What is our primary use case?
The primary use case is for virtualization of the Windows environment for our organization.
It has performed wonderfully. Over the course of the last 10 years, we have implemented vSphere Hypervisor and moved from five percent virtualization up to a current rate of about 85 percent, for our Windows environment.
The mission-critical apps we use it for are for production facilities, as well as optimizers for the machine equipment that is at those production facilities. There are ancillary systems in our corporate data centers that are used for the internal customer-facing apps, to work with the business intelligence piece, which can monitor metrics as well as capacity planning, ordering, and business warehousing. All of these business-critical functions run on vSphere Hypervisor.
How has it helped my organization?
We are able to increase the density of the virtualized servers and, with the increased density we have a lot of page sharing as well as memory sharing. We see performance increases from Server 2012 and forward; 2003 is debatable. There were negligible differences in 2012 but we did see benchmark performance improvement from utilizing Hypervisor and the increased density that comes with it.
What is most valuable?
The most valuable feature is its stability. There are a lot of product enhancements that come out regularly but, generally, the stability the solution provides is the most important to me, as I like to go home and sleep at night.
It is absolutely simple and efficient to manage. We can bring in people who have never been exposed to vSphere or virtualized environments and they're still able to support it from a server standpoint. The training time as well as the adoption rate, for a junior technician or somebody coming right out of college, is very good.
Sometimes, the talent pool is hard to fill so having that stability and ease of use is very important to us.
What needs improvement?
VMware has expanded, from a corporate standpoint, to where they have gotten very large. I have noticed, in the last couple of years, the breaking apart of specific added benefits and charging license upcharges for them. That would be the only negative thing that I have to say: As a large consumer of the Hypervisor, we have a hard time justifying the cost of utilizing the extra products, especially when it's a couple of grand here and there, a couple of hundred dollars here and there. It's hard for an IT administrator or an architect to sell to upper management. When they're seeing so much ROI from the Hypervisor, it's hard to show them that there is extra value in the additional products that can be tied on top.
I would really like to see an assessment of which products are actually going to be beneficial to charge for, and that they then continue to keep some of the products bundled in with the initial Hypervisor.
There are some competitive vendors out there who are sticking to the original model that VMware seemed to have, which includes a lot of additional features and functionality in the initial pricing, and I think they are gaining a lot of market share based on the fact that they are keeping their licensing simple. The only argument I have with VMware is that, when I ask our VMware team about a new solution, I hear comments like, "For a nominal fee we can upgrade your license and you can have that." For the large number of Hypervisors and the scale we have, it's frustrating to hear that I have to go ask for additional money for very small, additional features that I think should be included.
I respect that VMware has to grow and there are some features that they should not bundle in and that they should ask more money for. So I would like to see an analysis of sales and what's included and what the consumption rate is. I think they could dial it in a little bit better to where they have more bundled solutions.
Unfortunately, I think the type of model that VMware is moving toward is having an a la carte type of fee list. There are so many products that start with a "v" that I tend to get drowned with all the capabilities and I have to pick the particular thing I want to go after. Whereas, if there were more bundled services, or a package that included more bundled services, I might be able to swing that more easily than asking for money here and there.
What do I think about the scalability of the solution?
We're able to scale with density. I think that's the most important part. The clusters are allowed to go to so many nodes. We don't even touch the number of nodes per cluster. We traditionally have multiple fault zones in the data center, really for a comfort level, not because of a technological level. I know we could push the equipment a little bit harder but we generally like to keep things in a comfort zone that is constantly moving northward. So scalability is limitless and we have not really touched the capabilities yet, but we know the capabilities are there when we are ready to use them.
How was the initial setup?
The environment has changed hands several times over the years. Currently, I work to architect any new deployments but I was not involved in the initial bringing in of GSX, when the company first adopted virtualization, roughly 10 years ago. I have turned the environment over two or three times since I've been here. Now we have new staff in my group who are constantly evolving and changing with the adoption of new architecture and business cases for the Hypevisor and other products in the suite that complement it.
What was our ROI?
It's hard to calculate the ROI but I know that in our main, corporate data center we have gone from 700-plus Hewlett Packard servers down to fewer than 50 physical servers for the Hypervisor. We still have some legacy physicals that have not been virtualized yet but, over the course of this current refresh and into next year, those should go away.
In addition, in our paper mills and pulp mills we have heavily adopted virtualization, and in our box plants, where we make container boxes for shipments, we have seen a ratio of five servers down to one, and that's over a couple of hundred sites.
While an actual ROI number is hard to calculate, if you think about the yearly maintenance on all of those systems, it's very vast and deep. It also allows us the portability to expand rapidly and add virtual machines with virtually no overhead, once the initial architecture has been built.
What other advice do I have?
If you are not already virtualizing, existing-wise, you are doing yourself a severe disservice. Anybody who is continuing down the road of physical servers, any justifications that they think they have, should be challenged. If you have an environment that is all physical servers, a very easy win would be to present virtualization and denser workloads to your management. That would definitely make you look good in your career. I really don't see any negatives to moving to virtualization, even at a 100-percent adoption rate. We have yet to find a workload that is unable to run successfully in a virtualized manner, with the proper configurations and tuning.
We have not quite adopted vSphere 6.5 or 6.7. We do have some locations that have 6.5. On the radar will be utilizing the encryption capabilities, but as of yet, we have not really implemented that. We have a large organization so we move at a little bit of a slower pace. But implementing that is on the very near horizon, at least for our external-facing systems, as well as some internal.
We are also investigating the VMware Cloud on AWS initiative. That will probably be in the 2019 forum for dabbling or moving a percentage. With our being a manufacturing company, we move a little bit slower in adopting newer technologies and we have not really built the framework for a cloud initiative yet, but that will be something we investigate shortly.
I would definitely rate vSphere a 10. If you rate the Hypervisor alone, it's a 10. It has been one of the staples of technology for the last 15 years, and the key player for virtualization, for the whole industry during that time - or since Dell spun VMware off, or created the organization. It has been the premium, platinum product for Hypervisor. There are a few other players in the industry, but they are nipping at the heels, and that's about it. I do think that VMware is going to continue to lead, as far as Hypervisor goes, for the foreseeable future.