What is our primary use case?
Currently, we're doing migrations for older versions of Windows, both in the Azure Cloud and on-prem in our VMware vCenters. We use this tool to do comparisons between the current and future workloads and what would they look like, based on the usage. So, it is kind of a rightsizing exercise or rightsizing, either downsizing or upsizing, depending on the requirements. We just put all that information into Turbonomic, and it builds us out a new VM, exactly the size that we need, based on the trending and analysis. Then, you can also put in some factors, saying, "Look, it was Windows 2008, and we're going to windows 2019, or whatever. We're going to grow the database by X amount." This tool helps you do some of the analysis in order for you to get the right size right out-of-the-box. We love that.
I oversee a lot of stuff, so I don't really get an opportunity to go in there to point and click. We have people who do that.
It is doing Azure Cloud and VMware. Turbonomic understands the resource relationships at each of these layers and the risks to performance for each. You can compartmentalize your most critical workloads to make sure that they are getting the required resources so the business can continue to run, especially when we get hit by a lot of work at once.
How has it helped my organization?
The solution provides us with a single platform that manages the full application stack. In our decision to go with this solution, this was critical. We had so many vCenters and physical clusters out there. We had virtual and physical machines all over the place. Turbonomic was the way we were able to centralize all our vCenters and get a good picture of what is going on in the environment. It was all over the place without it, so there was no way that we could centralize and work on getting off of some of the older hardware platforms that we were on and start moving to converged, then eventually hyper-converged. This tool allowed us to take a good look at our environment and decide how we were going to size those workloads into those new areas, off of the old blade chassis and old standalone systems, to the more modern hyper-converged systems.
In our organization, it is optimizing application performance as a continuous process that is beyond human scale. The reason is because there are times of the year that we have these big hits. It is like if you were Verizon and you were to sell all your cell phones during the Christmas time. Well, we have a very similar thing here at our company, where we have a period of time we basically shut the business down. We have to give critical resources to critical applications, giving them the resources that they need in order to function. In order for us to do that, we are able to take critical workloads and put them off into their own area, then determine how much we have to take from the rest of the resources, which we take from the rest of the systems in order to put it into new clusters or systems. That is super critical for us every year.
We use it for management and rightsizing of our platforms, specifically for migration activities, because we're always doing it. The migration has been the biggest thing that I personally use.
What is most valuable?
There are a number of tools that we use in it. Some of the things that I request are the data dumps. They write some kind of scripts or something inside there where they are actually able to pull for me CSV files. Then, I can go in, take all that information, and build a master gold list for my migration activities.
Everything that I ask for, I get. I don't know what they are clicking nor do I know what they're doing, but when I request it, I get it. There are all sorts of different ideas and scenarios that I put forth to the developers.
Turbonomic provides specific actions that prevent resource starvation. While I'm not in there banging around on the tool all the time, I can tell you that I do very much benefit from it. On Monday, I was getting additional information from the Turbonomic guys.
We use the solution’s automation mode to continuously assure application performance by having the software manage resources in real-time.
What needs improvement?
There are a few things that we did notice. It does kind of seem to run away from itself a little bit. It does seem to have a mind of its own sometimes. It goes out there and just kind of goes crazy. There needs to be something that kind of throttles things back a little bit. I have personally seen where we've been working on things, then pulled servers out of the VMware cluster and found that Turbonomic was still trying to ship resources to and from that node. So, there has to be some kind of throttling or ability for it to not be so buggy in that area. Because we've pulled nodes out of a cluster into maintenance mode, then brought it back up, and it tried to put workloads on that outside of a cluster. There may be something that is available for this, but it seems very kludgy to me.
I would like an easier to use interface for somebody like me, who just goes in there and needs to run simple things. Maybe that exists, but I don't know about it. Also, maybe I should be a bit more trained on it instead of depending on someone else to do it on my behalf.
There are some things that probably could be made a little easier. I know that there is a lot of terminology in the application. Sometimes applications come up with their own weird terminology for things, and it seems to me that is what Turbonomic did.
For how long have I used the solution?
What do I think about the stability of the solution?
I have never had a problem with it. The product is a little over anxious at times.
What do I think about the scalability of the solution?
When we did the rollout in that phased approach, it was not difficult at all to roll in new technologies. They converged and hyper-converged into Turbonomic. So, it's definitely scalable. It moved right into the company pretty easily.
There are quite a few people using it, mostly for operations type of work. There are probably 25 users from operations, support, the performance team, and performance planning.
How are customer service and technical support?
I have worked personally with Turbonomic, one of their guys, on some of this stuff. I haven't talked to him in a while, but he helped us develop a lot. The support for Turbonomic is incredible.
Their technical support is excellent. By far, they are probably the best. It's probably why I am sitting here talking today, because I have to give these guys top props. I think the employee enthusiasm about this product is absolutely top-notch. It would probably be a great place to work.
I've worked with the Tier 1 support and their consultants. We had a consultant here for a year who was absolutely a top-notch fellow. He just became part of the team. He wanted to learn how we were doing things and tool the application to do what we needed it to do, which he did. He also left great instructions. A lot of his legacy is still there and being used today.
Which solution did I use previously and why did I switch?
We were using a combination of vROps and VMware. We were also using BMC TrueSight, which we still use today. There are a couple of others out there, because the network team uses a few things. There is all sorts of stuff that I think they were kind of hog-tying together to make them work.
Some of these solutions are ingrained in our processes and have been around literally forever. So, there isn't the staff or the resources right now to rewrite a lot of these things. Currently, we do kind of a side-by-side comparison, and I believe some folks have written some ways to integrate the new data from Turbonomic into the old way of doing things. That's just a culture change at the company. It's just a big place that has been around for a long time, which works slowly.
This solution was brought to us by one of our AVPs. She had worked at HPE, and we didn't know about it. She said, "Let's look at this," because apparently she used it at HPE. We looked at it, and said, "Ah, this is great." Then, we went with it.
Turbonomic is more customizable with a lot more features.
Even though you can turn on automation in VMware, it's not very good. It's kludgy and has a tendency to break things, where the autobalance of workload management that Turbonomic does within VMware is much better than the VMware tools which are designed for this. That may change, because VMware seems to be doing great with this. However, for right now, Turbonomic is the only way to go.
TrueSight is just straight up what you see is what you get.
How was the initial setup?
I went to the training when they first rolled it out, but I wasn't involved in the setup.
They did the setup in sections. So, they started off with the lower environments and some of the clusters out there that really needed a lot of attention, mostly blade servers and such. So, it was a gradual rollout. I think the entire rollout was somewhere in the area of a year to a year and a half. However, to get it fully running, where we could use this solution to our benefit, that was at least six months.
We use scheduling in real-time for implementing the solution’s actions as well as manual execution, which is when we schedule for a later date to change activities. We have had to enable or disable certain things. It seems to do that just fine.
What about the implementation team?
We used Turbonomic for everything to do with the setup. On our side, it required about five FTEs, who were engineering and operations personnel. There were folks who were creating the design and where it would be rolled out. That design was passed down to the operations folks who were actually implementing everything. So, it was done in phases.
We only have two engineers doing maintenance, a primary and a backup, and this is like their extracurricular activity.
What was our ROI?
The ROI would be in the return of hardware, specifically for a lot of the older hardware where we start to go into the converged systems. That is where we are seeing our ROI. We are getting rid of that old, junky hardware, starting to integrate and align things into one specific way of managing all our workloads, but not on old hardware. If anything, the ROI is end of life hardware elimination.
Also, we see ROI in extended support agreements (ESA) for old software. Migration activities seem to be where Turbonomic has really benefited us the most. It's one click and done. We have new machines ready to go with Turbonomic, which are properly sized instead of somebody sitting there with a spreadsheet and guessing. So, my return on investment would certainly be on currency, from a software and hardware perspective.
Turbonomic provides a proactive approach to avoiding performance degradation. Our capacity and performance team use this solution as part of other tools that they utilize.
The solution provides application-driven prioritization, with its AppDynamics integration, to show us how top business applications and transactions are performing. If Turbonomic comes back and tells us, "Hey, this application needs more resources. Or, you're coming up onto a period where it will need more resources. Start planning now." We have certainly used it for that and will continue to use it for that. We have actually used it for troubleshooting a couple of times, saving us 25 percent when it comes to performance-based issues.
We have seen a 25 percent reduction in tickets opened for application issues.
Turbonomic has definitely helped to save human resource time and cost involved in monitoring and optimizing our estate. It has automated a lot of things. We have saved 30 to 35 percent in human resource time and cost, which is pretty substantial. We don't have a big workforce here, so we have to use all the automation we can get.
Which other solutions did I evaluate?
We did an architectural review. We had to look at other options, but I don't know exactly what those were.
Some of the tools which already exist are not that great. They really need to up their game if they're going to keep up with something like Turbonomic.
What other advice do I have?
If you have a big shop, and it's scattered all over the place, then definitely take a look at this tool. Make sure you take a look at this tool because there is probably fit for purpose licensing for any size organization. It's a great automation process.
Turbonomic shows application metrics and estimates the impact of taking a suggested action based on its input from AppDynamics. So, we plug it into AppDynamics, then AppDynamics and Turbonomic seem to work together for that.
It knows what business-critical applications we have, but I don't think it manages anything specifically within the application itself. It is mostly just resource-driven.
As money starts to get tight and budgets start to get really scrutinized, I think people are going to have to start looking at using Turbonomic to help optimize cloud operations to reduce cloud costs.
We are going to continue to use it going forward. I just don't know at what level. There are a lot of changes being made to the infrastructure, so it's going to depend on the tools and things that become available, like VCF as well as all the products that they have built-in through vROps, enhanced vROps, and things that already come with the software.
I would rate it an eight (out of 10). Personally, there is a lot that it does that a regular person like me does not have the time to sit down and dig into it. We expect things to be a little bit more automated. That is why I gave it an eight. I would give it a 10 (out of 10) if I got in there and it's like, look, click, click, and click. However, I don't know if there is that kind of a comfort level here yet to just let this thing go and have its day with the place.
Which deployment model are you using for this solution?