Today, IT professionals have many choices on how to solve server/storage/networking requirements in the datacenter and at the edge. Hyperconverged appliances and software-defined storage solutions have emerged as an excellent alternative to traditional 3-tiered architecture and many have already deployed these solutions. If you deployed these solutions some time ago ago then you are now likely to be looking to replace or upgrade your software. Therefore, I thought it might be helpful to give a little background on the market and technologies available, as well as some questions that might help you navigate this important decision.
In this three-part blog I’ll start by addressing the two primary deployment models, identifying examples of solutions IT pros are likely to be migrating from and challenging the reader to think about their needs today and in the future. In the second part, I’ll offer some questions for you to consider as you’re trying to figure out which direction to go. And in part three, I’ll tell you about a great offer we have going on right now to help customers make the switch from their existing environment to a StorMagic solution.
Though still early in the adoption cycle, many people will be surprised to know that software-defined storage and hyperconverged infrastructure has been around for over 10 years now. The market has moved on significantly over those 10 years with end users having a broad choice today, typically segmented into 2 areas: build-your-own and pre-built appliances.
Pre-built appliances are an option for customers that prefer a simple “drop in” approach and don’t want to spend time configuring and testing their new hyperconverged solutions. If customers have the in-house capability or a trusted partner to build-their-own solution, they can typically lower their costs through the deployment of commodity servers and repurposing spare server and storage assets.
The build-your-own and pre-built approaches are equally valid – in fact both these options are available with StorMagic. I am not going to go into a detailed and complex analysis of each and will simply suggest you take the time to consider each one if you want to ensure you spend your organization’s hard-earned cash wisely.
Typically, refresh cycles for these types of systems are 3-5 years and many of you may be up against needed to make a platform decision for the replacement. In some cases, you may stick with your existing vendor(s) and do a like-for-like upgrade – taking advantage of the improvements in server/storage and software technology that has occurred. But, it’s also a good time to look around and see what else might be available. Additionally, some existing installed products in the market are no longer available and the end of support may hit in the middle of your next refresh cycle. Examples include VMware VSA and certain editions of HPE StoreVirtual 4000(formally LeftHand).)
Some of the key changes since the time that you originally deployed your existing hyperconverged or software-defined storage solution are the ever-increasing speed of processors and memory, increasing speed and capacity of storage and the ever-decreasing cost of both. Put simply, today you can do a great deal more with less. Today just two servers and the combination of magnetic disks and flash/memory caching can comfortably replace a configuration of six servers and high performance magnetic disks from 6 years ago, at a fraction of the cost.
Many organizations are not keeping up with the pace of change and at the point of refresh massively over-provision their compute and storage. It’s something like “hey, I’ve got a 4 server system now, surely I need this 3 node hyperconverged appliance cluster as a minimum”. They don’t carry out detailed analysis of their performance requirements, don’t consider whether they’ll actually use some of the high-end scale-out and management features and just go deploy. The reality is their performance requirements have typically only mildly increased by perhaps 25%, while the new system they purchase delivers 5-10x the performance of the previous system.
StorMagic has seen example after example of spending 2-4x on a system that has 10x the performance required. As a real-world example, we entered too late into a project at a 7 site law firm that spent $40,000 per site on a hyperconverged appliance cluster that could have been configured for less than $15,000 with StorMagic and two industry standard servers. We were too late to win the deal but it did bring home yet again how prevalent the issue of over-provisioning is.
Keep a lookout for part two of this blog series to get some questions to ask as you are assembling your true requirements for your upgrade project.