The Linux Business Case: Clusters, Part 2

By Tom Quinn

(Back to article)

One of the most common mistakes made by first time cluster buyers is to not take the physical demands of the cluster into account prior to procurement.

In one example, a company purchased a cluster from a "white-box" vendor who sold them the system as specified by the customer. After arrival, the customer realized he not only didn't have the air conditioning required to keep the system at a nominal operating temperature, but he also only had one-third the power needed just to turn the system on!

A cluster creates strains on your infrastructure you may not think of otherwise. The following provides a list of elements you should consider before bringing a cluster into your environment:

Power. Each cluster node can consume between 100-to-600 watts of power. Take that and multiply it by the number of nodes you are planning on including in the cluster. Add on margin for switching, storage, UPS systems, and you could end up with a system requiring tens of kilowatts on up!

For comparison, a typical wall outlet is rated for 1.8 kilowatts. In most cases, standard infrastructure wiring needs to be augmented to meet the needs of the cluster. Circuits need to be added, receptacles modified, additional wiring run, etc.

Cooling. Consumption of power generates heat -- the enemy of silicon. This heat needs to go somewhere or else the cluster will heat up, continue to generate more heat, and eventually start to fail based on components exceeding their thermal limits.

You should work carefully with your system provider to ensure your facilities provide adequate cooling capacity and air flow prior to making your decision to buy a system.

Floor space. Clusters can be big. A cluster rack takes up almost five square feet and can be over six feet tall. Make sure you have room in your lab or data center.

Weight. A full rack of cluster nodes is heavy. Up to 2000 pounds (but usually lighter). Many floors, particularly in older labs and buildings, were not designed to take that type of load. Check with your landlord. If you can't have a waterbed, you probably can't have a cluster.

Getting Started

The first step in your cluster migration is selecting the right cluster vendor. Work with a cluster vendor with proven cluster expertise in designing and building turnkey solutions.

A cluster vendor with industry-recognized success and training programs, for example, can mean better system reliability and can provide the user with the tools and skills needed to receive the highest return for investment possible.

But now that the thrill of the chase is over, it's time for the onerous task of cleaning your catch.

Turning the Key

Your system is being delivered to you and you, or your vendor, will need to install it at your facility. If everyone's done their homework, the system should go in smoothly and the hardware should work without too many challenges. Your team will start to get their hands dirty having to learn new tools, new concepts, and new configurations.

In many cases, the system will come preloaded with software and clustering middleware that can be used to verify proper integration. By using mature and comprehensive cluster management tools, administrators can be empowered over these complex systems, lowering the knowledge barriers to adopting the technology.

An organization's administrator costs can be greatly reduced when the appropriate management software and hardware are deployed. With proper training and after some trial and error, you will get your application up and running faster than ever before and start producing the results you want.

Tom Quinn is director of Government Business Development at Linux Networx, a provider of Linux-based cluster computing systems.