Data Center Capacity Planning

 

Name Data Center Capacity Planning
Modelers Kaveh Dianati
Contact kaveh.dianati.15@ucl.ac.uk
Client/Participant Telecomputing, Norway
Client Type Corporation
The Issue You Tackled

TeleComputing is a leading provider of centralized IT operation services, based in Norway www.telecomputing.com. The project developed a System Dynamics model that would enable continuing estimation of the point in time at which the processing capacity limit of the company’s data center in Oslo would be reached. The work was done in close cooperation with the Chief Technical Officer of the company, with regular in-depth meetings where model progress and results were presented and potential improvements discussed.

What You Actually Did

The model includes a relatively detailed representation of the development and growth in the company’s main assets, namely servers and storage spindles, and shows how the company’s growth objectives, together with its service level and inventory policies, lead to the historic and likely future growth patterns in its assets. Once the model was shown to replicate the observed behavior patterns for the right reasons, it was used to test future business development scenarios, and plan future investment in data-center capacity, valued at many millions of Norwegian Krone.

The Results

Previously, the likely future availability of processing power in the data center had been projected by using intuitive extrapolation. However, the model demonstrated that such extrapolations are completely misleading, because of a shift in the system’s behavior mode due to advances in technology. Consequently, the model fundamentally changed the company’s estimates concerning the project’s central question – when the limit of the data center capacity would be reached, and how much to invest, when, to ensure that capacity will continue to be adequate, but without excessive investment.Development of the System Dynamics model also discovered faulty data, and led to the devising of useful time-charted performance indicators, which resulted in otherwise unattainable insights.

 

Related Publications A System Dynamics Approach to Data Center Capacity Planning – A Case Study (Full Thesis) Download
A System Dynamics Approach to Data Center Capacity Planning – A Case Study (ISDC Conference Paper) Download
A System Dynamics Approach to Data Center Capacity Planning – A Case Study (Power Point Slides) Download

 

Did You Know?

Updates on the case

The authors contacted the CTO of the company for follow up three years after their final deliverable. The positive and valuable outcomes that emerged throughout the project, together with the ability to see their business performance in a visual, structured and rigorous manner, led the client to engage with the project enthusiastically from the start. As he remarked “I can confirm that the projections [from the model] are definitely true. We now know that virtualizing the majority of our servers means there’s no need to expand the data center for the foreseeable future, as we had previously expected we would need to do!

Although of substantial commercial value, this project was carried out as a requirement of the European Master in System Dynamics (EMSD) program at the University of Bergen, Norway, under the supervision of Professor Pål Davidsen.

Comments are closed.