Laboratory for Internet and Innovative Technologies

On-premise server performance depends on several parameters. Server’s hardware resources, operating system (OS) and runtime environment are persistent during server and service life cycle; they provide constant performance for even server payload. This feature changes if the server migrates in a dynamic multi-tenant cloud. The server’s hardware resources usually are shared among several tenants which impacts server overall performance. In this paper we analyze what runtime environment and OS achieves the best performance for web services in cloud PaaS layer for peak loads, particularly when the increased load happens due to huge number of small messages or by huge message sizes. We propose a middleware strategy to survive the peak loads of huge number of small messages and also a model to transform huge messages into smaller chunks and send to the server as separated sub messages.


Sasko Ristov, Marjan Gusev, and Goran Velkoski


Cloud Computing, Operating System, Performance, Virtualization, Web Services, SOAP Messages

Full Paper

The technical report is published as a paper in Proceedings of the 9th Conf on Informatics and Information Technologies, (Editors Bakeva and Gjorgjevikj) CiiT 2012, Bitola, Macedonia, 2012