the performance of the routing
approach. Therefore, understanding the content of
query is also important and the challenge to achieve
good retrieval is that there is no global information.
We focused our research on the unstructured peer-topeer
networks because most of the real-world large
peer-to-peer systems are unstructured.
In this paper, we present a query-feedback and
query similarity based peer-to-peer query routing
approach in unstructured peer-to-peer networks. The
approach uses information from the query feedback
data and the query content to improve the query
routing in unstructured peer-to-peer network with very
minimal cost and information.
In this paper, we propose a query-feedback based
routing approach for efficient routing in unstructured
peer-to-peer network. In our approach, each node
maintains a table that stores information about past
queries that have been forward and answered. Each
past query represents a number of peers. Each
incoming query will be compared with the past queries
in the table and similar and reliable peers will be
selected to be routed to have an efficient routing
approach. Experimental results show that our
proposed approach outperforms other feedback based
query routing technique in unstructured peer-to-peer
networks.
1. Introduction
Peer-to-peer networks have become an active
research topic in computer science in recent years. It
has gained popularity since the existence of peer to
peer application such as Napster [1] and Gnutella [2,
3]. The advancement of high-speed internet access has
even transform peer to peer application to be very easy
for the user to use, search and share resources.
Unstructured networks do not enforce any
structure in the network and it is popular due to its
robustness and scalability. Unstructured network has
its own share of problems when flooding is used for
query routing which is computationally expensive and
demanding large communication overhead. Therefore,
it demands a very efficient search routing mechanism
to overcome this problem that uses smaller overhead
and messages, reduced query time and network traffic.
In order to achieve this, selective query routing is
needed. The challenge of selecting peers for query
routing in unstructured peer-to-peer network is that,
the network is dynamic and no administration
imposed.
The number of the retrieved results is also need to
be considered to justify the performance of the routing
approach. Therefore, understanding the content of
query is also important and the challenge to achieve
good retrieval is that there is no global information.
We focused our research on the unstructured peer-topeer
networks because most of the real-world large
peer-to-peer systems are unstructured.
In this paper, we present a query-feedback and
query similarity based peer-to-peer query routing
approach in unstructured peer-to-peer networks. The
approach uses information from the query feedback
data and the query content to improve the query
routing in unstructured peer-to-peer network with very
minimal cost and information.