While attending VMworld 2017 I presented on some of the operational aspects of hyper-converged infrastructure. I believe the key takeaways were:
- Hyper-converged is more than just storage to gain the real benefits
- Hyper-converged has a difference scalability model (more linear)
- Hyper-converged requires a difference organizational structure to be successful
- Hyper-converged performance and availability are policy based instead of location based.
You can watch the video here:
One question I was asked after the presentation was about scalability. It was a really good question so I wanted to answer here. Let’s assume that you start with a 3 node cluster. After three years you add 12 nodes to a total of 15 nodes. At some point newer hardware types are available does that mean that you now need to buy 15 brand new nodes running into a major investment instead of incremental growth?
- The answer is yes and no. At some point you do have to make the three node investment but you should do it long before the end of life for your current cluster so you can organically grow new hardware before end of life. This should be taken into account on your growth models.
Thanks to all who attended.