Browse Prior Art Database

Redirection of services depending on service

IP.com Disclosure Number: IPCOM000033274D
Original Publication Date: 2004-Dec-03
Included in the Prior Art Database: 2004-Dec-03
Document File: 1 page(s) / 37K

Publishing Venue

IBM

Abstract

The goal of this article is to introduce a simple but effective solution to optimize performance of service execution in a client-service architecture, using different execution modes. An autonomic approach is proposed to execute a service in two or more alternative ways, based on the client's request, the current workload of the server and/or other environment parameters.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 52% of the total text.

Page 1 of 1

Redirection of services depending on service

In a client-service architecture, a server application can use information about its own state and the client's request to choose between different implementation flavours of the same service. The implemented flavours differ for the level of information provided to the calling clients. This is because the more the service is tailored to the needs of the client, the higher is the server processing cost and the lower is the client one. Conversely, if the service is too generic, and does not address the client's specific requirements, the server processing tends to be low while the client one may be high.

    Therefore, it is possible to define different implementations of the same service to move the processing from the server to the client depending on the server workload. If the workload of the server is high, the server can use a light solution based on algorithms that use less server resources, but with less "precision" in the service response. This means for instance that if the service is a request for an information update, the amount of information sent to the client may not be exactly tailored to the client's needs, and some information that the client already knows may be sent again. In this case the service elapsed time on the server is reduced, but the service is not tailored to the requesting client, which may need to apply a specific processing to adapt the results of the service to its special needs. On the other hand, if the workload of the server is low, the server can apply more processing to provide a service response that is completely tailored to the client needs, reducing the time required to process this response on the client itself. Workload thresholds can be used to define the service mode to be used when addressing a client request....