Pavilion HyperParallel Flash Array Review

Proactive and responsive support, straightforward to set up, highly scalable in terms of performance and capacity


What is our primary use case?

Our use case is for internal company engineering operations and DevOps. Both our DevOps and engineering operations teams have a lot of compute and data-intensive activities associated with them.

How has it helped my organization?

Performance-wise, this product is faster than pretty much anything we've seen. In terms of the density and how it compares, what we have in-house is not very extensive in terms of other things we use, but in terms of our research and actually, what we have used, the density is much higher than anything else we've seen.

We can basically store the entire company's data inside of one unit, when the unit is properly configured. As it is now, it's equivalent to replacing three or four racks of equipment. The density is incredibly high.

This solution provides us with flexibility in our storage operations. It's software-defined storage, so we can allocate capacity however we want. It uses thin provisioning, which is convenient for us, and all sorts of other enterprise features that come with it that we haven't used quite yet. But, we can imagine we'll be taking advantage of them as the usage against the unit rises.

Our use case is primarily about performance, so consolidation has not saved us in terms of costs or capital expenditures. Our implementation of the product is an add-on to what's currently at the company. We've taken data out of the existing infrastructure and just moved it. The migration has allowed us to use it a lot faster, but we haven't gone through a consolidation exercise where we've gotten rid of the old equipment and now just depend on the new unit.

Absolutely, we are able to run more virtual machines on our existing infrastructure. With respect to storage management, we've reduced the amount of work that was required. In fact, we can eliminate most of the staff that has been dedicated to doing that in the old equipment. Now, we need very few people to administer the entire company using Pavilion. We can basically have one person manage all of the company's engineering data.

In terms of cost savings, in our situation, the cost we're saving is not headcount but rather, engineering time spent doing those kinds of activities. Where we may have had to spend a lot more time administering storage and IT equipment, we now have to spend much less time doing it, even though the headcount dedicated to IT is the same. Basically, opportunity costs have improved dramatically, as we've been able to assign staff to more value-added tasks.

We probably had three people spending between 25% and 50% of their time doing related activities, whereas now, we have one person spending perhaps 10% of their time.

What is most valuable?

We find it valuable that you can scale the capacity and the performance independently.

The high performance is very valuable, as well as the enterprise reliability features.

What needs improvement?

In our current configuration, we can only run the line controllers in high availability, active-standby mode, whereas we would like to see active-active implemented. That would get us more performance with a given number of line cards.

Their global namespace support is coming, and I believe it is based on NFS 4.1. We have a mix of both Linux and Windows usage in the company, and getting an NFS 4.1 client with Windows currently is difficult because I don't think that's supported. This is not an issue with the Pavilion product directly, versus more of the general environment. We would essentially like to see a Windows NFS 4.1 client supported so that we can take advantage of the Pavilion feature from both platforms.

Having a little more ease of use with the NFS global namespace vis-a-vis Windows would be an improvement.

For how long have I used the solution?

We installed the Pavilion Hyperparallel Flash Array fairly recently, about four months ago.

What do I think about the stability of the solution?

The reliability has been rock solid so far. We've only had it for four months so that's not quite a long runtime, but it's been perfect so far.

What do I think about the scalability of the solution?

This is a single product but you can configure the capacity or the performance independently. You can configure the capacity by adding more SSDs into the product, or you can configure the performance by adding more line cards.

This product allows us to start small and scale up, which is very important because we don't want to spend unneeded CapEx if there is no usage driving the extra capacity, beyond what we need.

If a product were implemented where the minimum configuration has much more capacity or performance than you need initially, you're committed to an outlay that you may not need for quite some time. It means that you make a larger CapEx purchase and then carry whatever OpEx licensing, year over year, for capacity and/or performance that you're not going to use for a while. In Pavilion's case, they allow you to start small and scale up, so we began with a minimal configuration that was suitable for our needs, and thereby we're not wasting or using funds until we actually need the additional capacity or performance.

Right now, we are still building out our engineering environment so it's being used lightly, but we have plans to use it quite heavily. In terms of scaling up capacity, that'll be a little bit hard to predict, but I could reasonably see us double the capacity. Perhaps in the worst case, quadruple it. With respect to increasing performance, we'll address it on an as-needed basis.

How are customer service and technical support?

The service support has been excellent. Whenever we have had an issue, or for example, we haven't had an issue but whenever we have accidentally misconfigured something, the service organization has automatically been notified of that. In these cases, they've gotten back to us within 20 minutes to inquire about what's going on. At that point, we've realized our mistake, and we've been very happy with that.

Support is incredibly proactive and responsive. The reason that they're responsive is that the unit is monitored by their service center.

Which solution did I use previously and why did I switch?

Prior to Pavilion, we had a solution that was implemented by our IT department, and it consisted of racks of hard drives distributed across multiple sites for backups and things like that. We switched because for our use case, we required a lot more performance, moving data from storage to the applications that needed it and back, because the applications that were using this data also generate a lot of data. The legacy solution was really not designed to do that.

How was the initial setup?

I did not personally set it up but I have somebody who reports to me, and they said it was very straightforward. We had a deployment plan, although it was quite simple.

Everything was running within one or two days. This included bringing the unit up and having everything configured. Most of the time was spent applying the software updates. Once that was complete, it came right up and we put it online.

In terms of stability and any issues that came up during the deployment, there were none. Everything came up as advertised.

What about the implementation team?

Our in-house team was responsible for the deployment. They were trained by Pavilion and the training was very straightforward in that they just use the unit.

What was our ROI?

It is too early in our build-out to judge whether we're getting ROI or not, but we fully anticipate that we will.

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

This solution provides us with DAS performance and SAN manageability at an affordable price, although in our use case, we were mostly concerned about performance. As such, we were a little less sensitive to price. That said, for the performance we get, it's a lower cost than anything else we've researched.

With respect to pricing, my advice is that you should have an understanding of what your OpEx goals are and analyze the Pavilion offering in terms of those, to make sure you can support the equipment year over year. Then, the initial outlay for the equipment is probably going to be very company-dependent.

There are no costs beyond the initial outlay for the equipment and the annual maintenance, which is a combination of the capacity and the line card that you have in the system.

Which other solutions did I evaluate?

We've evaluated other solutions but the shortcomings of the others were that they did not scale capacity and performance as easily as the Pavilion solution did. The competitors also used SSDs and NVMe over fabric.

What other advice do I have?

We are planning to use HyperOS 3.0 but we have not upgraded to it yet.

Our data ingestion rates have not been impacted one way or another, as it is not quite applicable to our use case.

This unit is so dense and the performance is so high that the advice I would give is that if performance is a critical factor for your use case, then you really ought to look at this unit. Performance and scalability are the two key things you really need to look at because scaling is incredibly easy to achieve. If that's important and you want to avoid a lot of costs in purchasing extra equipment and more networking as you scale up, and the same is true for capacity, then definitely give Pavilion a look because you really don't have to change anything to get that implemented.

I would rate this solution a nine out of ten.

Which deployment model are you using for this solution?

On-premises
**Disclosure: IT Central Station contacted the reviewer to collect the review and to validate authenticity. The reviewer was referred by the vendor, but the review is not subject to editing or approval by the vendor.
Learn what your peers think about Pavilion HyperParallel Flash Array. Get advice and tips from experienced pros sharing their opinions. Updated: September 2021.
536,114 professionals have used our research since 2012.
Add a Comment
ITCS user
Guest