The State of the System i: The Analysts Speak

Reporting on an otherwise stellar quarter, IBM noted in its recent quarterly earnings report that System i sales were down 22 percent, quarter-to-quarter, from 2005. To be fair, it was a tough comparison, since the System i, bolstered by pent-up demand brought on with delivery on several major announcements, hit pay dirt in the prior year, with sales up 25 percent over 2004. That said, the negative 22 percent number left a few of us feeling more than a little disappointment, especially in light of the fact that the System z line was up 25 percent, the System p was up 10 percent, and even the System x was up 4 percent. A little envy seems warranted.

System i fans are a hardy and loyal bunch, and used to, if not immune to, the ups and downs (with more downs than ups) of platform revenue numbers in recent years. There’s little evidence that the roller coaster ride has had dramatic impact on the loyalty factor. On the other hand, System i bigots are still waiting hopefully for some sign that IBM’s message is getting through to the general IT public.

Is that likely to happen? Where do we stand and where are we going? Let’s take a look at what some of the analysts think.

On the subject of the third quarter plunge, Bob Djurdjevic, president of Annex Research comments that, “On the one hand, you can say 2005, which had been the turnaround year for the System i, was a tough year to compare to. On the other hand, if it was really a turnaround, it needs to go on–not just be a one-year thing. It seems that the momentum has stalled.”

The cause for this, says Djurdjevic, is that IBM has not been able to overcome the stigma of having the System i perceived as an old system in a world where X64 servers seem to dominate. “The System i just isn’t getting enough marketing buzz.”

But Djurdjevic does not fault Mark Shearer, general manager of the System i division, or anyone else in the System i management team. On the contrary, Djurdjevic believes that Shearer should get high marks for his accomplishments since he came on the scene a couple of years ago. He doesn’t believe the problem stems from Big Blue’s top management, either. “In fact, the problem may be below Shearer. I find in traveling to other countries and talking with people that IBM is not reaching down into the trenches. My impression is that the challenges that IBM faces are down below–not at the top. Not everybody knows where they want to march, and why.”

If this is true, some might still argue that this is a top management problem. But why quibble. It’s a problem in need of a solution, and fast.

Djurdjevic suggests that the answer will lie in IBM bringing a set of new “killer apps” to the platform. That, not just increased awareness through advertising, he says, would continue the momentum started in 2005. But he’s just not seeing those killer apps materialize. “As far as the new workloads go, the only thing new I’ve been hearing this year is this IP Telephony. It’s certainly good, but it’s going to take more than one of these kinds of applications to create the new buzz and excitement and to get new workloads onto the platform.”

IBM also needs to change the way it looks at the business, says Djurdjevic. Instead of viewing the business through the prism of multiple hardware stacks, he points out, IBM needs to recognize that workloads can be generated in a number of different ways. “Partnering with other parts of IBM like software and services to come up with packaged solutions would be another approach,” he says. “The customer may not even need to know what the hardware is because he’s buying a solution. Whether it’s the System i or the System p or the System x really doesn’t matter.”

Djurdjevic goes so far as to predict that “next year this time, we won’t be talking any more about the x and the p and the i. I think we’re going to be talking more about the business solutions and less about the platforms they run on. Over time, we probably are not going to be counting the revenues the same way either, because it will become impossible to do so. It will all be integrated inside the services or some other contract that IBM signs.”

Wayne Kernochen, senior analyst with Illuminata, recently heard the IBM execs explain the System i third quarter numbers, and accepted as reasonable their explanation that the past few years have seen several ups and down in platform revenue, and that we may be getting back to a long-term norm as opposed to the kind of results several major new announcements are made.

“On the other hand,” Kernochen adds, “we’ve been saying for the last decade that it would be nice if they got some new customers, and it may be an indication that they are not getting enough traction with new customers to get on a higher revenue rate.” Why? The answer is still visibility, according to Kernochen. “As far as I can see, the System i is nice and visible to the System i people, but it’s not visible to anyone else. So, when CIOs are asked what system to pick for a particular solution, they are not going to go for the System i.”

How can that problem, which we’ve all been talking about for years, be overcome? Kernochen suggests that IBM should be pushing harder to develop a stronger indirect market with ISVs who market to very tight verticals. As an example, he notes one ISV selling an ERP system designed just for car dealerships. Other companies that market portals and databases, for example, would be good candidates.

Databases? Yes, Kernochen says. “I follow several smaller database vendors, and the reason they survive in spite of the fact that in most large companies, Oracle has become the de facto standard. But these database companies say to the vertical ISVs, ‘Your customers don’t care what’s underneath as long as it runs and doesn’t cost too much, so we’re really a good platform to develop on and run on.'” He points out that companies that use this strategy have done better than the market overall and held off Oracle just fine by taking that approach. “I’ve got to believe that, at least for medium-sized companies, the System i should be able to do this. But I don’t get the impression that IBM has pulled in a lot of new vertical ISVs.”

Kernochen predicts that the fourth quarter numbers for the System i will be “sideways” and that, given the absence of any new major announcements on the table, the platform “will be running on its own momentum.”

What will happen in the coming year or two? Kernochen points out that “the System p is getting more and more like the System i. Now that hypervisor has gotten through everybody’s product line, what really are the differences? It’s not the chip set, and it’s not the hypervisor anymore.” He observes that a certain amount of the workload is now turning to Unix and Windows. The major difference with the System i is that the data management is integrated into i5/OS. Then the question is, how much are they using that any more, and can that be passed on to a Linux-type operating system? “While we have been focusing on other things, there is a creeping commonality with the System p, the robustness of which is getting up there. It’s still not near the System i, but it’s moving in that direction. At some point, IBM itself will look at this and say, ‘I think we’ll save money and merge the two lines.'”

Laura DiDio, research fellow with Yankee Group, says that her company’s research shows that the server market is, in general, “a little soft” right now, but that by the spring of 2007, sales will pick up. When markets are soft, she notes, large companies like IBM, Dell, Hewlett-Packard, and Fujitsu often decide to give sales a little boost by offering some incentives that might ordinarily be associated with a larger mainframe sale, so perhaps IBM should consider being more aggressive in that area. “They’ve been facing some still competition from HP, which has been doing very well.”

Because of the increased competition, right now it’s a buyer’s market, DiDio says, so each of the vendors has to do whatever they can to give incentives to customers to buy. Furthermore, she predicts that we will continue to see pricing pressure in the hardware market. “IBM and the big vendors have to find a way to present customers with a deal that’s very difficult for them to pass up.” She suggests even bundling extra services and customer support, with onsite consulting and/or training.

Perhaps most telling about what IBM could be doing to improve the marketing and visibility of the System i is DiDio’s report that the other server vendors are constantly bombarding industry analysts’ desks with announcements on hardware, software, applications, third party tools, and partnerships. “I can’t even absorb it all,” she says.

But not IBM. “If IBM is really serious about the System i, they really have to put a lot of marketing muscle behind it and turn up the volume. I just don’t see a lot of press on it, and I don’t know why that is.” She notes that Sun Microsystems, which has a lot more pricing pressure on it and a lot less money to spend on marketing, makes sure it is sending out the monthly analyst communiqués and newsletters. The same, she says, is true of HP, which has regular quarterly audio conferences to update analysts. “But I don’t see a lot of that from IBM. Why? You’d have to ask them. I would think they would be doing a lot in this area. I know what crosses my desk, and I know what their competitors are doing.”

The most optimistic of all is Aberdeen research director, Peter Kastner. “On the whole,” he says, “it’s full steam ahead.” To its critics, Kastner says, the System i is “the platform that won’t die.” Instead, System i users continue to invest in hardware, software, and new applications, while IBM continues to add new technology. “Hardware continues on a better, faster, smaller curve, and the biggest trend right now is multiple cores. IBM is one of the leaders in quad-core machines. The platform operating system and applications can easily take advantage of that technology, and I think we just go forward getting more cores per chip set.” [Editor’s note: this is precisely what I suggested a number of months ago that IBM do immediately to make the System i more competitive–an idea that IBM shot down. –TPM]

Kastner also believes that IBM’s support for service oriented architecture (SOA) will be a big boost for the System i, allowing customers to create composite applications that preserve the investment in the mission-critical, back-end, green screen applications while delivering the kind of Web services and user interfaces that modern applications can benefit from.

“Next year this time, we’ll be talking about SOA on the System i,” Kastner says. “It’s a journey that many System i shops are beginning to take.” And, in addition, information management will be a hot topic, he predicts. “Companies realize that as they create services in a SOA, they are going to have to do something about information management. What has been a pretty quiet information management space is not picking up as people realize that, in a service oriented architecture, you really need to think about where data is and when and how it’s duplicated and synchronized.”

In short, there is no real consensus among these analysts about the future of the System i platform, other than it’s not going away any time soon; nor are there agreed-upon pearls of wisdom to direct IBM’s strategy for the platform. But looking ahead to 2007, Kernochen’s thoughts may sum up best what most of us think: “The optimist in me says, ‘This time for sure.’ The pessimist in me says every year at this time, ‘Not this year.’ My real prediction is that the System i will be going on as before and will not see either a major uptick or a major downtick. But I live in hope.”