Sun CTO Greg Papadopoulos on Google's soft underbelly, Microsoft, HP and other topics

Sun CTO Greg Papadopoulos on Google's soft underbelly, Microsoft, HP and other topics

Summary: Sun CTO Greg Papadopoulos shed some light on a number of topics, including Google’s soft underbelly, Microsoft, HP, Dell and IBM in an interview at the Vortex 2005 conference.

TOPICS: Oracle

Sun CTO Greg Papadopoulos shed some light on a number of topics, including Google’s soft underbelly, Microsoft, HP, Dell and IBM in an interview at the Vortex 2005 conference.

Papadopoulos "speculated" on the where the Google/Sun relationship might end up: "If you look at what is possible now in the Google-like experience, when you start to add in search, portal and services…and the incredible experimentation takes place inside Google model…if you come back and look at center of it, what do you have to add to get to the desktop PC experience? It’s a collection of services from Net--email, photos, office productivity."

He described Google’s soft underbelly as having mostly anonymous users in contrast to eBay, which has a strong identity system and network of introductions and rating of identity. I’m not sure if he is implying that Google will move quickly into developing a social network and commerce engine, but Google Base  and Google Purchase (a PayPal competitor) points in that direction.

In addition, Papadopoulos said Google lacks software distribution, which Sun is helping to alleviate by  distributing the Google Toolbar with its Java software when people download the latter from Sun's Web site. 


Vortex 2005 co-host John Gallant and Sun CTO Greg Papadopoulos

On the hardware side, Google has several hundred thousand to a half a million servers in its grid, Papadopoulos said. Sun’s Opteron systems and new processors, such as Niagara,  are an "appealing set of hardware technology," and delivering software as a service with display-over-IP thin client technology works well with that model, he said.

Papadopoulos also talked about Sun’s vision for the future, moving from shrink wrap software to services and from client/server to utility computing. Reflective of the shift to new network-centric models, he said: “I want to be the Google of something, not the Microsoft or SAP.” Perhaps, but so far Sun is ahead of the curve with its vision of commoditizing infrastructure and trailing in terms of generating growth in its business, compared to Google, Microsoft or SAP.

The $1 per CPU utility, which Sun--as an infrastructure provider--is doing to stimulate the market and show how the business and computing model is possible, is still mainly confined to technical workloads. “I’m not so concerned with what’s happening now from a business model sense. We aren’t going to make a lot of money in the near term. What matters is the hardware and software stack, how to do it securely and efficiently. It’s also a long lead time to get ISVs there,” Papadopoulos said. He is also banking that the million-fold increase demand on the Net from sensors networks will be an area where Sun sell its infrastructure.

On the subject of Microsoft making a transition to services, which Ray Ozzie discussed at Vortex yesterday, Papadopoulos said that he wasn’t sure that Gates and company could transition its business model, which depends on lucrative upgrades to Windows and Office. “[Microsoft] will project software as a service in the Microsoft image, ultimately tying things back to the data model and identity system in .Net and ultimately in Longhorn. [Microsoft] holds the data, index, provides security and access—it’s the centerpiece of Longhorn, which is a closed and assured information platform.”
He also believes that the commoditization of computing, which has the virtue of transparent pricing (e.g., $1 per CPU), and the associated utility computing lower switching costs, doesn’t favor a Microsoft’s more closed system and higher margin model. Sun’s model, he said, is to drive commoditization, which can also mean lower margins.

Bill Coleman, CEO of Cassatt and a former Sun executive, said that in his experience CIOs are concerned about Sun’s economic model and view the company as a tactical, not a strategic partner. He countered Papadopoulos’ notion that computing will be a commodity like oil or electricity. He believes that computing will be more like the appliance market, with a few suppliers living off of gross margins below 20 percent. He asked Papadopoulos how Sun could afford its R&D (currently $2 billion annually) and distribution costs at that low margin, and surmised that Sun could go the way of SGI unless it has either dramatic increases in volume or dramatically lower R&D and distribution costs.

Papadopoulos disagreed that commoditizing computing would result in a 20 percent margin business. “There are a lot of rewards for getting over solving complexity problem,” he said. “We should collectively be embarrassed about how bad it is in computing, the operational complexity and power wasted. Of the $2.5 trillion [spent in IT], only a few hundred billion are for hardware and software.” 

In the short run, rewards exist for large and small companies that can automate IT.  Longer term, those who can deliver a more complete solution will reap more of the rewards, which is how Sun currently differs from Cassatt, which has software that automates application infrastructure management on Linux and Windows systems. Margins will vary depending on service levels and other kinds of services built on top the commodity infrastructure.

Papadopoulos offered his take on how IBM, HP and Dell fit into the emerging world of services and utility grids. He suggested that IBM Global Services would have to get rid of a lot of people who currently “support complexity” when commoditization occurs. 

HP lacks the more complete stack to compete. “At some level in the computer systems business, [they] have to decide if they are going to be a broker and follow Dell, who has perfected the model but lag in the next wave of innovation, or have enough control of systems, operating systems and middleware,” Papadopoulos said. “HP assets are diffused enough and they are not in control of enough to have complete solution.”

Dell is weakening in higher end servers because of a divergence from building desktops and servers similarly and focusing on processor speeds. “Now it’s about server power consumption, floor space, manageability and serviceability, areas where Dell is historically weak,” Papadopoulos said. In contrast to Sun, HP views having a desktop and server business as a strategic asset, especially on component cost, according to Ann Livermore, executive vice president of HP’s Technology Solutions Group.

The numbers for Q2 2005 tell a somewhat different story today. According to Gartner, for x86 servers using Intel Xeon or AMD Opteron, HP had revenue of $2 billion, Dell $1.3 billion and IBM $1 billion. Sun, a laggard in the x86 market, had $108 million, but grew revenue at 192 percent. Sun’s Galaxy servers are impressive, but it remains to be seen how much share the company can take away from others.

Papadopoulos also took IBM to task for its open source strategy. "IBM uses open source as a foil...if they are really serious [about open source] where are WebSphere, Tivoli and DB2  It's bull. Everything [at Sun] is going open source," he said. However, Sun has a lot of sorting out to do to move all of its code into open source. "It's more of a physics problem. It took a long time to get Solaris out. The physics is around IP concerns that go in it. There are over a couple hundred different copyright holders in Solaris, and we had to negotiate with each of those copyright holders."

Open sourcing is a good strategy for Sun in its quest to be the commodity infrastructure providing for the expanding Net, but IBM is generating lots of revenue from its proprietary software. If there were a diminishing return, as in owning an operating system, IBM would be glad to join Sun in the open source everything model.

Topic: Oracle

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


1 comment
Log in or register to join the discussion
  • Commodities sell on the basis of price.

    If everyone is selling the same thing, then anyone ordering would take the cheapest.

    Sun is a small company, needing more revenue. If they succeed in commoditizing what they are selling and they have much larger competitors, they'll lose.

    Otherwise, they have to distinguish themselves on features, and will be able to retain a comparatively small niche.

    Sun's R&D expenditures are under attack in the stock market, and a defense based on commoditized products is going to be ineffective. Dell's model is suited to a commodity product.

    And then, not expecting revenues soon from Sun's current strategy? How much patience are shareholders expected to have?

    Predictions of the future are sales projections. Sun isn't predicting much success for itself, is it?!
    Anton Philidor