Through good times and bad, scalability is one of the most essential challenges facing business and IT decision makers who must be able to meet the demands of their users regardless of market or economic conditions. Perhaps nowhere is that more apparent than the data center, which not only costs a lot of money to build but also to manage, support, and update. As such, there are several considerations for those who wish to have an efficient and high performance data center that will also scale both up and down depending on their business.
At this point, it really is important to understand the business case for data center scale. Inevitably, prolonged economic challenges lead many organizations to delay purchases on capital improvement of which data centers are certainly one of the greatest ongoing expenses. Many companies must now begin to replace their aging infrastructures, which will cost money, and so it is important to learn from the past by designing an infrastructure that is purpose-built for interoperability between different devices, resources, and protocols. Ensuring the highest level of interoperability is a good way to avoid a complete infrastructure overhaul (that is likely overdue because of economic conditions) and instead be able to replace only what is needed to maintain existing services or even offer more advanced services for the business.
In addition to interoperability, a scalable data center can also save money by allowing the business to avoid over-capacity by converging and consolidating infrastructure when business is contracting. Excess equipment still requires power, cooling, maintenance, etc. so less idle equipment equates directly to more savings. Of course, scaling up is equally important but this is where the advantages are more obvious. Simply put, a truly scalable data center infrastructure should be able to accommodate additional applications, services, and resources with only a minimum of additional hardware and often none at all. Sometimes, it isn’t even the business that is growing but the increased demand for new services such as mobile, social, video, etc. Again, adding these services to a converged (more on that in a moment) and scalable infrastructure shouldn’t require any much if anything in terms of additional hardware.
Scalability first requires a clear strategy for data center convergence and consolidation. Pooling server, storage, and network resources is essential to the concept of data center scalability because it allows IT departments more flexibility in the management and support of all IT services. In a converged infrastructure there are fewer machines providing more services which not only means fewer machines to manage but also less time to provision new services by simply leveraging existing resources. Converged infrastructures also make it easier to retire services that are no longer needed without the need to also retire any otherwise perfectly good hardware (often referred to as zombie servers since most of us are afraid to actually turn them off when they’ve been decommissioned for fear they may be doing something).
However, it is important to note that a converged infrastructure and scalable data center probably would not be possible were it not for the emergence of two important and still emerging technologies. Convergence and hence scalability are arguably dependent on virtualization, or the ability to use a single piece of hardware for many different purposes. Virtualization provides IT with the ability to scale based on existing capacity as opposed to adding new capacity. In other words, new applications and services are generally created through partitioning and provisioning as opposed to purchasing and deploying. As such, the process generally takes less time and costs less money. Fortunately, today’s leading data center solutions are specifically designed for virtualization which means it’s really just a matter choosing to develop the capability and ultimately take advantage.
Another technology that is exceedingly important to the concept of a scalable data centers is cloud computing. However, it must be said that cloud computing is not actually a part of the data center but rather a means of extending the data center by leveraging Internet-based applications and services. However, no all clouds are created equal. Today, it’s possible to “outsource” many traditional data center services to the cloud but it is also possible to replace a data center completely with so-called infrastructure-as-a-service (IaaS) which may in fact be the ultimate in scalability since it rarely if ever requires any capital expenditures (OpEx replaces CapEx) and the customer only pays for what they use based on the needs of the business.
So what’s it take to design a data center that will scale with your business? The simple truth is not much. Mostly it’s a matter of strategic thinking as well as a good understanding of the current technologies for convergence, virtualization, and cloud computing.