W3C home > Mailing lists > Public > www-ws-arch@w3.org > January 2004

RE: peer to peer discovery

From: Katia Sycara <katia@cs.cmu.edu>
Date: Thu, 15 Jan 2004 16:08:10 -0500
To: 'Francis McCabe' <fgm@fla.fujitsu.com>, "'Champion, Mike'" <Mike.Champion@SoftwareAG-USA.com>, dbooth@w3.org
Cc: www-ws-arch@w3.org, katia@cs.cmu.edu
Message-ID: <005501c3dbab$aea5ae10$d1bd0280@scs.ad.cs.cmu.edu>
Here is the P2P discovery text to be put in section 3.1.4  




P2P does not rely on a discovery service but the requesters and other peers
interact to propagate a service request. Peer-to-Peer (P2P) computing
provides an alternative that does not rely on centralized registries; rather
it allows Web services to discover each other dynamically. Under this view,
a Web service is a node in a network of peers, which may or may not be Web
services. At discovery time a requesting Web service queries its neighbors
in the network. If any one of them matches the request, then it replies,
otherwise it queries its own neighboring peers and the query propagates
through the network. Such architecture does not need a centralized registry
since any node will respond to the queries it receives. P2P architectures do
not have a single point of failure, such as a centralized registry.
Furthermore, each node contains its own indexing of the existing Web
services. Finally, nodes contact each other directly, so there are no delays
with the propagation of new information.


The reliability provided by the high connectivity of P2P systems comes with
performance costs and lack of guarantees of predicting the path of
propagation. Any node in the P2P network has to provide the resources needed
to guarantee query propagations and response routing, which in turn means
that most of the time the node acts as a relayer of information that may be
of no interest to the node itself. This results in inefficiencies and large
overhead especially as the nodes become more numerous and connectivity
increases.  Furthermore, there is no guarantee that a request will spread
across the entire network, therefore there is no guarantee to find the
providers of a service. Because of their respective advantages and
disadvantages, P2P systems and centralized registries strike different
trade-offs that make them appropriate in different situations. P2P systems
are more appropriate in dynamic environments such as ubiquitous computing,
while centralized registries may be more appropriate in static environments
where information does not change frequently.


Received on Thursday, 15 January 2004 16:08:33 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:41:10 UTC