Is Linux finding more success in specific vertical markets?
One inhibitor to adoption has been that when companies look at the solution stack [operating system, middleware, database, applications] they have built on Solaris, for example, and they question whether they are going to be able to migrate the whole solution stack to Linux. Because migrating 90% of that solution stack is not good enough. So, that's one of the reasons we have to be smart about the industries to pursue.
The financial industry has been interesting because they don't necessarily have these dependencies on external vendors. They have their own solution stacks they created writing C and C++ applications. For them, the process is a matter of working with them to move these applications over. So that's worked pretty well.
Now it's expanding because we've got IBM Websphere on board, Oracle, BEA and those applications. Now that they are there with Linux, the scope of applications available has dramatically increased. So, Linux is big with the financials?
Eight of 10 major commercial banks are using Linux. Now the government is the next big hunting ground for most people. It's nice to have a target market where budgets go up. Early Linux penetration into the enterprise was in basic functions like print and file serving. What are customers using Linux for today?
We are beyond doubt that it is used at every level. It replaces mainframes. It replaces the heavy duty multi-processor Sun systems, etc. All the supercomputing done today is done on large scaled-out clusters of Linux machines.
The notion of going from two- and four-processor machines to eight- and 16-processor machines and eventually 32, 128, we think that movement is slowing down. People are finding that deploying 10 $3,000 machines from Intel gives them an enormous amount of ability. You can cluster on that and there's no longer a single point of failure.
This is all about scaling and off-the-shelf cheap. If you talk to any CIO today. That is all virtual. If they need more capacity for their database, they plug in more hardware. The software never needs to know, just adds to the capacity. Now we're talking about utilization.
There are banks that have a whole range of machines and during the day they are customer facing ecommerce site and at night, they process checks. And in the morning, they reconfigure them back. That's the mecca they are looking for.
That doesn't mean the machines have a lot of mainframe characteristics. There are a lot of them there. At the same time, we are aware of projects. Linux not penetrated the 64-way, 128-way; that's still the domain of the Unixes. Short of that, there's nothing that doesn't happen on Linux today. What misconceptions do some enterprises still have about Linux?
The meaning of free. We have two products here: Red Hat Linux 8.0 and Red Hat Linux Advanced Server. Over time, there will be an enterprise line of Linux products and a consumer line. On the consumer line, you would be able to download it free from our Web site and deploy it to as many machines as you want to. The role of that distribution is to give people an accurate insight as to where the community is at any point in time. We'll release it two, three, four times a year and put all the latest stuff in there and make it available. Improvements will be made that way.
If we released that for enterprises every four to six months, Oracle would say they are still working on applications for the first one, for example. The situation that arose from that is that the customer would have to buy Red Hat Linux 7.2 to support DB2, 7.3 for Oracle support and SAP is only available for 7.1 and there's no Oracle database for 7.1. It was a mix-match of stuff that never worked. What we said was from the same development tree, we branch off a version that we call enterprise Linux, which is our Advanced Server. We will make sure many third-party applications work on it and that it's stable. Is there a common theme to issues you hear from customers who want to use Linux for mission-critical work?
Not many. Companies want to get to know us and approach us and say 'here's my plan, I want your help.' We don't talk to many people any more who need convincing that we spend our time talking to people who need convincing. That's where the money is. IBM is saying that and HP too.
Some people are asking about their multiple Linux distributions, how they can all stay in synch. That's a question of people applying what they saw in the Unix world to Linux. But what people don't know there is that if you take IBM's version of Unix and Sun's version of Unix, those companies are responsible for the whole stack of stuff that went into these distributions. They made have had the same roots at some point, but after they branch, they never met any more. SuSE and Red Hat engineers work together today, part of the same groups. It all branches from the same development tree. In the open source world, that doesn't apply.
The other question we get is now that you are pretty dominant, are you going to be the next Microsoft and the next monopoly in your space. There's no way we could be that. The way it works is with Microsoft and their new licensing program for example, where some customers may have to pay 30% more, people were upset because there is no other way to go. You could not find another Windows provider.
But in the Linux world, once we hand you the software, it's yours, not ours. There's no license tied to it. If you're not happy with us, call IBM and see if they can service this and they'll be happy to take your business. You cannot be locked in the way you are with proprietary software, so that's a good thing for enterprises. It's a strong sales argument and gives companies control back. What's your opinion on the impact of Microsoft's recent licensing changes to Linux?
It certainly drives people to Linux. We have been selling to the CIO types and once we have confidence with these people, they bring us into new opportunities.
Click here for part one of SearchEnterpriseLinux.com's interview with Red Hat.
FOR MORE INFORMATION: