The value of management in a commoditized world

PUBLISHER’S NOTE: Please note that, as of 9/29/03, all of your valued Network World Fusion newsletters will be delivered to you from nwfnews.com. If you use filters to manage your newsletters based on domain name, please adjust accordingly.

In the past, many hardware vendors viewed software simply as a way to “push more iron.” Although that’s still reality to some extent, the dynamics have changed.

When you evaluate hardware, the speeds-and-feeds discussions are still there, but in many cases the hardware has become more of a commodity. So while speeds and feeds are important, many hardware vendors are realizing that the differentiator for their hardware is better management of their hardware.

In the server market, vendors are bundling management software or embedding management capabilities in their servers. Examples of this are IBM Director and HP Insight Manager. Veritas and Intel recently announced that Veritas’ OpForce is pre-installed on Intel’s blade servers and Itanium servers.

The increased presence of specialized management appliances takes a different angle on this issue. In these cases, it is the management functionality that is being purchased, not the hardware itself. The hardware is important from the perspective of providing redundancy, continuity of service, and scalability. Beyond that, the brand of the hardware is not important, nor particularly discussed. The operating system is typically a Linux variant – which is also not necessarily discussed. What is important for appliances is that they perform a specialized management function, and that they do not take any existing system resources to run. There are appliances that monitor and manage servers, test loads, fight spam and more. The hardware is merely the mechanism for delivering the management service.

Interestingly, some of the newer trends like virtualization and grid computing also subscribe to this notion of commoditization.

Virtualization looks at computing resources, whether it’s storage or CPUs, as a pool of resources. The group of resources is treated as a single entity rather than as discrete components. It doesn’t necessarily matter that an application is run on server number 99567; virtualization will allocate a job to the systems that are in the best position to complete the job in an optimal amount of time. Much like we use commodities like sugar or flour, it usually doesn’t matter where the flour was milled, we just scoop it out of bag. In essence, the individual servers or storage devices are used as a commodity that must be managed. 

Grid computing also takes a similar view of infrastructure resources. The premise of grid computing is that resources could be comprised of excess capacity in many different systems and locales, potentially even in different domains. For CPU-intensive applications, such as life sciences applications, they can be run on a grid structure that uses whatever computing power can be applied to a problem. The focus is placed on the end result, rather than the means (which computers are used and where they are) of getting the end result.

So why are we seeing so much commoditization of infrastructure, even in the face of vendors trying to differentiate their products? As the industry moves toward service-oriented architectures and on-demand environments, this commoditization is a necessity. Delivering a service becomes the focus and priority, as opposed to keeping server 2345 up and running. The infrastructure becomes nothing more than a means of delivering that service.

On a final note: although commoditization occurs when looking at the service level, the discrete resource information must still be accessible to properly manage that commoditized infrastructure. The devil is in the details when it comes to management.

Commoditization is not something to be afraid of, as long as it’s properly managed.