Although the IBM Software Group was officially announced on Jan. 10, 1995, it was actually put in place in 1993. Now, 10 years later, it is the leading software developer in the world. eWEEK Senior Writer Darryl K. Taft sat down with Steve Mills, senior vice president and group executive of the IBM Software Group, to get his take on the groups origins and future—particularly given its latest acquisition of Rational Software Corp. The interview spanned two conversations, including one at the companys recent developerWorks Live! Conference in New Orleans.
Where did the idea for an IBM software group come from?
The idea of creating a true software business within IBM predates the creation of the Software Group. We began to build out our portfolio in the early 90s when Janet Perna went off to build the DB2 workstation offering in 1991. We delivered in 1993. We delivered products like MQ and what had been our VisualAge tool set in the 1992, 1993 time frame. So there were some activities that had taken place prior to the creation of the software group, which occurred in 1995.
I think one of the important things that occurred was that Lou Gerstner came into the company, and although not a technologist, it was fairly obvious to him that if you wanted to have a software business you were going to have to have a lot of specialization in the marketplace beyond your developers to be successful. It was sort of an applied technology. You just dont buy it and plug it in; youve got to put it to use. It requires skill and expertise to sell it and implement it. So there was a real need for field structure in support of a growing software business. And Lou saw the economic benefits of a growing software business, so he bought into the idea of having one. He pushed the sales leaders in the company to begin to build up a software sales capability in 1994. That didnt go very quickly. He then tapped John Thompson to help form the Software Group in 1995.
So in 1995 we then collected up the field sales structure and we made it a part of the software group and we built a sales team that the software group managed and created the sales plans and incentives and career paths and everything else for. That began in 1995. And today we have 14,000 people in customer-facing field jobs. And that includes the classical sales reps, the technical support specialists, the IT architects, telesales and 2,500 laboratory services people that are part of my organization doing implementation work. So its the largest direct customer-facing sales team in the world that is dedicated entirely to software. Continuously weve built out the portfolio through combinations of build from within as well as acquisitions that weve made. So its taken quite a long time. In the early days it was getting the foundation in place.
How do you think your message has resonated with developers?
The developers just dont show up, they come to you because they see the market is buying your products, that your products are becoming popular and that youre a market maker. The core runtime offerings—WebSphere, DB2, MQ, Notes—are critical to build a large base around. Then you attract developers to that base. Developers are attracted to something. The something is not a tool; its a business thing. They choose the tools that match the nature of the application youre trying to build and the application youre trying to build it for.
The popularity of Linux is helping us greatly. Linux is taking the place of Solaris as the popular operating system in universities.
What do you make of Microsoft backing off on its .Net naming and branding scheme?
Its kind of obvious that they came up with this sort of architecture brand name. You cant buy .Net. Its kind of an umbrella thought. And whenever you launch those kinds of things, youre stuck with making it mean something. Customers want to know, “What do I buy?” And I think they struggled with the problem of what to tell customers to buy. They were calling everything such and such .Net, like SQL Server .Net, Windows Server .Net, which has been changed. So I think they struggled with some brand imagery and concluded that it was getting in the way of clear communication about what the customer should buy today. You dont want to launch one of these architecture brand names and have people wait years to get the products that conform to the architecture. And Microsoft is still a long way away from delivering all of the .Net services that make up the .Net vision.
Do you plan to make hay of it?
Its easy to make hay out of it. For customers who say theyre going to do .Net, we just ask them, “What are you going to do? Whats the application? What problem are you trying to solve? Youre not trying to be .Net. Microsofts trying to be .Net; youre trying to be a bank.” Then you remind them that Microsofts still delivering the pieces, they have provided all the functions, youre still in a hybrid environment, you want to build anything to any scale on the server and communicate with any servers you use the old COM services and the old tool set. You have to mix these things back and forth, and you look at the road map and see that its not complete.
Its not a hard play; they kind of set it up for us. They throw you a softball, and you hit back at them.
Well, many say the same can be said for WebSphere—that its a complex brand made up of hundreds of disparate products.
There are quite a few WebSphere products, but they all contain WebSphere. And were having great success. WebSphere has been growing double-digit in the market. So they can be critical, but were doing just fine.
And E-business On Demand?
E-business was a theme that we launched in the 1996 time frame. And we see On Demand as a theme—a label on an era and a set of business characteristics, its not a label on technology. And just as we have pursued the e-business theme for many years we think on-demand has long legs to it. And as always you continue to calibrate the market. And a lot of us were trying to understand how you describe what the customers are trying to do. If you can put a face on it you can cause them to identify with you more. And thats the reason for the name.
You see Microsoft as possibly your most fierce competitor, yet you are in lockstep with them on the various Web services standards and organizations. How can you maintain that balance of competition and cooperation?
Both companies were seeking something in the Web services effort. From our perspective we were anxious to get a set of open interfaces created that enable any client to communicate with any server. And whats possible today is that a Microsoft Office or Visual Basic client can interoperate very easily with a WebSphere server. You can populate WebSphere from Office and WebSphere can populate Office and do so seamlessly. Web services standards have been quite powerful in allowing us to move to this open client structure, which is one of the things we were seeking in the Web services effort. Microsoft has, over the years, rejected in participating in interoperability standards. They joined the original OMG CORBA effort, and then they left it. They wanted to do it themselves. They got involved with Java in the beginning, and then they left it. They wanted to do it themselves. So theyve participated in standards activities in the past but have often then left the process and therefore never deeply incorporated any interoperability standards in their own product set. And XML Web services is a way for them to get connection to other things
Now, the use of XML is going to happen anyway. Theres nothing more open than XML. Its truly declarative in nature; its not programmatic in nature. It would have been accessible to them anyway, so we dont feel were sharing some deep dark secrets with them. And the potential to have better interoperability from client to server and server to server was the benefit we received. And you cant deny their participation in the marketplace, so our approach to this was very pragmatic. And our view is that weve derived a lot of benefit from it. Were better off going to customers with a proposal that can link to their Microsoft Windows environment, rather than to isolate their environment. So that was the logic and its paying off, and Im quite happy with it.
Its just a standards thing.
Do Web services take over where CORBA left off?
Web services are different. The promise of CORBA is fulfilled through J2EE [Java 2 Enterprise Edition]. It is the logical derivative of CORBA, which is the logical derivative of DCE. CORBA is son of J2EE. J2EE is son of CORBA. And thats at a much lower structural level. Web services sits way up here. And its descriptors. Who are you? Describing you, your business, how you do business, etc., so somebody else can interrogate that database, and on a machine-to-machine basis you can coordinate between the two.
Web services is not the perfect technology or perfect answer to every issue. Itll improve as Java and other mechanisms have improved.
How has IBM been able to capitalize on technology built elsewhere, like Java and XML, even to a greater degree than the companies that invented the technology? Sun invented Java and had a hand in the origins of XML, yet IBM has leveraged these technologies to a greater degree.
This is why I say J2EE is son of CORBA. Java is certainly a creative and interesting piece of work as far as a programming language is concerned. And we embrace the language. We were doing a lot of work in the C and C++ area, and Java is a derivative of that. So we saw Java and its compatibilities with the C world as providing a hybrid language that would provide both ease of application development as well as good scalability with applications written on servers. We liked the language. And then given that the CORBA work was essentially C++-based, we recognized that Java could provide a substitute for C++ and that a new more modular, easier-to-work-with set of interoperability services could be crafted around Java. The language could be extended, the runtime services could be enhanced to create an EJB container architecture as the core of the J2EE structure. So we saw a path that we could use Java for what we were trying to use C++ and Object Request Broker Architecture for. Sun didnt see that. But Sun was never working on an interoperability strategy. Being interoperable with what? Suns a hardware company. They want to sell hardware boxes. Were trying to sell integrated systems. We have a software business that stands alone from the hardware business. So we saw the value of the language and the value of the environment and what it can be turned into to give us a kind of integration layer based on a single language that we could put tooling around that could be very powerful. They didnt see that. But then again, theyre not even in the software market. Its not genetic to them; they dont think software. Theyre not into that abstraction that is software.
XMLs a different animal. Weve been in the markup language business since the 1960s. We had GML and then SGML. When we shipped WebSphere in 1998, we had an XML parser built into WebSphere. Hardly anybody was talking about XML from an external perspective. We recognized the value of XML a long time ago as kind of this lingua franca for common descriptions. Weve been excited about it for a long time, and we jumped on it.
Does the Rational acquisition set IBM up to dominate the software industry?
Its an important acquisition to help extend our software business. And if you combine Rationals capabilities with the other products and capabilities we have, we believe we have the industry-leading portfolio.
Were clearly the biggest provider in middleware technologies and tools in the market. We are trying to define architectures that lead the industry with the kinds of applications we think businesses want to deploy. So we think were in a leadership position. Ive worked awfully hard to put all these pieces together to get it to this point. Theres a lot more to do, but its all part of the vision we had of the way applications could be if you have all the pieces working together properly.
Do you think Sun is now on track with the Web services standards, like WS-I?
Yes, we have always sought to have them involved. … But this is my personal view [regarding Microsoft]. It takes two. I think [Suns Scott] McNealy threw the first punch quite a while ago with his anti-Microsoft rhetoric going back years. Its one of the root cause problems here. He sort of started it as not only a company battle, but almost a personal battle. It became a me versus Bill [Gates] sort of thing. Like “I was going to retire but I decided to stay to help make the world safe from Microsoft for my children.” It just fueled all of this.
So what would be your message in terms of competing with Microsoft?
We compete with them on the nature of our infrastructure technology versus theirs. So we do Windows, we do a variety of Unixes, and we do IBM mainframe across our entire portfolio. We do scalable, reliable, secure, recoverable transactions. We have the ability to integrate systems from a process flow standpoint to do context passing and state-managed transaction flow across N-number of connected systems. We have an N-phased commit transaction engine in WebSphere that runs cross platform. I think when you look at the technology you can see very clear and distinct differentiation between the characteristic of what IBM delivers versus what Microsoft delivers. If the customer only wants simple connecting from a Windows system, calls from data somewhere else, or they want to initiate the system to do something and send them some data, well the basic connectivity/Web services technologies that were talking about here can cause a Windows system—which is uniquely .Net—to cause those things to happen.
If you want those things to happen under transactional control, you now have a more interesting set of problems to deal with. If you attempt to do that in a Microsoft world, you come out of the newer .Net technologies and into COM [Microsofts Component Object Model] and leverage the COM services to provide a level of transactional control at that exit point off your Windows server. … And Ive yet to see a medium or large business process flow that you could anchor to a Windows front-end environment and have any sense of real process integrity to it. So in some respects because they remain so isolated to a single platform, a world that is buying into the idea of heterogeneity and multisystems, a world that demands that everything legacy be tied together is a challenging environment for them, because their principal focus has been on departmental two-tier client/server implementations. This new world is really a tough for them to adapt into. And standards are really the first step in that process. There are other questions to be answered in terms of the infrastructure they provide and whether or not they ever intend to take any code beyond the Windows environment. We dont know how you do these scenarios without having code running in multiple places.
How strategic is Eclipse to IBMs tools strategy?
Extremely. The characteristic of the tools market has traditionally been fragmented, and tools that exist in one environment do not always exist in others. So availability of tools has been different. Developers have to determine if they get to use tools they are comfortable with or do they have to change tool sets and go through another learning curve around a new set of tools. So the phenomenon of openness and heterogeneity and flexibility is either encouraged by open tooling environments or discouraged for lack thereof. So the Eclipse effort was designed to open up the market for lots of participation around a common set of constructs related to a workbench.
Your competitors always try to knock the fact that the IBM Software Group relies on IBM Global Services to fix problems with the products. To what degree do you leverage IGS?
To a degree. But theres an illusion in the marketplace that IBM sells its products through Global Services. We all laugh at that. I have 10,000 full-time dedicated salespeople worldwide that sell software and only software—they dont get paid a nickel on hardware and they dont get paid a nickel on services. I have to do that for a reason, because theres nobody in IGS that gets up in the morning thinking about selling my products. They satisfy demand for skill, thats their job. And I have 2,000 people in the software group that do services, in addition to the global services people. And I have those 2,000 people because I have to do all the new product support, and I have to train Global Services. So those 2,000 people do projects as subcontractors to IGS in addition to doing projects without IGS. They also have to enable the other systems integrators. Since IGS is 10 percent of the market, I have to deal with the other 90 percent. Were not IGS-centric. I dont need IGS in order to sell or deploy anything. And in turn they dont need me. We team together all the time, but were not hardwired.
Who do you most hate to see in a competitive situation?
Theyre all so annoying [laughter] and they all have different tactics. Companies that dont show up are the hardest to beat. Microsoft is a very challenging company to compete against because they dont show up. Youre competing against the customer people who have bought into the Microsoft story that you can do anything with Microsoft. And therefore the contest is very often not a debate. Increasingly were seeing more and more Microsoft people show up in accounts and we like that. We like to compete head-to-head, our people versus their people because now we believe its a fair fight, in terms of us being able to demonstrate our capability versus their capability with the customer sitting in the neutral chair. Microsoft is uniquely challenging because theyre a marketing machine. Theyre getting a lot of credit for things that they may do in the future that they dont do today. But code talks. I love those types of competitive situations where you show up with your code and I show up with my code and we bake it off.
Latest IBM News: