Red Hat is sticking to its Red Hat Enterprise Linux 5 roadmap, because it's still drawing in new customers and retaining the old ones, said Tim Yeaton, vice president of enterprise applications, in a one-on-one interview with SearchEnterpriseLinux.com.
In this Q&A, Yeaton explained why Oracle's Linux can't make the grade and the Microsoft-Novell brand of interoperability is lacking. He also discussed the learning curve for virtual appliances and the upcoming Sybase-Red Hat virtual appliance.
SearchEnterpriseLinux.com: At Red Hat Summit, I noticed that your message and roadmap hasn't changed much since October, even though competition has increased and come at Red Hat in new ways.
Tim Yeaton: The proof is in the numbers. As we announced in our earnings call, we had over 54,000 net new customers in 2006, and we averaged 10,000 to 15,000 new customers a quarter. For the year, 99 out of our top 100 accounts were renewed, and the one [Oracle] who didn't renew it, well it was obvious why - - they decided to go into business for themselves. [He's referring to Oracle.]
Oracle is coming at Red Hat hard with its discounted support offering for its Red Hat distribution knockoff, announcing customer wins and gaining momentum. What's Red Hat's reaction?
Yeaton: There are well in excess of 50 Red Hat derivatives in the world, but I can't ID a lot of them that pursue the same kind of business model as Red Hat. Also, there are innovations in our release model that are impossible to replicate.
We have the ability now to extend our support model to allow a customer to rely on a point release for a period of time. Technically, to them it looks like a point release but functions like a full stream for a certain number of years. Red Hat supports roughly eight or nine active streams.
There's building binary compatibility, and then there is assuring ISVs you have achieved it. That takes years of competency. It is very easy to introduce new capabilities and break a little thing here and a little there, and before you know it the ISVs are all over you saying they can't rely on that commitment anymore.
If a company is trying to take apart and put together our product, it is a hard endeavor. But on nine streams, it is very complex if [the company] does not have the same testing infrastructure that we have in place.
On the other side is the Microsoft-Novell partnership. Last week, they also revealed new customer wins. Their particular strength seems to be Windows-Linux interoperability.
Yeaton: We had a customer advisory board [meeting], and interoperability was on the agenda.
Their first issue was application development interoperability. With new applications, can they interoperate with the old? As customers go to Web services, will their applications interact with all clients at once? We have a very clear statement on that: JBoss. JBoss is multi-platform and will always be for multiple platforms.
The second issue is data interoperability. We hear from a lot of customers who have their data locked up in these application silos. This is the area where the Meta Matrix acquisition fits in. Next area we are addressing is user interoperability in areas like ID management. We have had very strong capabilities in this space since the Netscape acquisition.
The most important thing in interoperability, however, is to conform to open standards. There are good open standards in LDAP today. I think a couple of vendors with one-off proprietary variants think the way to go about interoperability is to build a proprietary pipeline between two proprietary applications. We have also been doing open standards work with Active Directory since Netscape.
Was file-and-print interoperability with Samba brought up?
Yeaton: Yes. The last area of interoperability at Red Hat is systems services, file-and-print services. Samba is really the standard, and we have since brought over the rest of the Samba team into Red Hat.
We have to remind the marketplace that interoperability is on the customer's terms and not the vendor's terms, and we are focused on doing that.
How informed are customers today about virtual appliances?
Yeaton: It varies. For the most part, this is a case where many customers understand the concept; but until they actually see the offerings and can play around with them, they won't know for sure.
We know [virtual appliances] will be compelling to customers, but I think there is going to be a lot of education that is still going to have to happen. Our partnership with Sybase will roll out in the second half of year. In that intervening time we will build out the program so that it makes sense to partners who want to do this approach.
How will data center management change when virtual appliances do begin to pervade the data center?
Yeaton: There is running multiple virtual instances in application-specific way, and then there is building virtual appliances like the one with Sybase.
In one scenario, the power of the model is the high rate of consistency in the images that you deploy as virtual instances. Those scenarios are actually more regionally manageable because we have consistency at the image level.
If, on the other hand, people are building very different applications, configurations and those kinds of things [in virtual appliances], you can make some of it better with underlying management tools. In the end, however, if you are going to have every application management tool and database manage tool as virtual instances, you can actually introduce more complexity than you were planning for.
What is Red Hat's virtual appliance roadmap?
Yeaton: We certainly expect to see the virtual appliance space as a paradigm that a lot of partners and ISVs will be interested in. The model might not apply to all of them, but there has been interest.
The first mission is to build the platform, which we have done, and the business model that enables it to be an effective deployment vehicle. That [latter] part we will do over the course of the next few months. At the same time, we'll be engaging our ISVs and Red Hat Exchange partners to gauge their interest.