A Quantitative Approach to ERP Server Sizing

By ServerWatch Staff (Send Email)
Posted Mar 30, 2010


Under-sizing server platforms for an environment may lead to mid-life refresh, resulting significant disruption to business operations, added logistical complexity, and an increase in total cost of ownership. Intel IT developed a quantitative model to assist with server sizing by analyzing the effects of many different factors on enterprise resource planning (ERP) server utilization. Our analysis shows that the headroom advantage offered by four-socket servers can increase flexibility and help avoid mid-life refresh in situations where there is greater uncertainty in workload growth forecasts and larger demand peaks. This headroom is also important when targeting longer server refresh cycles of four years or more and when using active-active cluster designs to provide high availability.


"Server sizing and selection is a critical element of Intel IT’s enterprise resource planning (ERP) infrastructure strategy, which makes use of both two-socket and four-socket servers. Under-sizing servers may result in the need for mid-life server refresh, which can result in significant disruption to business operations, added logistical complexity, and increased total cost of ownership. "We developed a quantitative model to assist with server sizing by analyzing the effects of various factors on the utilization of the primary server resources: compute, memory, and I/O. The model is based on the Intel IT approach to sizing servers for our ERP environment." Server sizing and selection is a critical element of Intel IT enterprise resource planning (ERP) infrastructure strategy, which makes use of both two-socket and four-socket servers.

Read the Full Story at IDC

Page 1 of 1


Comment and Contribute

Your name/nickname

Your email

(Maximum characters: 1200). You have characters left.