You mentioned a possible partnership with EMC. Is that something thats in the works?
We have a very strong partnership today around virtualization with VMware, and we are a Tier 1 OS support platform for the VMware team. And were the only Unix that has that distinction. VMware doesnt support HP-UX or AIX [IBMs Unix variant]. So what were really seeing is a split in the marketplace away from proprietary Unix, and its the open-source Unix and the open-source community in general [that are] really giving us opportunities, and the VMware example is a good one.
Weve got more to do with EMC. I think we see the market quite similarly. And at the end of the day, they also believe in innovation. We have a lot in common with them.
So you see the possibility that the partnership with EMC will go beyond VMware and into storage.
Were the No. 1 platform for EMC storage today. … So I think theres all kinds of opportunity in the marketplace.
Youve talked a lot about Niagara and SPARC development. Where do the Galaxy systems fit in with your overall plan?
Its very definitely at the core. … A fast general-purpose system displaces all kinds of custom hardware. AMD has been tremendously successful in the marketplace, and I believe weve played a very strong role in that. We think theres lots of innovation to come, but the customers going to be able to make the choice at the point of purchase—do I want to put a SPARC chip in this or an AMD chip in this? But for the most part, the systems will otherwise be identical. Well still have the same mechanicals, the same supply chain, the same standard interfaces, the same software platforms—its going to be up to the customer to choose.
Intel seems to have made somewhat of a comeback with its "summer of servers." Are you concerned at all that your AMD-only [x86] product line might be limiting to you?
Intels made some really interesting press announcements, but at the end of the day, we want to build what customers want to buy, and what weve heard consistently from customers—and this is not from Intel or AMD—is they prefer SPARC and they prefer Opteron. And the numbers bear that out. … Given the share gains, I think thats pretty much irrefutable.
Going forward, we want to continue to listen to customers and present them with the choices they require, and if Intel is able to get themselves back into the game, by all means wed be interested in looking forward. But again, were going to pretty much stay focused on what customers are telling us, and what theyve told us—and theyve spent their dollars to prove the point, AMD has made tremendous share gains over the past few years—is thats where they think the markets headed.
With SPARC and Opteron, we can present customers with a lot of choice, but thats not to forestall any future choice. Were pro-customer, were not anti-vendor.
How much of an advance do you expect Niagara 2 to be over Niagara?
It will be epic. … This is not an incremental innovation. This is a doubling in thread count, a massive improvement in memory bandwidth [and] overall performance numbers that are just staggering, so we expect it to be very disruptive.
I had a very interesting conversation with an executive in our sales team not too long ago who was yelling at me over the phone because he had such a huge performance advantage over some of our competitors that he thought wed made a tragic mistake in under-pricing Niagara. He said, Dont you understand what a big advantage we have? Why cant you up-price this thing? If all I do is sell this to my own customers, my revenue might go down. And I said, Well, you just made the point. Dont spend all your time selling it to our customers, try selling it to other peoples customers.
Its selling into the PA-RISC install base that [was] for all intents and purposes abandoned. Selling to the Power install base because the Power6 transition doesnt look like its happening anytime soon. Selling into the legacy Xeon marketplace because customers can see a 10-to-1 compression in their power footprints and their data center space consumption. I think the markets really opening up in front of us. <<p>When do you expect Niagara to become a multiple processor architecture?
I dont think weve articulated a date. … We have Niagara 2 systems up and running today, and they are wicked fast. And the good news is, all the portfolio of applications that are out there running Solaris today, they just run [on Niagara 2]. No new work. A lot of peers in the marketplace seem to want to introduce the next new generation chip set, but then they always break the application binaries, which ISVs hate, because they have to go back and hit the reset button and rebuild their systems.
One of the great things about Sun—one of our little secret weapons—is that customers dont have to do any work to get our new upgrade. Thats why were seeing such a big adoption of Solaris 10, and frankly such a big adoption of [UltraSPARC] IV+ and Niagara systems.
That will now start to apply to our storage systems as well, with Thumper as really the first foray into that space, and that will begin to spawn a whole series and breadth of NAS offerings.
As well, youll see Solaris start to play a much more prominent role in the traditional archiving and tape offerings.
What are you seeing regarding trends in corporate buying? Some analysts believe its beginning to slow.
Weve seen no evidence of slowing whatever globally. What weve seen is a slowing of system purchases from vendors who are end-of-lifing their platforms. You should expect to see HPs PA-RISC customers slow down, as they did last quarter, because they end-of-lifed it. … If you announce that youre going to kill something in the marketplace, people just arent going to buy as much of it.
On the Dell side, I think weve seen a slowing of point-product purchases. A slowing of customers saying, Oh, procurement is in charge, when all they care about is a cheap box. And now I think were moving toward IT leaders being in charge, because also procurement was able to go to a white box vendor—you now find yourself trying to support 17 different service processors and 14 different drive manufacturers because procurement was able to build your data center out of the cheapest parts available. But now youve been saddled with the operational costs of running it and that absolutely dwarfs the purchase price.
I think were seeing a movement in the pendulum away from procurement as the architect of the data center toward IT leaders and technologists architecting data centers.
I think thats representative of a split in the marketplace. Customers are really turning into those for whom procurement will forever be in charge and customers for whom architects and business innovators will be in charge. There are a lot of companies out there who make their living selling to customers for whom procurement is in charge. Those folks just care about the cheapest possible box once a month, or the cheapest possible printers or the cheapest possible PCs. Thats not our core market. Our core market are those network operators or service providers who make money off of innovation in the data center.
Theres the traditional enterprise and the emerging next-generation service operators. My advice to a customer looking at [CRM (customer relationship management)] solutions is, rather than exploring how you can architect your own, why dont you go to Salesforce.com? So guess who were going to be focusing our time on? Were going to do our best to try and be the worlds best vendor and partner to Salesforce.com.
But that means were going to stay out of trying to sell to flower shops and shopping malls or dentists offices, [all] of whom, by the way, purchase a tremendous amount of IT in aggregate, but over the long term, theyll stop buying IT, theyll start buying services from companies like eBay and Yahoo and Salesforce.com, who are Suns customers.
All that said, if you hang out on Wall Street … those are companies the make their living on the basis of technology advantage, and our technology and our road map deliver competitive advantage to those folks. Thats why theyre interested.
Whats going to be your message to Wall Street, given the amount of success Sun has had in the past with it?
The message is going to be a very simple one, which is, were still listening. Were there to give folks an update on the innovation and the technology road map and were here to say, "Thank you for the feedback youve given us historically, and were still open for business and were still looking for the next wave and weve got eyes and ears open and we want to know what we can do better."
Sun executives have said the plan is to open-source the entire software stack. Can you give us more specifics as far as timelines and frameworks?
As soon as humanely possible. Im not just being glib as I say that. I want you to know why it takes time for us to open-source software. We are not dragging our feet in the least. If anything, I have our legal teams working day and night to get this stuff done, but at the end of the day, I must be able to tell a customer, I will indemnify you for the technology you use to run your hospital. I will indemnify you for the technology you will use to run your carrier fleet or to run your bank network.
I cannot deliver to customers technology I cannot indemnify. …[Do] you remember when BlackBerry was being sued, or more recently when EchoStar was being sued for the usage of Tivo? They actually threatened to turn off peoples Tivo systems. I cant risk that disruption to the customer base, so what we do at Sun is we indemnify our users when they use our products. Im not going to open-source any technology, or introduce any product in to the marketplace, that I cannot protect at that level.
I know a lot of other vendors who have kind of skipped over that and said it doesnt matter. Thats not our customer base. So I want to be sure that for every product we deliver—whether it is Java or Solaris or any element of our microprocessor code—that we will open-source all of our products, we will open-source them under OSI-approved licenses, we will work with the community to establish the most effective governance and cooperative development model, and then at the end of the day, I will protect our customers.
Can you give us an idea about what software is next in line?
The Java platform seems like its kind of topical. Were looking at engaging with the community to pick the right license. Theres all kinds of choice we have in front of us. … Were going to continue looking at what the right options are.
The net of all this to Sun? The market is definitely growing. We can see the opportunity is definitely growing. The world is not going to consumer fewer network infrastructures next year than they did this year. The whole world is tipping toward the Internet. To me, that means the whole world is tipping toward Sun.
Check out eWEEK.coms for the latest news, views and analysis on servers, switches and networking protocols for the enterprise and small businesses.