There is a lot of noise with our sales teams competing with a not-so-new style of computing. Hyper-converged is one of the new techno-buzz words our customers are considering with the expectation that it will be lower price and/or lower cost. Lets talk about both....
First, HDS offers several hyper-converged solutions. Some as containers/appliances and others that are more open and extensible. HDS does not shy away from the hyper-converged discussion.
Secondly, there are sweet spots and sour spots for appliances and some hyper-converged solutions. We cannot be seduced by a lower price when considering a new converged platform solution. Remember that only 20-25% of the total cost of a VM is in the initial purchase price (HW and SW). Many people are short-sighted and choose a lower cost of acquisition, only to find that long-term ownership costs are much higher, much higher than other alternatives that did not have the 'hyper' prefix.
So the opportunity is to work to understand the 'true and total' cost of a VM, using traditional DIY VM, converged systems, or hyper-converged systems. Beyond the purchase price are many other factors that impact total cost, including power, cooling, scale-up, scale-out, management, wasted resources etc. In doing these comparisons, there are some clear situations where hyper-converged solutions tend to provide a lower-cost of ownership (TCO) over time. These conditions are:
- Lots of small VMs (1-2 vCPU, small memory and disk). We tend to see that when a large percentage of the total VM count is this type, HC may be a good fit for these many smaller systems
- With lots of small VM, they should be static, not growing much related to memory or storage. The static nature makes it easier to fit them in containers
- In these small VM systems, predictable workloads, probably with organic growth
The real issue with HC solutions is related to un-correlated growth, and the resulting wasted resources that can emerge. That will be covered in part 2 of this topic. I have created a position paper on this topic, and will spend the next few blog entries putting some 'meat-on-the bones' of this outline.