IT often faces new challenges that can be stunning in size and impact. These typically come in the form of large-scale...
deployments or new projects that the infrastructure was never designed to handle. Virtualization eases the prospect of expanding to accommodate large-scale deployments, but you're still left with a dilemma: When it comes to your existing IT infrastructure components, should you start from scratch, upgrade or redesign?
Start completely from scratch
Scrapping everything and starting over would be ideal for many administrators, as this option is a complete undo for your data center. While virtualization has given us the virtual undo button, what we're talking about here is hardware. This might include your hosts, networking or storage frames. All of those pieces cost money, and it's unlikely you'll simply be able to trade in the equipment. Even if you could swap out old equipment for new hardware, that would involve some level of outage, which could be problematic. Rip and replace only works in very rare cases and often comes with its own set of challenges.
Upgrade or expand what you have
This leaves two options: Try to expand what you have or redesign around what you have. While these options might sound the same, they are very different, and it's important to understand the effects of your decision. Expansion is often the lower cost alternative and can be done quicker, but it comes with a few downsides.
One of the key factors to look at when you expand any part of your data center infrastructure is how far past its base you're looking to expand. Most IT infrastructure components are designed for some level of expansion, but they all have limits. Excessive expansion beyond the base limits of your hardware could compromise the entire infrastructure.
Even if you aren't constrained by hardware limits, you've just used up that additional cushion for growth. Now, when the next request comes along, even if it's small, you won't have the resources available to handle it.
Redesign existing infrastructure
A redesign means making changes to your IT infrastructure components and allocations to better meet your needs. Focus on where your primary pain points occur: storage, networking and compute. The key to success is not only understanding how a redesign will affect your existing infrastructure, but also focusing on the application at hand without losing sight of your overall data center vision.
Even if you only focus on one of those pain points, it's a complex undertaking. Storage, for example, is costly and affects multiple items in your data center. Storage redesign isn't simply an expansion; it might mean reallocating tiers of storage, including local storage options, to better fit your new applications. And all of those options can affect bandwidth and compute. Your application needs should be the driving factor in your decisions. This will help you come to different conclusions as you look to replace your hardware. You might want to move some of your existing production workloads to the newer equipment for easier expansion and connectivity with your application and move some lower-tier items to older equipment.
Virtualization gives you enough flexibility to move things with limited or no downtime, which better enables you to redesign, add and, in some cases, decommission IT infrastructure components as it best fits your application focus. Avoid getting caught up in the details of one aspect. For any expansion to succeed, you need to look at the overall data center picture and keep that in focus.
Protect your virtualization infrastructure
Modernize your data center with hyper-converged infrastructure and software-defined networking
Incorporate microservices and containers