HPE Synergy Review

We use it for public cloud right now. It covers a plethora of workloads. We have everything from the small, mom-and-pop, five-user accounting shop running QuickBooks, all the way up to data analytics running machine-learning on it.

The main improvement has been around being able to expand the environment without building out additional infrastructure. We have the blade chassis set up. We throw an extra blade in there and we've just expanded our compute and ram. It's made it much easier for us to grow the infrastructure, once we put the foundation in place.

In terms of managing our IT landscape, it helps by reducing complexity. That's really is what it comes down to. We've taken one of our virtualization guys and turned him into a Synergy guy, and he's been able to understand the networking, he's been able to understand the OneView, he's been able to understand the Synergy. He manages all of those rather than our having to have multiple teams associated with it.

The solution has helped us to implement new business requirements quickly. We deployed the 8180 processor from Intel with it, which helped us empower much faster compute. We have actually been able to pick up more of those machine-learning clients.

In addition, our IT infrastructure team is more efficient. We have fewer teams, fewer people associated with it. It has also affected the productivity of our development team, now. At the beginning, they didn't know how to utilize it, but now that they've been able to bridge into it a little bit more, it's sped them up.

For example, we run our own panel and customer-interface platforms which are built in-house. They bring stats from the compute node and they bring stats from the RAM utilization into a customer portal that is API-driven. That one was always very hard because all of the infrastructure that we'd stand up was a bunch of 1Us. When a new cluster would come in, they would have to write the code for that cluster and it would take them a week or two just to put the right code together and deliver it to the customer. Now they've written it for the whole Synergy platform, so we just slide a blade in and it comes online.

The solution has also decreased our deployment time. On average, we deploy a private cloud in about two weeks: Buying the nodes, racking and stacking, and then applying VMware to it. From there we, would install the OSs, etc. We've cut out the racking and stacking. We've cut out a lot of the physical deploy time. Now, we buy the blades and put them in an open slot. The networking is already set up, the infrastructure is already set up, so it's just VMware and the OSs, which cuts us down to about 60 percent of the time it used to take.

The most valuable feature is the control of the overall solution all in one box. With it being all together, it has really taken down the complexity of a multi-environment. Switches being top of rack, storage being separate. It's moved it all into one box.

There is room for improvement in the setup.

It's stable. It hasn't given us a problem yet. No outages.

It's very scalable. We like the idea that we can put four chassis in one of our racks, and we can connect up to 25 chassis, so the scalability to us, and being able to sync all those into one management portal, is unheard of. You can't really sync that many blades and chassis together in any other platform.

Technical support is good. We had a struggle at the beginning. What happened was that we reached out and they didn't really understand what we were trying to do. We got more of the leadership from HPE involved and said, "Hey, this is what we're trying to accomplish. We're trying to set up a solution." They put some more advanced technical support in place and were able to solve it in under 24 hours. That helped us in initially setting up the solution.

Our previous solution was 1U Dell EMC equipment. We were buying new equipment. we were also buying the 1U servers from HPE. But we needed a solution which had flexibility.

The initial setup was complex. We had a couple of bright engineers working on it, and they figured out a lot of things that they don't know. They don't always have experience around certain platforms. The problem was that they couldn't find documentation easily, to walk them through setup when they just didn't know the platform.

We used a reseller, TIG. Our experience with them was great. They facilitated getting us pretty much all the information we needed.

We have seen ROI. Building it to be so powerful, we had six racks of 1 to 2U servers of Dell EMC equipment spread out throughout the data center - six racks of gear. We were able to condense down into four blades, because they are so powerful. That's a huge ROI and savings for us.

It has also reduced our cost of operations. We have fewer teams focusing on the overall solution. We have a team of two right now, compared to the team of six or seven people assigned to the whole cloud solution in the past. From an operations perspective, we're probably saving a good $150,000 a year.

The solution has reduced our IT infrastructure costs, cutting out those racks. A rack costs me the square footage. Also, if I'm using a rack for my cloud, I can't sell it to a client. Moving six racks into one has saved me a ton of overhead on the infrastructure side.

In terms of TCO, I think we'll get there faster. While the blade is more expensive than a 1U-server - I'm going to buy a 1U server for $10,000 and it's going to support X number of clients - ultimately we get to the TCO faster because we're able to pack more into the density. That means our price per gig can be lower, but we make more margin on it, overall. We're able to get to that TCO faster because of the density that Synergy provides.

VMware is part of the cost. We bought the chassis, we bought the solution. The blades are roughly running us $60,000 a pop right now.

We vetted out everything under the sun from the UCS side, the Cisco side, the Dell EMC side. They had new chassis coming out, chassis which hadn't been released yet, which they were showing me. 

We really landed on Synergy because it was the most flexible out of all the solutions. We were able to get the power we were looking for in the chassis. Our data center is really built on density. We can run 52 kW in a rack, but with all the other platforms we were limited on the amount of power we could push in there. We liked Synergy because we could put four chassis in, running two 8280 processors, and running 1.5 to 3 terabytes of RAM per blade, making it a super powerful, 42 to 50 kW cabinet. We couldn't do that with any other platform.

Most of the time, people look at our solution - the way we built Synergy - and they say, "Man, we don't need that much power. You built it with the highest processors you can buy, the most RAM you can put in there." I tell them it doesn't really matter, from a Synergy perspective, if you need a smaller solution, less RAM. Synergy is a good play for them, even though they don't need the powerhouse, because Synergy is a platform is solid. They could get gold processors, lower processors, put 512 of RAM in it, and the TCO for them would still be as good as we're getting.

The biggest lesson we've learned from using the solution is that we should have gotten on it sooner. That is the main one. We weren't focusing heavily on our density before. Now we are. The thing we've learned is that for every new solution, whether it's storage or networking or whatever it is, we need to focus on the density side of things.

We did a big vetting of HPE's Pointnext services which helps move to the solution. We spent a ton of time on it because we really wanted to make it work. Unfortunately, since we're a service provider, the model didn't work for us, but we may use it for storage, so it could work out.

What does it not have? That's the harder question. Right now, it has everything we need. We don't really see anything, it doesn't have.

I'd rate Synergy a nine out of ten because, after it was done, it was almost perfect. It was getting it done, at the beginning, that was our biggest struggle.