The general manager of IBM Systems Storage explains a controversial boast and the way he sees the storage market shaking out.
Andy Monshaw, general manager of IBM Systems Storage, rankled a few journalists and analysts last summer when he called a press conference on a busy Monday morning and basically declared IBM "king of the world" in the data storage business.
His contention was based on a single marketing report by IDC's Dave Reinsel that melded both disk and tape hardware sales into one apples-and-oranges-type report, crowning Big Blue as the world's No. 1 data storage company. In the report, which lumped together total disk and tape storage hardware numbers for 2006, IDC said IBM earned 22.2 percent of combined revenue ($28.6 billion) over Hewlett-Packard's 20.9 percent and EMC's 13.2 percent.
IDC's report, which was commissioned by IBM, took in all spinning-disk servers with three or more drives, as IDC's methodology required, Reinsel said. However, the report didn't differentiate between internal and external disk storage--very different markets for IBM's various competitors.
It took a while for schedules to be connected, but eWEEK Senior Writer Chris Preimesberger--who wrote a column criticizing IBM's actions--
finally got the chance to sit down with Monshaw at eWEEK's San Francisco offices. Preimesberger and Monshaw discussed the report and its implications, as well as the storage issues that are top of mind for enterprise IT storage managers.
What were you thinking last June? The press conference seemed to be a lot of unnecessary crowing for a company as well-respected as IBM.
When we talk about our business, we try to define it in the way the clients view the business--not in a subsegment that we might be particularly strong in. In the conversations we're having with clients now, 99.9 percent are not about external disks. In fact, extremely rarely will a client want to talk about features and functions on a disk. They want to talk abut continuity solutions, security, availability, retention, policies, security management for an infrastructure--things like that.
That's why when we report our earnings, report our progress in the market, we talk in a broader sense, because that's what our portfolio carries. I think that was a point of view that perhaps you didn't share with us.
That was then. Let's move forward.
Will the storage market eventually belong to the specialists or the generalists?
First and foremost, everybody wants a piece of this space. The server guys are coming in one way, the switch guys want to come up, the software people want to come in across and the niche storage players want to go up. The simple, short answer is: It will go to the generalists of the information infrastructure capabilities.
I'll give you some examples. First of all, no argument: We will undercall the demand in the marketplace for the next 10 years. There is an uncapped demand; in fact, a lot of the demand we're calling isn't even showing up in the IT space--like digital video surveillance, which is largely managed in the security budgets. It doesn't show up in the infrastructures of the clients; it will, but not yet.
Here's the way I see this whole thing evolving: Every time there is a unique market concern, there is a rush [within the industry] with point solutions that, after implementation, clients are finding out, 'Well, that really wasn't the problem. The problem was broader.'
Take security, for example. You know, Wall Street loses some tapes; all of a sudden, you have all these different security encryption appliances. They start to get deployed. They don't scale, but more importantly than that, that wasn't the problem. Encrypting the data is industry standard; managing the keys is the problem.
I was with a client yesterday--a large telco from Central Europe. I said, 'Let's talk a bit about security; do you have a problem?' And they said, 'We not only have a problem, we have such a distributed security set of solutions, that we may end up in a position where I have to say to a client, 'I have your data, but I can't unlock it.'
This is a big, big problem. The approach from the generalist in the information infrastructure space is: Map the whole thing out. Look at not only how you start with the immediate concern, the encryption on tapes, but make sure that when you bring it to market, it's integrated with key management and with the software.
I could give you lots of examples. Solid-state drives will be the next one. People will espouse deploying this really cool technology; however, it doesn't last today in its deployment phase. It's about a system, not a technology. Solid-state drives are only as performance-fast as the slowest piece in the I/O path. What you don't want to do is deploy a technology that is more costly and won't produce the performance that clients are looking for.
There are a hundred of these examples. Storage management is another one. This hyper-growth in the market is creating complexity in the infrastructures. Complexity in the infrastructures is not about having 27 different management tools; it's about having one tool with open standards that manages everything.