In his first 18 months as general manager of IBM's worldwide Linux team, Jim Stallings has seen dozens of corporate shops and governments around the world push Linux from the periphery to the heart of their IT strategies - never a bad thing for job security when you have the word 'Linux' in your title.
But with the open-source operating system now more firmly entrenched, the next challenge for Stallings and his team is to help Linux drive IBM's companywide On Demand initiative deeper into corporate accounts with one hand, while fighting the ongoing server battle against Windows with the other.
An 18-year veteran of IBM before he left the company in 1996, Stallings worked in several capacities with Big Blue including that of vice-president in charge of worldwide sales for the AS/400 division. Upon his return in 2002, Stallings served as a vice president in charge of IBM's integrated accounts division, and in January 2003, assumed his current position.
A graduate of the US Naval Academy and former Marine captain, Stallings took time to sit down with InfoWorld editor-at-large Ed Scannell to discuss what roads he sees Linux traveling both in the United States and around the world.
You have been running IBM's Linux operation for 18 months now. What have been the high points and low points along the way?
I suppose I have some in both categories. What I am most struck by, both in the US and outside, is the level of sophistication that governments and major users have around Linux. First, every one of them now has a Linux strategy in place, and they are mostly long-term strategies.
They have gone from edge-of-network solutions to where they now have dedicated application development machines, web services machines, and they are pushing right into the middle of the datacentre as fast as we can help them do that. I did not expect them to move at this speed 18 months ago.
But it is not just the speed of adoption, but the size of the businesses doing so. IDC and Gartner recently came out and said Linux-based server shipments are approaching $1bn a quarter in revenues.
That is largely on servers, what about the desktop?
Well I had made some predictions that it would mature at a much slower rate mostly because of the (lack) of applications. But with companies like Novell, which acquired SuSE and Ximian, and Red Hat now stepping up their (desktop) efforts, a lot of things are starting to play out faster than I thought they might.
What do you think is the value-add Linux offers in on demand environments?
If you look at some of the market projections for Linux out to say 2007 or 2008 on the server side, user choices for operating systems will be clearly either Windows or Linux. Windows will be the de facto standard but it is declining, and Linux is emerging as the fastest growing standard.
In 2007 there will be as many Windows server shipments as Linux server shipments in that year. But the difference is, one is an open operating system and the other is not. So where this matters to on demand is the ability to be much more interoperable. So it is not just about Windows or Linux, but about having an environment where users can exploit the best of the applications and workloads for either one of these environments.
So this support we have for Linux sets users up to really exploit a lot of this open framework that is gradually becoming a very big standard in the market. I think users will come to realise they have great flexibility with this open environment for things like different workloads and the fact they can consolidate them on different types of servers and not have to recompile or rewrite. That gives them a lot of flexibility and speed for on-demand strategies and deployments.
What is the next natural evolution for Linux? Is it just bigger, faster, stronger, or something not so obvious?
The bigger, faster, stronger piece for Linux is on a linear path. It is becoming very resilient and as robust as enterprise-class Unix. I see that continuing release after release.
So far Linux has been growing by taking market share mainly from Unix. Is that trend also continuing, or is it now taking more share from Windows?
Clearly, for the last year or more is has been a Unix-to-Linux migration, and it is also about moving primarily to Intel platforms. But that is not the only story. The other story, particularly in the enterprise side, has been consolidating on the mainframes and things like our pSeries and iSeries, which, too, are functioning more as consolidation platforms.
The other thing we are seeing is, over the past six months Microsoft said that by the end of 2004 it would not support the NT server. So a lot of the business partners like Novell came to us and said "Hey let's give them an alternative such as a Linux migration path." So we saw in the first quarter of this year some 60,000 of these migrations take place and it is still accelerating in the second quarter.
Since Novell acquired SuSE and Ximian, what sort of difference in the market have you seen? And has this placed any added pressure on your relationship with Red Hat?
We have had a really good relationship with SuSE and that has extended to Novell without any disruptions in that we have a lot of the same channel partners. And there are still a lot of NetWare machines supported by (IBM's) AS/400s. We obviously still have to compete in markets where we have some common products.
But what is different is, users who were wondering about these companies (SuSE and Red Hat) sustaining themselves financially don't have to think about that any more. It has given Linux this stamp of credibility and sustainability. Also, it gives users a channel to go acquire Linux solutions. Prior to the acquisition SuSE for the most part ran on our mainframes and non-Intel platforms, and Red Hat ran on Intel platforms. Now, both run across our entire server line.
In your recent travels to India, what sorts of things have you picked up on regarding their use of Linux and what we might learn from that in the US market?
In India what struck me was the enthusiasm with which the entire ISV community is pursuing the expansion of their knowledge about Linux and open standards, and how it could drive their future growth.
The Indian (Linux) market itself is about $6bn a year, but they are serving other markets, predominantly North America, that is worth many billions of dollars more in services. So they are thinking long term about the exportability of their applications and the support they can bring to these other markets using the infrastructure they have built.
So as in the United States, the services business there is growing rapidly. As rapidly as the applications business there?
Absolutely. And the reason is someone has to support all of this new stuff going in. Linux-based software is going to require all of the interdiction services, the maintenance support, as well as all the help desk stuff. One of the things we actually helped them with is the opening and design of their help center that supports 29 states plus the national government.
Are Indian-based companies learning how to run a profitable businesses selling and supporting Linux-based applications and services?
Absolutely. When we opened our Centre of Competency in Bangalore, India, we had there four mid-size ISVs, all of which were profitable. They sold into places like the banking and government markets and they all had their products running on a variety of different platforms. You have to remember that these companies typically sell their products into the government and commercial companies that are serving the external markets. We aren't talking much about deploying Linux-based desktops into the consumer base yet. We are talking about deploying Linux-based desktops that run help desks that are supporting major airlines and banks, who have outsourced that to India.
So how expensive is free software these days? As people start to integrate more Linux-based systems into complex heterogeneous environments, are service and support costs rapidly rising?
Users are doing more with Linux than they have ever done, so naturally the footprint for maintenance and support gets bigger. If you look at our services business, (Linux) is several hundred million worth of business and still growing. The services business grew 89% year to year. But I would say that on a per server basis, their users' prices are going down simply because, with larger footprints, you can discount more.
Are you seeing more interest in Linux-based laptops and other mobile devices among corporate users over the past year or two?
Yes, and that is one area that gets overlooked a lot. It is about 30 million systems with about 23 million desktops and seven million servers. And if you count the embedded Linux devices, it is in the neighborhood of one billion devices. So the area of the most explosive growth are mobile devices that have some sort of intelligent operating system sitting on a micro chip that is connected back to a network and/or server.
Are you working more closely with Microsoft these days to support Windows-Linux-Unix environments - or less so?
That is a very important agenda item for them. I mean the whole reason why Linux is becoming a choice for users is that they are saying this isn't about one or the other, it is about both. So our approach on this has been all about creating an open framework. You will see this in our server shipments through our Squadrons Initiative, where we are shipping our servers with OS/400, AIX, Linux, and Windows so you can run different classes of workloads. Users are making this a huge deal, and so it is something that Microsoft is just going to have to deal with.
Ed Scannell writes for InfoWorld