Shared Infrastructure allows you to pool resources across your data center to minimize costs, reduce complexity and shrink your footprint, all while being easily scalable and powerful. Given the budget pressures in the average IT department, you can’t afford to miss out on these advantages, but you also have to be careful to approach shared infrastructure the right way.—
The most important thing to understand is how shared infrastructure can reduce total cost of ownership. By pooling storage, cooling, power, and compute resources you can save nearly 20 percent across a traditional three-year hardware cycle. You do this in three ways:
- minimizing rack and data center space;
- cutting back on cooling costs; and
- reducing power costs.
Shared infrastructure pools resources that would once sit on separate racks into a confined space. Shared infrastructure can put 160% more nodes in the same space as traditional computing. Reclaiming floor space offers the obvious benefits of increasing the lifecycle of your datacenter building and freeing up space for other projects. Fewer racks can also ease management and, most importantly, feed into the other two advantages—reducing power and cooling costs.
With fewer racks, a smaller number of larger fans can be used to cool your servers. The increased heat from greater compute density is offset by the more efficient ways to cool a smaller space. Your building has fewer hot spots and targeted cooling efforts will allow for more efficient energy use.
Shared infrastructure also reduces the cost to power the servers themselves by creating effective resource utilization. One of the dangers of traditional architecture is over-provisioning resources. By concentrating on individual nodes and layers of your architecture, you can easily fall into the trap of over-provisioning or over spending on one or more layers of your architecture, as this video shows:
Shared infrastructure reduces management time and complexity
By cutting down on over-provisioning, you are also cutting down on power costs and racks, saving money all around. You also can save management time and complexity. Rather than spending your effort on individually optimizing each node in your architecture, shared infrastructure allows you to concentrate on business processes as they exist across all the layers. This gives much clearer insight into what your business is doing, keeps your staff from working on bloated or over-provisioned projects, and streamlines the type of equipment you’re putting to use. Rather than using different equipment from different vendors for every part of your architecture, you have common equipment with common management needs, power and cooling requirements, and service needs. This creates a much more predictable and streamlined management model.
Of course, with any change in architecture comes some common pitfalls or issues that you need to work through to get the right kind of savings and management benefits you’re looking for. And not every type of task is well-suited for shared architecture. When considering shared architecture you need to consider a couple of important issues:
- Whenever you are trying to stop over-provisioning of resources, it is possible you will under-provision instead. A solid shared architecture depends on a good understanding of the resources a process will be using to make sure you get the right mix of resources pooled together. Shared architecture is highly scalable, but if you get the mix of resources wrong you can lower the efficiency of the architecture.
- Big data and growing compliance issues are making storage within nodes a major factor. Making sure your storage needs are met in each node is vital to making the process work.
The nice thing about shared architecture is that the process of setting it up helps clarify many of the issues inside your business processes as well as your architecture. You’ll get a clear picture of your data center and it will reveal more ways to save money and run it efficiently. And no one can afford to miss that.
Tags: Data Center,Storage,Technology