Sun Microsystems Inc. CEO Jonathan Schwartz thinks that the economic downturn and Wall Street meltdown will make IT managers more open to change than they may ever have been before. And that is going to benefit Sun and its open-source strategy, contends Schwartz, who is also the company's president and a high-profile blogger. In an interview with Computerworld last Thursday, he spoke about the economy, technology, innovation and servers. Excerpts follow:
What are you doing to help your customers with the economic problems? We are preconfigured for the downturn. If you think about the discretionary expenses that go into operating a data center, first and foremost there's the physical plant itself — the physical space, the power consumption, the HVAC. So all the work that we do around energy efficiency, and on getting optimal performance — a lot of folks say, 'Gee, Sun, why are you so focused on the environment?' It's because the environment ends up being a huge operating expense for our customers. And to the extent that we can help them lower their environmental impact, we're also lowering the economic impact on their businesses. That's clearly Job 1.
The second element of discretionary expense is software licensing, and probably the single biggest software license that our customers have to pay for is [to buy] proprietary databases. Second on that list are proprietary application servers and an application infrastructure. I just was with a customer this morning who didn't recognize that he had roughly 2,000 developers working with MySQL because it wasn't a purchase standard [within his organization], but it had become the de facto [database] standard. He didn't recognize that he could get that level of productivity [from an open-source database].
GlassFish — the same is true for the application server marketplace. OpenSolaris — now that it is multivendor and multiplatform and the source is available, those environments where you don't need support don't have to pay for it. And then we enable customers that want to subscribe in production environments to pay for the supported version. Across the board, whether it is hardware or software, you are going to see some great systems innovations coming from Sun. The [server] platform that we are launching Monday is a good example of what happens when you bring all of that together: You end up reducing customer spending. That, by the way, doesn't mean that you reduce your revenue; it means you take share from your competition.
Do you really expect customers in the near term — with the economic downturn — to, say, swap out an Oracle database and replace it with MySQL? Unquestionably. Now that doesn't mean they are leaving Oracle — Oracle is a fantastic company, and they've built a fantastic database. But there is no longer one-size-fits-all in the enterprise database marketplace.
Let me ask about MySQL. One of its co-founders, David Axmark, just said that he's leaving Sun, so there are indications that there may be disillusionment or trouble within the MySQL ranks. Can you address that? Sun is a big company. We've got people who are leaving and starting companies, we've got folks that are coming back into Sun. Healthy organizations have folks who are bringing in new ideas, and occasionally, folks after their 25th year of contribution decide that they want to do something else.
There are different talents for different parts of the job. Now that MySQL is going into enterprise mission-critical environments, they are taking a very different approach to how they build for scale. They are going to probably focus a lot more on stability and platform efficiency, and that may come at the expense — although hopefully not — of being on the wild and woolly edge of having a new [software] rev every three days. So that is not always going to sit well with everybody, but that's also why you don't ask everybody to do the same job.
IDC's server revenue numbers for the second quarter show a decline for Sun in the amount of hardware that was shipped and even in Unix market share. What's your take on that? We are predominantly a systems company — 75%-80% of all the revenue at Sun comes as a result of selling systems. And systems are the combination of silicon innovation, hardware engineering and packaging, as well as software design, so you really have to look at the overall picture. We pare our systems business into a few separate lines. By far and away, the fastest growing is the Niagara CMT [Chip Multithreading] line. That is over a billion-dollar business, and it is growing 60%-plus a year; it is arguably the fastest-growing business that we've ever built at Sun. It is completely on fire.
What has been slow for us are the very high end systems — the enterprise systems space. And frankly, that's been slow for everybody. That's not the growth part of the marketplace. Between us and IBM, we give as good as we get. But at the end of the day, the systems marketplace isn't about to shrink. Virtualization and [server] consolidation will make things more efficient, but we just think there is ample opportunity out there, and the best way for us to get after that opportunity is just to keep innovating.
With Intel's new six-core Xeon chips, vendors of x86 systems will be going after the same parts of the market that you are: ERP, CRM, databases and consolidation projects. How does your new four-socket Sparc Enterprise T5440 system match up, and can it stay ahead of the x86 platforms? We also ship x86 systems, both Intel and AMD. And with the innovations we've had in Solaris, we've optimized for the highest-scale x64 systems in the marketplace. I think the T5440 is designed to optimize for a different set of variables.
We are going to go after the data center market in total, and we're going to present customers with a choice. Just as we run Solaris on a zSeries [mainframe], we're going to run Linux on our Sparc systems. We're going to give customers the best set of options available, and then they tell us what works. We don't go in presupposing for them that they will buy one set of products.
On your blog, you talked about aggressively expanding your customer base. How does the T5440 help you do that? It's a little unlikely that this server is going to be the first system that a new customer buys from Sun. I don't want to close off that option, [but] it's more likely that they pick up a one-socket Niagara system. Just on price point, you seldom spend $50,000 to $100,000 on your first server, and that's the price range that these [new] systems start at.
Niagara as a whole, though, gives us access to a market that really is representative of a very unique problem space. We don't see IBM with [its System p line] at all in the Niagara space. What brings new customers to Sun is differentiation and innovation. We want to be 50% faster, or 50% more energy efficient, or half the size. Those things, when added up across real large data centers, mean real money to real customers.
Sun has fostered a reputation as a disruptive company from a technology standpoint. But what will it mean to be disruptive going forward? You want to be careful. You want to be disruptive to the industry; you don't want to be disruptive to your customers. I'll give you a great example of the kind of disruption that the market is going to see from Sun in the next 12 months. We have been very aggressively promoting OpenSolaris in the marketplace, and there are a lot of storage vendors that have been really excited to embrace open-source operating systems — so long as they stay on servers.
As you have seen with Thumper — a 48TB storage platform based on the ZFS file system — we're planning on taking Solaris and extending all the skills and knowledge and ecosystem that we built in our server business to our storage business. That now means open-source platforms will be at the heart of open storage as it evolves as a market category, and we plan on being a leader there. That's very disruptive to the competition.
What role does FUD — fear, uncertainty and doubt — play in the server market these days? I'm asking this because Linux advocates don't seem to miss an opportunity to explain why that operating system will crush Solaris at some point. How do you counter that? We don't pay a lot of attention to that; we pay a lot of attention to customers. We are very well aligned with the Linux community. We're not the enemy of one another — and I know there are folks who get emotional about that now and then. But our focus is going after the proprietary vendors that are causing our customers a lot of grief and a lot of pain. And the more we focus on solving those customers' problems, the more they embrace open source.
Some people think that the current economic problems will accelerate the adoption of software-as-a-service technologies. What are your thoughts on that, and how are your products going to line up to support SaaS? The downturn has already caused customers to get ruthless about saving money. Customers under stress are open to change. And that is what I see from every customer I've spoken to, especially in the last 60 days. That means they are open to change in moving away from proprietary software vendors and proprietary storage vendors, more open to moving to software as a service, more open to moving to free software — and that, again, creates opportunity for Sun. I think the doors are going to be more open in the next year than they have ever been.
Where is this innovation going to come from? According to the National Venture Capital Association, there have been only six venture-backed IPOs this year; mergers and acquisitions are down as well. The fear is that investment dollars will dry up. Innovation rarely arises in a bubble. Someone clever once said that necessity is the mother of all invention, so believe me, people are becoming a lot more innovative as I speak. Why? Because they have to. If your budget just got cut 50%, I promise you everything is on the table. There is no better time to start a company than right now. It may be tough to find funding, but there is no better time to go look at the parade of legacy technologies that need to be replaced and the extraordinary interest from customers in entertaining new ideas.
If that's true, are you considering any changes in where your research dollars are being spent? In general, we're looking at ways to increase R&D. That doesn't mean across everything. It means to double down on those parts of the market that really represent clear revenue return. Coming back to the T5440, although [the Niagara platform] is more than billion-dollar business, you have to remember that [work on] the first silicon began in 2001. R&D takes patience, discipline and rigor. We're not going to make changes within the quarter or within the next six months that are just going to be episodic or ephemeral changes because there was a downturn.
In regard to IT automation, HP CEO Mark Hurd's observation is that while offshoring reduces costs, automation eliminates them. Do you agree with that? And if so, can you tell me what you're doing with automation and where you see Sun going with it in the next year or so? If you take a T5440, you can run 256 separate threads on a single machine. That means if you were running historically 256 separate machines, you can collapse them on to one. Our invention and our innovation start from the silicon and rise up through the system and then move through Solaris and OpenSolaris up through the application tier. Companies that are trying to apply that after the fact on somebody else's microprocessor or somebody else's operating system end up, from where I sit, being less efficient. They're dependent on other innovations for their automation to work.
Fixing a problem after it has been created is certainly one business, and it's a pretty good business. We just prefer to fix the problem before it's been created.