As admins seek to streamline data center operations, many are pushing for a transition from heterogeneous to homogeneous...
environments. This can be done as part of any hardware refresh cycle, but the switch still requires planning, and admins need to map out how to migrate workflows so no important data is lost in the process.
Heterogeneous data centers contain hardware from a variety of different vendors, while homogeneous -- or converged -- data centers rely on a single vendor's hardware and often come in the form of a converged .
There are both pros and cons to transitioning to a converged data center. The most touted advantage is that basing the data center around a single vendor's products often means there is a single point of contact for support.
Taking the converged data center approach can also improve operational efficiency. With a single vendor offering, admins do not need to know all of the nuances of multiple hardware offerings or management tools. Instead, the data center hardware is at least semi-uniform and admins can manage it from a common tool.
It is also possible for an organization to benefit from the economies of scale by transitioning to a homogeneous data center. If the organization is purchasing all of its hardware from one vendor, organizations may also get bulk pricing discounts.
Of course, homogeneous data centers face the problem of vendor lock-in. If the entire data center is based on a single vendor's products, then admins rely on the vendor for security patches, software updates and hardware pricing, removing a certain level of management flexibility.
Another potential disadvantage of using a single vendor is that admins may have to wait for new technology if their vendor of choice is slow to market with innovative products.
Key considerations for a converged data center
Once an organization commits to the adoption of a converged data center architecture, admins must focus on preventing data loss during the migration. A complete transition to a homogeneous data center encompasses all the data center systems, including those used for backup and recovery. It makes sense to bring the new backup and recovery systems online first so that the workloads are immediately protected when they are transitioned to the new architecture.
Admins should support premigration data backups as part of the transition -- at least for a certain period of time. This means allotting time for an extended transition period where both the old and new backup systems remain online. Legacy backups should not be fully retired until they are no longer needed.
Additionally, the organization needs a plan for how to migrate each individual workload to the new hardware. If the current data center is heterogeneous, then there may not be a migration plan for every workload. It will therefore be necessary to come up with a plan that not only addresses migration methods for each workload, but also the order in which admins transfer the workflows.
It may be tempting to migrate the most resource-starved or mission-critical workloads first, but it actually makes more sense to migrate the least important workloads first. This way, admins can test their migration techniques and validate post-migration performance before moving the organization's most important workloads.