HPE GreenLake Flex Capacity Review

We're able to report back on the value that we're getting out of the hardware platform


What is our primary use case?

We use it for automation, CI/CD; for transforming our development environment from old, static VMs and even private cloud VMs, into a more deployed, purpose-built system so that we can deploy/tear down, deploy/tear down as needed for each project.

How has it helped my organization?

It gives us compute and storage. The biggest thing that it gives us is reportability; that's the critical component. By giving us a monthly bill that is expandable based on usage, we're able to report back to the organization on the value that we're actually getting out of the hardware platform.

The best way that it's providing benefit to how the organization functions is that it has cut the legs out from under the executives who are trying to push all development into R&D, increasing our costs. I believe it was five times more expensive to do that.

In the next four years over the lifespan of the first segment, it's going to be saving us approximately $1.2 million in just hardware purchases and software support costs. 

In addition, we're seeing about a 15 percent reduction in the time it takes to deploy IT projects. For example, when it comes to deploying to R&D, which is the purpose of the stack, by going with an all-SSD array, and with the automation which is built into the platform, we are able to get machines out to the developers at a much faster rate. We are able to destroy them and then re-deploy them to the developers again. It's significantly faster than on our old, spinning disk configuration.

In terms of the administrative aspects of our IT operations, GreenLake has definitely assisted in some ways because it comes with Datacenter Care Services. They're there fixing it before we even know something is broken.

What is most valuable?

It is significantly less expensive than going to cloud for development.

What needs improvement?

In the GreenLake portal, I would really like to see more granularity in the costing model. That's the one limitation that we do have: We have a lot of departments. We want to be able to break down, in a granular way, that this department is using X amount of resources, that department is using Y amount of resources. 

Currently, it just looks at the platform as a whole, from a hardware perspective only. I would like to see virtual machines. I want to know we're using this many virtual machines and they're using that percentage of the resources. I want to be able to tag it and say, "This is this department's fees, this is that department's fees." That way I could charge back internally to my own departments. I need to be able to cross-charge.

Just charging me, the customer, doesn't work well. It's almost never is it going to be a single customer who's actually using it. When it's IT, we're distributing it out to a lot of people, and we need to be able to pick out and show that, all the way back to the units. That way, we could really justify why they should give us an extra $10,000 a month.

For how long have I used the solution?

We've only had it deployed for three months.

What do I think about the stability of the solution?

For our particular configuration, it is very stable. The fact that they let us use the networking equipment of our choice and wrap it into the program is very helpful. That really is very important for stability when it comes to iSCSE communications for storage

But the Synergy frames themselves are significantly easier to manage than the c7000s and there are a lot fewer parts to break. We'd constantly have parts break in our c7000s. For stability purposes and for resiliency and reliability, it is much better. And the Nimble array that we're using has a lot of built-in fault tolerance, which is also very helpful.

What do I think about the scalability of the solution?

The scalability of the service is amazing. The fact that the Synergy frames can scale up so quickly and so easily, compared to a c7000 is great. We don't have to buy additional Virtual Connects, we don't need to keep on buying Onboard Administrators. That really helps with the scalability. 

I love the fact that, for the Nimble array with the GreenLake service specifically, there's no additional cost for them to upgrade the head units. If we max out the capacity of our Nimble storage array, if we need 20 terabytes more, they will just come out and swap the entire array heads, giving us the ability to add space and only charging us for that space, not for swapping out the heads.

How are customer service and technical support?

Technical support is fantastic.

If you previously used a different solution, which one did you use and why did you switch?

We were using c7000 chassis with VMware.

We were doing the same thing over and over again and it was costing us more money. We would let support contracts lapse and things would not be covered and then, of course, they would break. It would always be a rush. We looked at it and said, "This is a solution where you're always covered by support, there's no way of getting around that, so we can't have someone in management nickel and dime us out of support." We've got 5,000 to 6,000 VMs running and we have developers all over the world - over 1,200 of them who have to access these things 24 hours a day. Without having that support it was always killing us.

This service enabled us to make sure that we have support at all times, no matter what.

How was the initial setup?

The initial setup was very straightforward.

What about the implementation team?

We used an HPE partner. They did a great job.

What was our ROI?

We haven't seen ROI yet. That being said, we are already canceling some of our other software, and that will definitely show on our next fiscal year.

What's my experience with pricing, setup cost, and licensing?

Our cost is approximately $25,000 a month.

Which other solutions did I evaluate?

We also looked at Dell EMC for their pay-as-you-go model, but because of the density of our virtual machine deployments, running up to a 100 VMs on a single blade, we really couldn't go with Dell EMC. Their pay-as-you-go model is only for hyperconverged, and hyperconverged does not work well with the densities that we tend to run.

What other advice do I have?

If your management is pushing you to go all-cloud, then really look at this service as an alternative, because you're getting all of the OpEx savings that you would in the cloud play, but at a much lower price. We were comparing this to the cloud, to Azure, and the price for Azure would have been four times the price of this, and without the level of performance that we're getting out of this platform.

The service hasn't affected our organization's capacity management efforts and it hasn't eliminated the need for over-provisioning.

The biggest lesson we have learned from using the solution, the biggest thing, is that you have to make sure that you get it fully implemented before you let your DevOps guys touch it. If they touch it during the implementation process, they will mess it up.

I would rate GreenLake at eight out of ten. It is giving us a lot of what we need. The only issue I have is with the lack of some reporting features. If they were to get those reporting features in the portal, I would give it a ten.

Disclosure: I am a real user, and this review is based on my own experience and opinions.
Add a Comment
Guest
Sign Up with Email