In-Depth

Multi-OS Hardware: The Next Step in Server Technology (Part 1 of 2)

Running multiple operating systems on a single server brings big benefits. In this first of a two-part discussion, we look at two of the three main benefits of multi-OS environments. Next week we'll

• By Vish Mulchand and Jim Lofink
• 08/12/2003

But this is just the start. Once a single server can support two operating systems, why not three or even four? Why can’t multiple operating systems be combined onto a single system?

This may be the technical trend of the future, but why would an organization want to buy a system that can support multiple operating systems? This week and next we'll examine the benefits of a multi-OS server, including:

• Reduced costs

• Optimal resource utilization and flexibility

• OS-independent application choices

When considering each of these benefits in detail, organizations in the future will ask why a server doesn’t support multiple operating systems.

This week we begin by looking at costs, resource utilization, and IT flexibility.

Benefit #1: Reduced Costs

One of the hottest trends in the data center these days is IT consolidation. In other words, meeting the need to “do more with less.” Companies are under pressure to reduce their IT expenditures. At the same time, from a business perspective, organizations are looking to IT to be more flexible and responsive, and to provide a competitive advantage. This has resulted in countless companies undertaking major consolidation projects. Through consolidation, organizations are looking to:

• Reduce costs (lower TCO)

• Enhance business agility

• Improve quality of service

• Reduce risks

Companies start consolidation projects by looking at their hard dollar costs. Typically, the first option considered is the reduction of the number of data centers and server consolidation. There are all kinds of success stories relating how companies saved millions of dollars per year while benefiting from improved performance and a simplified infrastructure. If consolidation can generate such great benefits, is there a way that it can be taken a step further?

The answer is: yes. By consolidating two different types of servers (e.g., UNIX and Windows) onto a common set of larger servers, an organization can reduce the number of servers in the infrastructure. This results in additional savings by further reducing hardware costs, facility expenditures (floor space, power, cooling, etc.), and administration expenses.

In addition, though you have two different operating environments (now on one server instead of two), IT organizations benefit from:

-- Simplification: It is easier to install additional cards and set up a new partition for a new application or test environment regardless of the OS than it is to bring in a new system.

-- Easier administration: Administrators now only have to be trained and know one hardware environment.

-- Support. Organizations only have one number to call (from a hardware standpoint) regardless of the OS instead of multiple points for different operating systems.

Benefit #2: Optimal Resource Utilization and Flexibility

One reason IT organizations consolidate servers is to optimize the utilization of their environment. Typically, when a company purchases servers, they buy extra capacity (such as more memory or processors) to accommodate peak demands (month end, year end, business hours, etc.). This results in a server running normally at 20 to 50 percent of capacity. As the number of servers increases, the extra capacity (unutilized resources) also increases. At any given point, every data center has underutilized servers.

While a data center needs to plan for peak loads and growth, IT organizations look at these extra resources as an unutilized investment. When money is tight, it is difficult to justify why these resources are not being used to their fullest. In addition, because these are often individual servers, it is difficult (if not impossible) to take idle resources from one application and apply them to a bottlenecked application. As a result, IT organizations want to consolidate applications onto larger servers, separated into their own environments through partitioning. This enables IT to reconfigure its environment to match business needs, optimizing the performance of applications and maximizing server utilization.

The same philosophical approach holds true for a multi-OS server environment. Why should flexibility be limited to individual operating system environments? An IT department must look at the resources in totality regardless of the operating system.

Furthermore, while it is important to address system demands at any given moment, it is critical to be able to rapidly respond to business changes. New market opportunities, cost-cutting initiatives, and mergers and acquisitions can alter the mix of applications and operating systems.

As the demands on given applications and the mix of applications change, so does the use of different operating systems. Because the operating systems were traditionally managed from a “silo” approach, new demands on a given operating system required additional hardware purchases. For example, in an application migration from UNIX to Windows using a traditional approach requires additional hardware for the Windows environment. The UNIX environment's additional resources (now that users have moved off the system) would be wasted.

Avoiding such waste is a key driver to a multi-OS server. Such a server could simply be reconfigured, so IT would simply move resources from the UNIX environment into the Windows environment. IT has the flexibility to adjust the infrastructure in response to changes in the business, and the agility to respond more quickly to such changes since it avoids the effort of purchasing and implementing new equipment.

More Benefits, Plus the State of the Art

Next week we'll explore how multi-OS hardware affects application choices, and we'll examine the current state of this technology from three leading hardware vendors.