The MultiHttpServer : a parallel pull engine

Abstract

Internet agents require fast, concurrent access to many web pages. This service should be stable, central, and easy to access independently from the actual implementation language. This paper describes the MultiHttpServer (MHS), a parallel pull engine implemented in Java with a TCP/IP interface for communication with other programs. A second TCP/IP interface provides information for administration purposes. A simple config-file allows application-oriented tuning of the MultiHttpServer. An optional JAVA-Servlet can remotely start up the MultiHttpServer on demand. The focus of this report is on a user oriented description of functionality and usage of the system. Sample clients in several languages are discussed.