You introduced the z9 zSeries system. What has happened since the launch of the z9 last September? If you followed our fourth quarter, which is really the first full quarter we had that product in the market, we ended up having one of the best quarters in the history of this line of servers.
Didn't you have customers holding back on buying? In other words, wasn't some of that pent-up demand? No, I think a lot of it had to do with our customers growing. If you look at where we are serving customers in the enterprise, a lot of the big banks, a lot of the big public-sector customers, a lot of the people that run transactions online are growing. This gave them vertical room to grow. These customers can't hold back. The demand is there for this scale of a platform.
Where is your new-customer growth coming from? A lot of the new-customer growth is happening in Brazil, Russia, India and China. I just left China. There is a tremendous amount of growth there.
Most of your new mainframe growth is happening overseas? No, a lot is happening here. The small and medium business is growing. What was a medium-sized company five years ago is now a big company today. Those companies are all mainframe customers in the United States.
The z9 was your first mainframe to get Advanced Encryption Standard capability. Is AES prompting customers to use the mainframe in different ways? The No. 1 concern/question that I get from customers is about security. What they want to talk about is how do they exploit the full capability of the mainframe. Everything from key management, centralized management of encryption across the enterprise, AES, intrusion detection -- they want us to help them manage and exploit the capability for security on a mainframe. Most of our customers tell me it's one of the principal reasons they buy a mainframe -- because it's secure.
Is AES prompting customers to use the mainframe in different ways? The answer is yes. Customers are saying they want to run a variety of workloads on the mainframe. They have been gradually moving edge-of-network applications and workloads [running on Linux] onto the mainframe, and they are now bringing a lot of the Web server apps onto the mainframe. They feel that the edge-of-network [applications] are the first point of defense. So, yes, they are exploiting AES -- and are concerned about intrusions inside and outside, by the way.
A complaint by mainframe users over the years is the cost of software. Are you doing anything different to address software costs on mainframes? Do you have any new plans, new pricing? The newest idea is the ability for customers to run workloads on our specialty engines. Back in 2001, we introduced an IFL [Integrated Facility for Linux]. We extended that with zAAP [zSeries Application Assist Processor] for Java and in January, zIIP [Integrated Information Processor], which allows them to run DB2 on a specialty engine, so they get the advantage of mainframe pricing. That's been our strategy. It's turned out that the majority of the revenue and workloads that are coming to the mainframe are these new workloads -- about 60% of them are Java, Linux workloads.
What's the relative price savings if you run on a specialty engine? It varies. I hate to give you a number, but they do save money. The other thing is it works better because the engine is tuned for the application and workload, so they eliminate some inefficiency. They get a double savings in terms of more efficient running but also a better application because it's tuned to a specialty engine.
Are you developing new specialty engines, and if so, for what kinds of workloads? Our customers keep telling us that they are interested in us continuing this road map, which says any workload I have in the enterprise, they want to be able to run it on the mainframe, so we're open to that. It's totally in collaboration with customers.
IBM has announced a goal to have 20,000 professionals trained in zSeries skills by 2010. Where are you on that plan? We're ahead of schedule, and I'm really excited about it. I had no idea that the market would respond as well as it did. The goal is 20,000, and that's in general. We got a special challenge to get 10,000 [trained] in China. And we are well on the way of doing that. We are ahead of schedule. We will get to the 20,000 by 2010, and the biggest reason is our customers, our business partners, the universities are all working on this problem together. We've got a lot of people that are now coming to see the mainframe, know about the mainframe, learn, get a job working on the mainframe.
What I'm finding as I talk to some of the younger employees and younger people around us is they don't have legacy, they don't have biases, the distributed system era, the Unix era. ' What they are thinking about is security, resiliency, global reach, and they want to work on things that can do that. A lot of these kids grew up in the open world, they are very familiar with Linux and other open-source, and when they learn the mainframe runs five different operating systems, they are thrilled by that. So we think we are going to get the 20,000 easy.
So you are ahead of schedule? I think we will get to the 20,000 before 2010; in the next three years or so we will get to the 20,000.
What are customers telling you they want in terms of innovation? What are the things that you are hearing the most? The thing I hear the most is around security. I think customers [are] putting in place very rigorous programs around security.
In terms of security, what do customers want in addition to what you have already given them? The first thing they want is to help understand the full capabilities that we've actually sold them. The other thing that they are asking is to help them run enterprisewide security, but using our platform as the platform that manages security for all other platforms across the entire enterprise, no matter where they are. Out of those conversations, we're learning new things that we need to go invent and build for them.
Tell me something you need to invent. Some of it is automation. Some things that people do as far as monitoring -- customers want it to be autonomic, and we're doing a lot of that. They want us to exploit every opportunity to make things automatic, to make the security be as intelligent as possible and not have to rely on individuals. The direction is in the area of autonomic and automation around security.