Automating the Data Center: The First Steps Make All the Difference
The past few years have brought tremendous technological advances to the data center: open systems, client/server technology, distributed platforms, and Internet-centered computing. While these changes have allowed organizations to take advantage of unprecedented new business opportunities, they have also presented IT management responsible for running the data center environments with a host of new challenges. There are more systems and platforms to support, a steady flow of emerging applications to incorporate, new processes to manage, and fewer qualified technical professionals available to meet these needs. In addition, IT executives are facing escalating pressure to facilitate the sharing of back office information across all lines of business, divisions, and departments, putting a heretofore unseen emphasis on enterprise infrastructure. In short, competitive mandates are forcing businesses to stay online 24 hours a day, 7 days a week. To complicate this issue even further, enterprises are expanding portions of their operations into Internet-based co-location facilities, presenting the data center with the added challenge of managing high-availability production operations in physical computing environments that don't belong to them. The upshot is that in the New Economy enterprise, the IT infrastructure is online all the time, and the network is the data center.
In response to these forces, increasing numbers of organizations are exploring automation techniques so their data centers can provide more sophisticated levels of computing at lower costs. Performance and event monitoring, storage and network management, job scheduling, and software distribution are just some of the data center components that are becoming targets for automation. When properly defined, planned, and implemented, automation of these components offers savings in staff time, an increase in system reliability, a faster resolution of problems, and less downtime.
But notice the caveat: When properly defined, planned, and implemented. Many automation efforts stall because the scope of the project is unrealistic and the components targeted are not strategic or, often, because the specific terms and benefits of automation have not been clearly delineated. Additionally, inherent complexities, such as managing new data center processes and staffing issues, are not addressed during the planning and implementation of automation initiatives. However, when data center automation projects have well-defined goals and parameters, and when the implementation methodology is systematically laid out to meet those goals, companies of all sizes can achieve cost-effective, predictable IT operations and make significant strides toward the ultimate aim of a "lights out" data center (one that operates with virtually no manual intervention).
Define for Success
So where to begin? The first step for any organization considering automation for their data center model is to define and set the parameters. What does this mean? It means understanding what automation really means and how it applies to your specific situation. Defining and setting automation and benefit parameters will help companies get expectations in line with what's achievable, as well as help them identify the technical and production components on which they should focus to best ensure success for their individual business imperatives.
In the broadest sense, automation can be defined as those processes or components that can be monitored and provisioned without human intervention. If you really think about it, in the context of data center operations, automation encompasses little more than usage and capacity monitoring and provisioning to support the availability of your system, its resources, and the performance of your product or application.
When considering usage and capacity monitoring and provisioning, the following summary technical components must be accounted for: hardware, operating system, network components, I/O, database, supporting storage and batched processes (that is, database loads and so on). Non-technical components include those processes associated with change management and disaster recovery.
When analyzing automation opportunities, it's very important to consider the production processes that support the technical processes; ultimately, these production procedures control the technology and the dynamic changes that automation will bring. For example, if you're implementing technology that will ultimately effect automated provisioning of storage, but bypass altogether the change-management policies and protocols that the data center has in place, you'll have a major problem. This lack of planning for change management is one of the biggest stumbling blocks encountered by IT management and one that's critical to consider before implementing new technologies.
Once automation has been defined, companies investigating data center automation must identify the strategic components most suitable for automation, critically analyze the complexity of those automation processes, and then assess the potential results as realistically as possible. Doing this successfully is no easy proposition, as many data center directors can attest. But by understanding the true definition of automation, identifying the components most appropriate for automation, and assessing the complexity involved in automating those components, businesses can realize the returns they expect and implement a successful automation initiative.