Open Sources | Rodrigues & Urlocker » Interview with Red Hat CTO, Brian Stevens

October 03, 2006 | Comments: (0)

Interview with Red Hat CTO, Brian Stevens

TAGS: Open Source

I was very fortunate to talk today with Brian Stevens, CTO of Red Hat. It's easy to fall into the trap of thinking of Red Hat as the "Dell of software," as some have opined. Or is this a trap at all? Is it something to be emulated, or shunned?

I asked Brian this and other questions. Here are my questions and my best efforts to capture his responses. I wasn't able to record the conversation, so please pardon any errors I make in my live transcription of the conversation.

Red Hat is somewhat anomalous as a technology company, and even as an open source technology company, in that you don't own most of the code you ship, and you don't employ most of the engineers who "work" for you. Given this, what do you do as CTO?

Red Hat now employs roughly 2000 people in the company, 600+ of which are engineers. Unlike in the past, those engineers are not exclusively or even predominately focused on "sustaining" technology. They're focused on innovation.

Our approach, then, to the type of development talent that we need is largely no different from the proprietary technology companies. We need people that can help us drive the open source technology forward, and then support it. We're no longer a pureplay "packager" of technology, but are actually innovating new code/technology.

Where we differ from a proprietary software company - and differ significantly, is in our model of development. Our development model is open. We put our 600+ engineers out in the open in the open source development community, where they are known for the code they write, not the company they work for. Ownership of the code isn't what drives us - creation of compelling new technology drives us. Our business model allows us to give away the code, deliver strong customer value, and still make money in the process.

It's technology tied to business strategy. It's a two-way street, too. The SLA [service level agreement] with Red Hat doesn't begin when a customer has a support issue. It begins when the customer helps drive the direction of the development in the first place. Our customers are co-creators with us, and that's the "ownership" we need to drive our business.

Do you think the Red Hat model would apply equally well to other areas of software?

Red Hat's model works because of the complexity of the technology we work with. An operating platform has a lot of moving parts, and customers are willing to pay to be insulated from that complexity.

I don't think you can take one finite element - like Apache - and make a business out of it [using our model]. You need product complexity.

So, does the Red Hat model lend itself well to JBoss' application server? Was that part of the decision calculus for the acquisition?

Absolutely. It satisfied that litmus test for us. Plus, it helped that JBoss' business model was aligned with ours - taking a proprietary product and opening it up would not have worked as smoothly as our integration with JBoss has. It's one thing to merge a few bits, but quite another (and arguably more difficult) to merge a drastically different business methodology and culture.

Tell me about Red Hat's Emerging Technologies group.

In 2001, when I joined Red Hat and built the enterprise engineering team, we weren't being as aggressive as we could have been about driving innovation in open source. We were mostly packaging existing open source software. We wanted to change that. So, we formed the Emerging Technologies group to to bring "future technology" to customers through open source.

Our focus areas include:

• Real-time
  
• Messaging
  
• Management
  
• Virtualization (Xen)

We're working to drive innovation in these four areas, in particular, to enhance our customers' experience with open source enterprise infrastructure.

How has this focus on innovation changed the company?

It's changed the culture in a very positive way. We have engineers who don't just want to find bugs - they want to define technology direction for the industry. Our focus on innovation has enabled engineers to find their "niche" within the company, whether in innovative or sustaining technology.

It's also enabled us to more aggressively drive value for our customers. We're no longer just leading in open source business thinking, but also in technology. We love that.

What is your biggest frustration with the open source model? Do you ever feel limited by what the company can deliver?

I was one of the architects at DEC for roughly 15 years, and our pace of innovation there was very slow compared to Red Hat. Code ownership doesn't seem to accelerate innovation, and probably inhibits it. I'm very happy with the open source development model, at least the way we practice it.

Which open source projects/technologies do you think Red Hat is best positioned to help improve, and which are most exciting to you?

The $100 PC is hugely exciting. We're getting to really change the way we think about a PC - these machines are flash memory-based with mesh networking, among other things. A very different machine from the PCs that sit on the desktops in the United States. We think this project will bleed into other markets, beyond just emerging ones, and technology. Again, this puts Red Hat on the vanguard of not just the "Linux desktop," but the desktop, period. That's very exciting.

Virtualization is huge, but largely untapped. VMware is fantastic but not as pervasive as it should be. Xen can drive virtualization to ubiquity. Xen will make SOA, grids, etc. real.

The last thing is AMQP (Advanced Message Queuing Protocol). [Brian then launched into an explanation of why it's cool, but he lost me. One interesting thing to me is that Red Hat has chosen to go for an Apache license with this, specifically so that it can be a pervasive messaging system, and not just "the open source messaging system."]

Red Hat on the desktop....I haven't heard much from Red Hat about the Linux desktop in a long time. Is it a focus for the company? If not, why not?

We think that the $100 PC is the best place for us to invest our desktop innovation dollars. It's the right thing to do because it opens up computer/Internet access to a huge group of people previously priced out of the market.

We also think it's the best way to disrupt Windows on the desktop. Having said that, we aren't trying to build a Windows clone. We're looking for fundamental ways to rethink the PC, and this disruption in terms of cost and technology is much more interesting than just doing another desktop OS. We want to completely change the paradigm of desktop computing. Again, we aim to be a technology leader, not a follower. The $100 PC helps us to do just that by forcing us to fundamentally rethink and retool the software for the PC.

How important is price to Red Hat's customers? Put another way, are you still slogging through the "Linux is cheaper than Windows is cheaper than Linux" TCO debate?

I don't see that. Cost matters to the CIOs I talk with, but usually it's the cost of deployment, under-utilization of software and hardware, manageability, etc. It's not the software licensing cost that most concerns them.

Our job, then, is not to re-create existing proprietary technology, but rather to enable IT to either dramatically lower costs, or to do things they've never been able to do before. We focus a lot on the latter opportunity, but also consistently drive costs lower through better manageability, virtualization, etc. We're the best value for a CIO's money, no question. CIO Insight honored us as such, two years running.

With the JBoss acquisition, Red Hat now has a robust stack: operating system + middleware. What's next? Applications?

For the foreseeable future, Red Hat will be focusing on the infrastructure elements of the software stack. Management, messaging, high availability, SOA, etc. It's a pretty big swath for us, and 99% of what we're hearing from our customers has to do with infrastructure. The better we can make it run for our customers, the happier they'll be.

What was the hardest thing, from a technology perspective, about the JBoss acquisition?

For the most part, there wasn't a lot of overlap between the two technology bases. JBoss sees things from the top-down (application layer), and we see things from the bottom-up (operating system), which differing perspectives have really helped us. So, the technology integration was actually quite seamless.

We assumed that integration of engineering teams would be the more difficult task, and were actually pleasantly surprised to find the teams gelling together quite well. We originally planned to have JBoss work as a separate division, but there was so much cross-pollenization that it quickly became clear that a tighter integration made more sense, and was culturally appealing to both sides.

Prior to the Red Hat acquisition, JBoss was developing (or talking about developing) all sorts of new extensions to its core application server product: an Enterprise Service Bus, content management system, portal, etc. How will Red Hat's ownership of JBoss change the JBoss product plans?

We're basically uniting our efforts around Red Hat's strategic technology direction: building the best platform and infrastructure on which to run applications, with some emphasis on Real-time, Messaging, Manageability, and Virtualization. There are some differences, of course - JBoss needs cross-platform support (though we want to make sure we improve our RHEL product to ensure JBoss runs fastest with enterprise Linux). But the two companies' visions were basically aligned.

Do you think Red Hat will ever get into the application space?

Not particularly. We create the best platform on which applications can run and compete with each other. That's our business.

Tell the truth: Do you wish you had a Mac?

No, because I already have two. I have a Mini at home and a core duo laptop sitting on my floor. I dual-boot on my PowerPC-based Mini and use Parallels on my Intel-based laptop. You actually can see a lot of Macs within Red Hat.

I knew there was a good reason I like Red Hat so much. :-)

Posted by Matt Asay on October 3, 2006 02:16 PM

RATE THIS ARTICLE: