Abstract
The evolution and increased use of business on the Internet has become a driving force in the deployment and management of Internet business applications. Consequently, many Application Service Providers (ASPs) have adopted a server-based model as a new Internet architecture designed to effectively deliver and control these Internet business applications. In this paper, we propose a speculative processing approach for the server-based model that increases the Quality-of-Service (QoS) of Internet applications that utilize this architecture. The Active Server-based Model (ASM) anticipates ASP web-application users' “click” actions to launch processing requests of the application during idle communication intervals as the user views the current screen. The ASP web-application predictor limits action predictions to a particular web-application domain, thus increasing its effectiveness and efficiency. Through a series of trace-driven simulations, we found a 23% increase in the Quality-of-Service associated with ASP web-application clients.