Most IT professionals have heard the hype, and with vendors frequently announcing new hyper-converged products, it may be time to decide -- is it smart to hyper-converge your environment?
According to Ben Woo, managing director at Neuralytix, there are several essential areas those considering hyper-convergence should consider. "In some ways, it really is like putting all your eggs in one basket," Woo said. That's why thinking about future plans is extremely important. Considerations such as how security might be affected and whether the organization will utilize the cloud could affect both the decision to hyper-converge an environment and which hyper-converged platform might best suit the organization.
The hyper-converged appliance versus hyper-converged software approach is another consideration first-time buyers need to take as well. Vendors that provide software options ship no hardware, but supply the management and virtualization technology so that a user can pair it with servers, storage and networking of their choice. This means it's less expensive and more flexible, but also more complicated to implement. Hyper-converged appliances, on the other hand, package and ship their software with pretested servers, storage, networking and a hypervisor.
According to Woo, however, the software component is more important than the underlying hardware. He said choosing between a software or hardware hyper-converged configuration really comes down to: "Do you have the skill set to configure your own hardware, or [do] you feel more comfortable with just buying a complete solution?"
One sentiment that should put potential hyper-converged buyers' minds at ease about the technology is that it's poised for much greater adoption going forward. According to Woo, the next biggest development in the hyper-converged market will be acceptance. "I would say that if an organization doesn't have a significant amount of hyper-converged infrastructure or some form of converged infrastructure within their data center, they're missing out on the boat, and they're spending too much money," he said.
Transcript - Is it time to hyper-converge your environment?
What are the benefits and the drawbacks of a hardware configuration versus a software configuration for hyper-convergence?
Ben Woo: It's a common question. It's actually a misconception. Ultimately, the hardware is somewhat the same. They're classic, what we call common off-the-shelf, or COTS, components. It's a server with disks inside. It's got some networking. Everything else is software. So a hardware solution is a delivery model, not an infrastructure model. It just says that somebody, the vendor of the software, has decided this particular configuration is going to be best for their software-defined data center solution, as opposed to one element that says just buy a software solution, you can tailor a little bit of what the hardware looks like. Ultimately, they're the exact same thing. It comes down to: do you have the time or desire, and do you have the skill set to configure your own hardware or whether you feel more comfortable with just buying a complete solution?
Vendor lock-in is one concern, but are there any other potential pitfalls of hyper-convergence?
Woo: Yeah, you mentioned performance earlier. I mean, the ability to be discreet about performance really can be a challenge sometimes because everything is in the same footprint. You may end up in certain situations where you have an excess of capacity. All right, so you need three separate nodes to achieve the storage and data capacity you need, but you might only be using the equivalent of half a node in terms of compute. So there are cost trade-offs like everything else, but ultimately, because of the cost points available, the excess or the overprovision of the capacity generally doesn't have a major impact with an organization.
What would you say are the top three things to consider for people who are looking to hyper-converge their environment?
Woo: First thing is what does your infrastructure look like today? Are you ready to make that transformation? Do you have a comfort level about bringing all these components together? In some ways it is really putting all your eggs in one basket. At the same time, it's infrastructure. You don't like it, replace it.
The second thing to look at is security. Security is not an issue so long as you realize that everything is generally somewhat insecure and you take that perspective. Just because you move everything on a hyper-converged infrastructure doesn't change the security level that you have.
And the third thing to look for is, with your hyper-converged infrastructure, can you expand into the cloud? Because ultimately it doesn't matter whether you're the smallest company or the largest enterprise, there will be a hybrid conversion of your traditional data center. You're going to have some on premises because you feel like you paid for something and you want to hog it. And on the other hand, there are some things that need to be processed in the cloud, and you really don't care that it's the cloud because the data originally linked in the cloud, it's going to be delivered in the cloud, so why bring it in-house, process and then send it back out? So I would say those are the top three things to consider.
What do you think are going to be the biggest developments in hyper-convergence going forward?
Woo: I think the No. 1 thing about hyper-converged infrastructure isn't about hyper-converged infrastructure. It's about acceptance. So if you look at technology, there is interest, acceptance, deployment, we call that the IAD model. I think we've gone past interest, we understand they exist, but we're not quite through the acceptance phase yet, so there are some early deployments of hyper-converged infrastructure. To be honest, as a market analyst, I look at the fact that it really doesn't move the needle too much right now, but I think the next three to five years, as we move towards the end of the decade, I would say that if an organization doesn't have a significant, and again relative word, amount of hyper-converged infrastructure or some form of converged infrastructure within their data center, they're missing out on the boat, and they're spending too much money, and the net effect is that they will lose competitive advantage.