IBM's strategy to compete in the rapidly changing storage market centers on its Spectrum Storage software, flash...
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
and storage virtualization offerings.
Andy Walls, CTO and chief architect for flash systems at IBM, focused on those three product areas during a recent interview on the company's data storage plans and overall industry trends.
There's been so much disruption in the storage industry, including Dell's acquisition of EMC. What do you see as the pillars of IBM's strategy in the rapidly changing storage market?
Andy Walls: From an IBM perspective, our pillars are threefold. One, we believe that our flash offerings are disruptive. Flash is fundamental to the IBM story. I hear it from internal executives, as well as customers, all the time -- and it's not just performance, although that's certainly part of it. When you get into analytics and new workloads, being able to have consistent, smooth performance without having to spend a lot of resources to do it is just fundamental.
We believe that software-defined is more than just hype now. We hear even our large customers all the time saying they want to take advantage of commodity hardware, hardware that they have, especially for colder storage, and be able to take software and make storage out of that software. So, we believe our software-defined offerings -- our Spectrum [Storage] family -- are absolutely key and central to our storage story.
And thirdly, our storage virtualization [allows] clients to have the same look and feel, regardless of what storage they [use]. It's becoming so important to be able to have a virtualization layer where they have the same copy services [and] the same features. They don't have to train their support staff every time they buy new storage. With our Storwize family and the portfolio we have there, that's a key part of our story.
You talked about storage software running on commodity hardware. Did IBM make the right move in selling its x86 server business to Lenovo?
Walls: We're able to focus now with our Power and our z servers, and we work with partners with our storage offerings. But you do make a good point. [With] our software-defined, we are intent on working on commodity servers and working with all the different suppliers of commodity hardware. That is part of our strength, since we did have x86. We do understand all the different servers. We really believe that we are in a special position to be able to work with different suppliers.
Is it really possible to run storage software on commodity hardware without doing any integration or tuning?
Walls: Yeah, some integration. It's always, 'Well, what do you define as commodity hardware?' Does that mean you can just go pick up any server that you want and run our software? The answer to that would be no. You couldn't do anything, but we will work with the key providers of x86 servers, and have a support list and work with our clients. Oftentimes, the clients who are doing this are not just doing one or two servers, but dozens or hundreds, or, in some cases, even thousands. And we will work with the clients to see what their requirements are, and make sure that we can support those servers and feature functions.
How do you see IBM's Cleversafe object storage acquisition fitting into your overall strategy this year?
Walls: We're really excited about object. One of the predictions we made was that cold storage is getting hot. It's a very, very important part of our portfolio to offer object that can deal with the tremendous amount of archive storage and colder storage out there. I can't go too deep into their plans.
IBM flash storage for archives
Do you foresee flash being used for archival storage?
Walls: Your active data should really be on flash. Now, let me define active data. I was dealing with a healthcare provider, and they said, 'Your X-ray from five years ago, you might consider that archive.' But they consider it active data, because if you walk into the doctor's office, and he needs to pull up your X-ray and look at it, that needs to be available with reasonable access, so that he can make a diagnosis in a short amount of time and go to the next patient.
As active data is broadening, the amount of data that should be on flash is really increasing. Now, if we expand that and say, 'Well, what about the pictures that you took last summer that you haven't looked at and probably won't?' If you needed to get to them, an extra second or two would not hurt. I think there's still a place for colder storage media, for tape or slow spinning drives.
But I think we're going to see more data that should be on flash, that we might have called archive at one point. If you want to do any real-time analytics, it needs to be accessed quickly. We might consider it with low access rate, but when you need it, you need it quickly. So, flash is going to creep into places we might have traditionally called archive. The advantages with power, support and other things are so good that it's going to be economical for clients to do that.
What type of flash will be useful for archival storage?
Walls: I think there's going to be a place for 3D [triple-level cell], especially. We're going to continue to see consumer sorts of devices become enterprise-class, and they could be very useful for this kind of storage.
IBM doesn't yet support those flash technologies. Are they in IBM's flash storage future?
Walls: We are certainly looking at and working on advanced technologies.
Do you foresee the end of disk, or at least high-performance hard disk drives?
Walls: About 25 years ago, I had a director in IBM who said that tape was dead. Here we are 25 years later, [and] tape is still an important part of an archival strategy. I don't see the death of disk in the next five years, for sure. What you're going to see is slow and high-capacity disks. Part of the problem is that the capacity increases that we've enjoyed over the last three or four decades are going to slow down. And yet, it's still a very inexpensive technology when done with lots of capacity in a small footprint, as long as you don't need to access it much.
SoftLayer helps IBM find its place among cloud providers
Where does cloud fit into IBM's storage strategy?
Walls: If you look at IBM as a company, [cloud] truly is a pillar of [more than] storage. It is the future, and it underpins many of our strategic imperatives. It underpins our Watson. It underpins our analytics offerings, being able to get to the cloud, and use compute and storage. I often thought that if I were starting a small company, I would want to focus on getting the application engineers and software folks instead of having to build out my own infrastructure. With the cloud offerings that IBM has -- SoftLayer and others -- you now have such rich functionality, where you can just go to the cloud and get what you need very quickly. You can start building your applications and using things, like Bluemix, to build up what you need. It offers such a return on investment.
And storage is very much part of our strategy. We listen and work with SoftLayer very closely and make sure that we can meet their needs, so that clients don't have to necessarily just buy our next-generation flash or our next-generation block storage. If they want, they can go to the cloud and use it, along with the server offerings we have there.
Aren't Amazon, Microsoft and Google dominating the cloud discussion with customers?
Walls: I wish that the question could have been phrased: Why would anybody want to go to Amazon and Google when they could use IBM? But I know what you mean. There's a lot of press with them and others. We need to really be able to show the strength we have. We have a huge number of clients and many data centers around the world that work with SoftLayer. Our advantage is the ability to use proven technologies, like our flash, our storage, our Watson and other applications as well, bringing it all together to give clients experience from IBM and a breadth of offerings. We really do have a good value proposition over the companies that you mentioned.
Overview of IBM's spectrum of storage
Analysts predict drastic change in 2016 storage market
IBM makes room for hyper-convergence in its storage strategy