- Graphical workstations
- Replacing the c7000 systems
- DevOps, using containers.
We bought it with the intent to replace the c7000 workstation blades. It is bigger than the workstation blades in core count, memory, and graphical capabilities. So, it has broadened us in that regard, and we have more capabilities.
It is very easy to reproduce. We have a bare metal kickstart and deployment process, which are very quick and easy. We haven't yet taken advantage of the Image Streamer capabilities of the product just yet.
It has affected the productivity of our development team in a bad way. When we first stood the hardware up 18 months ago, the image streaming capacity and capability were not very good at all. We had hoped that it would allow us to be more composable and be able to switch over from one version of an operating system to another version of an operating system. However, it wasn't ready for prime time yet. Therefore, we had to go back to a deployment of bare metal install. We are still waiting and trying to figure out how we can do the composable infrastructure.
We are all about graphics. I know that the Synergy has a sidecar on it, so innovation into graphics capabilities and more broadly used storage. Right now, we have 180 blades and 15 frames, but our only solution for storage is either onboard the blade itself or some network-based storage. We could probably bring some Fibre Channel into play. However, we would maybe like to see some innovation around the storage and those systems.
It still has some room to grow. For our solution, we need something between the c7000 and Synergy. Synergy is the high-end deployment, and we're still learning to how to do the composable infrastructure, so we can turn it around and make it look like this today and that tomorrow. The c7000 gave us a stable workstation remodel. We went from an Acura to a Maserati, needing something in between.
It is not very stable because of the image streaming and the issues that we had with it, which were more software-based. The hardware is very stable. The software which runs the hardware part is where we had our biggest problems with it.
It scales really well. We have 15 frames of Synergy with five racks. It has three frames per rack, so it scales really well.
On a scale of one to 10, the technical support is probably a six. It got better as we got deeper into the Tiered support. The initial support with its Tier 1 and 2 support weren't very useful or helpful at all. It was only when we engaged our Salesforce to help us get into the backdoor to talk to their Tier 3 team did we start to see things circulate.
It was communicated through to the sales team who supports us. We were looking for a replacement solution for the HPE c7000 workstation class blades, and this is what was recommended as a solution.
The initial setup was very complex. The cabling was such that it had to be done a certain way. It is one of those things that you want to not touch once it's stood up. That is our initial impression.
We did use Professional Services to help us deploy the Image Streamer and database back-end. They set it up and left. We had to reconfigure it and move it from rack to rack. That is when it broke, then we had to reengage those people again.
We have seen ROI. We have the ability to quickly compute. Instead of in the traditional way, one of the nice things that Synergy brings is that it is a big enough platform. We are putting a hypervisor on the system and carving off a portion of that to attach it to the graphics card, then the remainder of the CPU and memory are available for virtual compute.
The licensing is more around the software for RGS, because we are deploying bare metal installs. It is mainly the operating system and any lights-out management. So, licensing is minimum. We are licensing it annually.
When we made the purchase of the hardware, we added Professional Services to it.
It reduced our IT infrastructure cost by a little because of the amount of capacity and it is bigger. I only have 12 blades in Synergy, where in the c7000, I had 16 blades. I lost four systems for customers in the way that we deploy systems.
We like HPE because of the technology they bring to bear. Other competitors have blade systems, but they don't have graphics enablement. HPE also has a product called RGS, which we use very heavily. That partnership alone is huge. Thus, we have a huge investment in HPE.
Ask more questions to challenge the answers that the sales community gives. We went into the product thinking that we knew the product previously. That it just needed to be a maturation of the product, and it wasn't. I would poke more technical questions at them. Sometimes, with the bleeding edge, you have to be careful. We were one of the first adopters of it, and some of the bugs were still in it when we were ready for production work, so maybe it wasn't production ready.
It is a good, stable product. It will become more mature. Ask the questions of your sales team and the technical solution architect, ensuring that it's for you. Consider the total cost of ownership, and maybe starting out simpler.
It is touted to be composable. However, we haven't used it as composable. It is more of a replacement for an existing service right now.
The deployment time stayed about the same, because we deployed the same way.
It has added value to the process. It impacted us initially when we were deploying it. We were trying to troubleshoot the problem when we should have been under production, and we were about eight months behind deployment because of it. So, it impacted us in a negative way. With total cost of ownership, once we determine how to start using the composable infrastructure, it should add to that.