Improvements to My Organization
When we have certain standards for performance, the customer experience is much better as well. They expect to have that kind of performance maintained or improved in the future. If there's a glitch, for example, whether it's storage or network, that's where customers start complaining about performance and the business goes haywire after that, for a while, until we fix the problem.
NetApp maintains the very high performance that we want to have. We work very closely with their engineer to make sure that every update they have will line up with what we require, or to fix whenever we have found problems in the past. We don't want to run into issues where, this is the price to upgrade to a certain version, and there is a certain impact.
I like the performance aspect of EF Series. It basically provides everything that we are looking for as a solution, very low latency and very high performance. That's why we're using NetApp to run our business. For example, with databases. We run a lot of Oracle databases that rely on performance because we want to have the business application respond within a certain amount of time for business transactions.
Room for Improvement
The EF Series has the web services outside the box, not inside. It doesn’t have the full OS, like Data ONTAP. You have to have a proxy web service that interfaces with all of the EF Series and you develop your application through that. I would like to have that interface inside the EF Series, so we don't have any dependence on the proxy service.
I’d like to see bigger, faster, better hardware, of course. I think that is the way the hardware is trending anyway; bigger, faster CPU, better software, fewer bugs, all that stuff. Those are things I trust NetApp to do. It's a company that's doing it right to get all the hardware and software to work together seamlessly.
Nonetheless, there are a couple of other things that NetApp hasn't done right. For example, NetApp still relies on the SAS bus of the backbone, so you still have the SAS controller. That’s a bottleneck for doing the lightning speed of flash. That's the limitation of flash. The industry right now is moving to NVMe. That stuff actually goes directly to the bus. It's actually faster. That's the first one.
As I’ve mentioned, a full OS like ONTAP for the EF Series is another one; actually have a shell and people could work directly on that. It's easier than using a proxy command through another machine. It's a limitation for us to work on.
NetApp is famous for redundancy, data protection, replication and so on, with ONTAP. I’m looking for a solution in the EF Series; a solution to mirror the storage off the chassis, off the data center, such as SnapVault or SnapVolumes. They don't have that yet. That's something I keep comparing with ONTAP because we were probably the biggest customer back in the old days with NetApp for Data ONTAP. I'd like to see those features come on over to EF Series as well.
As you know, NetApp is very famous for HA and failover, so stability is not a problem with FAS in the past and then now the EF Series as well. They always have hot dual controllers that we can rely on if there are some issues with the hardware; it can still fail over to get the business going. We don't have any problems with that. However, we also have another layer where we rely on technologies such as Oracle. Oracle has multiple technologies to keep the business going and rely less on the hardware redundancy. For example, Oracle has GoldenGate technology, Data Guard and so on, so that’s another layer that we're using.
Our solution is actually focused on breaking the pieces into a manageable vertical, where we focus on performance. When it begins to have a little bit of latency, we start to split it out. That's scalability on the application side. Infrastructure-wise, we design that way because we know NetApp can meet certain requirements that we have. Beyond that, maybe not, so we have to design our application to work around that.
Customer Service and Technical Support
I use technical support every day. They’re very proactive. AutoSupport is always the key when getting NetApp; basically, NetApp is famous for AutoSupport. If you buy a NetApp product without AutoSupport, it's not NetApp. We are always relying on the proactive support from NetApp. They call me even if I didn't get a notification from my customers. They call me ahead of time. They page us. They work with our on-call team directly, where they page to the on-call center and then we're like, "Oh, NetApp paged us. What's going on? Oh, because a controller is misbehaving.” Even if the customer has not noticed the performance problem, NetApp is already on top of it. That's what we like about the support. I think we have platinum support, which helps.
Nowadays, we're more like a data center, so we don't necessarily do the hands-on installation but we do have people on-site. We do have the knowledge transfer to those people that do the necessary installation, and then me and my team do the rest when we have a consultant. For example, I don't see the issues every day but we work on them every day.
Initial setup very complex, depending on the solution you're looking at. We're looking at databases, so we're looking at an InfiniBand fabric kind of SAN. Being able to get that kind of setup right the first time is always a challenge. People don’t plug cables in all the way and you have cable problems. You don't necessarily see that until you get everything up and running: "Oh, we’re missing a path." Stuff like that.
Other Solutions Considered
At the time, 7-8 years ago, we were looking at flash, because we have a need to reduce performance lag. There were many solutions back then, such as Violin and all the vendors that provide flash storage, IBM, TMS, Nimbus, and so on. We were evaluating a lot of companies at the time. At the same time, NetApp purchased the company that was running the EF Series. We evaluated everything that we had at the same time. We came up with the solution to continue with NetApp because of the support and the relationship that we had with NetApp, rather than invest our time to go with another company that we don't necessarily know what they're doing.
For example, Nimbus; we don’t know what they're doing. They’re just a startup. We didn’t know if they were going to stay in business or not. Violin might also be gone, and then all of the time you invest working with that company, is also gone.
We use the enterprise support model, where we can rely on almost like a partner. We know that NetApp is a stable company that we can rely on.
In general, when I look for a vendor, the important criteria that I basically look for are the roadmap of the products, support and the customer base. When you have a lot of people that trust the company, you know that you are choosing the right solution.
The EF Series has a lot of lines, a big lineup. Look at your application’s performance requirements. The EF Series is all about performance. Choose the right line of product because you can have so much performance but if your application does not need that much, you waste a lot of money, especially if it’s flash. You waste of lot of money.
Disclosure: I am a real user, and this review is based on my own experience and opinions.
Dec 06 2016