NetApp AFF (All Flash FAS) Review

With SAP databases, there's significant performance improvement.

What is most valuable?

We decided to use the All-Flash because of speed. Most of the time, when we looked at the SAP database, what we found was, by using the All-Flash, we got almost 100% improvement on our jobs.

How has it helped my organization?

The best part about it is the density; otherwise, earlier, we used to use a lot of 300- or 600-GB disks. It saves space, saves power and makes us more efficient. The main thing is performance. If you can get the report done in half the time, it's good.

What needs improvement?

I would like to see the All-Flash FAS support virtualization better. I find that lacking in some areas; application and for disaster recovery. I know we have to do a lot of setup and we need to know exactly what needs to be done, but I would expect NetApp to make those best practices available automatically. Why do they say, “Do this, do this,” when they could say instead, “For DR, click this button”, which would automatically implement the best procedure, rather than having to figure it out yourself? That should be automated.

There are several other improvements that can be done, especially with the clustering. I don't know why we had to make back-end decisions. With software-defined networking, most of the decisions can be made at the front end. Right now, how NetApp works is, you get the data to the head, take it to the back end to make a decision and then pump it back. I just want to eliminate the switch in the back to the cluster. Why not make those decisions? Maybe they need to do something on the software-defined networking; maybe have some module in the switch to make the decision at the front-end, distribute the workload for the clusters in the back. I really don't like having another switch in the back. You know your data comes from this network.

What do I think about the stability of the solution?

So far, we have not had any major stability issues because I look for stability, then performance; the product has to be stable first, then comes the performance.

My uptime is 99.99%. Other people say “All five nines,” but I say, “Hey, when the CFO or the CEO wants access and it's down, it doesn't matter what you're doing.”

Stability is very, very important. The first thing is stability, then performance. Performance is important because performance is everyday work. Stability is like, you say nowadays, “IT infrastructure has to be like air. You don't look for air, right?” You can automatically breathe it like that. Storage has to exist all the time. That's the main criteria on stability.

What do I think about the scalability of the solution?

So far, I don't know the exact size that we have. I know we can add more storage. We just procured some more disk shelves to add. I don't know the limits. I probably need to go check out how large we can be.

Also, we're trying to keep our environment separated. That way, there's no contamination. There are also regulations and other things we have to worry about. If we're putting everything in one box, putting all the eggs in one basket, we need to be really careful about stability, performance, and making changes.

If we want to scale out in the future, I think the system is capable. We should not have problems; I hope that will happen.

How is customer service and technical support?

We might have used technical support a little bit but most of the time, it is working, so I don't think we made any calls. I don't think we are using it. We're paying for it but we're not using it much.

Our vendor was good, they did the initial setup; they helped through the setup. If you set it up right the first time, you probably don't have to mess with it a lot. If it is stable, there isn’t much else to do.

Which solutions did we use previously?

We were previously using NetApp with spinning drives, and we were also using some of the EMC DMX.

Now, we are using NetApp exclusively.

How was the initial setup?

Initial setup was pretty easy. I think it only took maybe half a day to do everything; put it in, power it, connect all the cables, configure it. I think we put it in production within like half a day; not difficult.

Which other solutions did I evaluate?

We did run the eval and our PoC through other vendors, other storage suppliers.

There were two other flash players, and we finally ended up going with NetApp All-Flash. The reason being the migration would be much easier. We added our existing cluster to the same cluster, so that we could do the migration whenever we are able to do it. We didn’t need a big downtime to migrate it.

Also, when we buy other technology, we have to have people to manage it. We need to decide whether, “OK, do I need to use the current talent pool to migrate to All-Flash, or bring in a new player where we have to support both?” It adds to the cost.

When we are selecting a vendor to work with, we look at whether they want to work according to our interest or according to the vendor’s interest, because we need to make sure they can support us in the long run; that they are reliable; and that they have good people who know the product and have a good attitude working with customers. Most of the technical knowledge and other things, you can acquire, but attitude is important.

What other advice do I have?

If you are a NetApp customer and considering a new technology, you need to look at the additional cost of doing things or administrating another thing. If you are completely moving from NetApp to a new vendor altogether, can they do everything? Transitioning from one storage to another takes a long time. At the end of the day, your servers and other things, they don't have anything there, like transient, that you can replace any time. But when it comes to storage, your storage is important.

If you give me the storage, I can do pretty much everything. If your data is available, you can figure out how to reroute it or do things with that, but if your data is not there, you have servers, everything is useless; network. Everything is useless. I still see people invest a lot of money on networking. I say, “Look, if the storage is not available, you don't need network; you don't need servers.” You need to look at your storage; it’s very critical. It has to be stable, perform well and you need to be able to protect it. If those things are there, you can take the storage anywhere and make it work. If you don't have compute, Amazon EC2 can give you compute, Azure can give you compute, but you need to protect your storage.

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