What’s driving the creation of the next-generation data center?

Technology is fascinating, but to understand where it’s heading you need to understand the business drivers and economic trends behind it. For example, why are enterprises deploying next-generation technology such as virtualization and grid computing? The key business driver behind all of these technologies is the need to reduce operational costs – specifically labor, hardware and software licensing fees – and inefficient processes. To reduce costs, and increase efficiencies, enterprises are dramatically re-architecting and consolidating the data center. 

Specifically, IT execs want to:

* Reduce the cost of management and lower the total cost of ownership (TCO), by combining multiple servers, databases, storage subsystems, and physical infrastructure in a single location. Research shows that a single person can effectively manage up to roughly 35 “entities” (servers, storage subsystems, whatever). Consolidation helps businesses trying to reach that level of utilization.

* Increase the utilization of individual components. While most networks operate at 50% to 80% utilization, average utilization of servers, CPUs, storage subsystems, and the like is closer to 15% to 20%. Many enterprises buy a new server every time developers start work on a new app, regardless of the utilization of existing servers. By consolidating server farms into clusters, this becomes unnecessary.

* Simplify process complexity. For example, a company with 50 different data centers in 50 locations may have as many processes for data backup and recovery. By reducing the number of sites, organizations can optimize their processes to ensure best practices are deployed consistently and cost-effectively.

* Reduce physical infrastructure costs. Decreasing the number of data centers can reduce investments in power, HVAC, and real estate, all of which can comprise as much as 40% of the cost of a data center.

So far, so good. But consolidation also has downsides. It dramatically increases an organization’s reliance on the network, and can decrease application performance – a non-trivial issue given the large percentage of employees that work outside of headquarters. As noted in our upcoming research report  “Bandwidth and Network Optimization” (see http://www.nemertes.com) only 13% of users work at headquarters, where the data center often resides. Moving application servers away from users (by consolidating them centrally) magnifies WAN dependencies: Now that the server’s across the WAN instead of down the hall, WAN problems can derail application performance.

Moreover, next-generation data centers have a greatly increased need for redundancy. A centralized data center is a centralized point of (network and infrastructure) vulnerability.

Finally, and most importantly, next-generation data centers need next-generation architectures. In a word, that means “virtualization” at every layer:  routing and security storage, and computing. Next issue, we’ll discuss specific technologies that allow you to convert most of your data center resources into virtual resource pools. And we’ll show you how grid computing features prominently in your data center’s future.

Andreas M. Antonopoulos is principal research analyst at Nemertes Research. He can be reached at andreas@nemertes.com