Utilization analysis of servers in a data centre

Authors:
Girish Keshav Palshikar;Amrit Lal Ahuja;Harrick M. Vin
Affiliations:
Tata Research Development and Design Centre (TRDDC), Tata Consultancy Services Limited, Pune, India;Tata Research Development and Design Centre (TRDDC), Tata Consultancy Services Limited, Pune, India;Tata Research Development and Design Centre (TRDDC), Tata Consultancy Services Limited, Pune, India
Venue:
ICDEM'10 Proceedings of the Second international conference on Data Engineering and Management
Year:
2010

Citing 8
Cited 0

Server Capacity Planning for Web Traffic Workload

IEEE Transactions on Knowledge and Data Engineering
On the Performance Regularity of Web Servers

World Wide Web
Understanding web server configuration issues

Software—Practice & Experience - Special issue: Web technologies
A server performance model for static Web workloads

ISPASS '00 Proceedings of the 2000 IEEE International Symposium on Performance Analysis of Systems and Software
An Analytical Approach to Performance Analysis of an Asynchronous Web Server

Simulation
Server Consolidation Using Performance Modeling

IT Professional
Profiling services for resource optimization and capacity planning in distributed systems

Cluster Computing
Web server performance analysis using histogram workload models

Computer Networks: The International Journal of Computer and Telecommunications Networking

Quantified Score

Hi-index	0.00

Visualization

Abstract

A data centre is a large centralized collection of IT infrastructure (servers, databases, application software etc.) for business functions. Analysis of the servers in a data centre with respect to their utilization characteristics is important for effective management the IT infrastructure - e.g., for reducing costs, improving reliability of business operations and capacity planning. Already saturated servers can crash with even a slight increase in workload and better usage of underutilized servers can save money and power. In this paper, we formalize several business questions related to utilization of servers in a data centre: already saturated, near saturation, under-utilized and interesting utilization patterns. We provide algorithms that use rigorous statistical techniques to automatically identify servers that satisfy such conditions. We present the results of utilization analysis of a real data centre.