RealTime IT News

Greg Papadopoulos, CTO, EVP of R&D, Sun

Greg PapadopoulosSun Microsystems is undergoing a dramatic transformation. For the last few years it has moved away from some of its proprietary approaches toward a more open source approach to its tools and technologies.

At its recent JavaOne developer conference, Sun finally made Java more palatable to Linux distribution by changing the license under which Java is distributed.

Solaris now has an open source project in OpenSolaris and is contributing an increasing amount of open source efforts in Solaris.

Sun's software emancipation extends across it's product lines including its Java Enterprise System and related tools.

At the helm of Sun's technology initiatives for the past eight years has been CTO Greg Papadopoulos. During his tenure, he has helped chart the company's technology direction, architecture and standards as the IT industry, and therefore the company, was undergoing rapid change and tremendous challenge.

With the recent promotion of Jonathan Schwartz to CEO of Sun, Papadopoulos has now also been given some additional responsibilities, which he explained to internetnews.com.

Q: What is the most exciting piece of upcoming technology innovation from Sun in your opinion and why?

The Sun Grid. It really is, finally, computing as a utility -- just like electricity.

You don't have to build your own power plant; you just pay for what you use. If you have a Web browser and a credit card, you have access to supercomputing power, at $1 per CPU-hour. Simple as that.

We build and manage the power plant so you don't have to.

One of the great things about the grid, aside from the ease of use, is that it will enable new kinds of communities to form. Most communities today are built around the notion of: "Here's my program. Download it."

But the Sun Grid provides a place where people can go and say, "Here's my code actually running. Go interact with it. Add value and prosper."

It's a place where anyone could go and create a Monte Carlo analysis of financial models, for example, test it, and actually deploy it commercially, essentially renting the infrastructure instead of building it.

Q: What non-Sun-originated technologies do you think hold the most promise inside current Sun initiatives?

As we look at the software industry today, we are seeing two trends: open source, where software developers are sharing bits (operating systems, Web servers, browsers, etc.) and software as a service, where companies such as Google, eBay, Amazon, PayPal, pull together bits and deliver a service.

The blending of these two technology trends -- what I call open services -- is where the industry is headed.

In order to have open services, however, we have to insist on open, standardized interfaces. By open and standardized, I mean that the creation and maintenance of the interface is democratic and the licensing terms do not preclude anyone from implementing the interface.

You know an open standard if the economic and social positives are there: multiple, competing implementations from a range of vendors and a low switching cost for consumers.

Here's how I see it working: If you create a service and people become dependent on that service, then others must be able to inspect your interfaces to create an alternative service.

In that way we'll be able to create healthy competition and not new monopolies. This will enable the new kinds of communities I mentioned earlier -- communities built around compute power rather than code.

And each service will become more valuable as additional services are able to interact with it.

Q: What do you see as the risks of Sun's open source initiatives? Or is it all good?

Managing the interface is one of the toughest challenges: keeping critical interfaces open to encourage competition and stimulate innovation.

We're very concerned about the impact of proprietary intellectual property on the global standardization system, and we're concerned that this hurts innovation and access for both developed and developing economies.

I believe, for instance, that it should be exceedingly difficult to gain patent protection on an interface. At the very least they should always be subject to reverse engineering.

We have to say, look, every electrical appliance gets to include a standard three-pronged plug. Nobody gets exclusive rights to that.

We have a stack -- servers to Solaris and the Java Enterprise System -- to allow developers to create open systems built on open standards and open source. Not open sourcing would greatly limit the markets we can participate in, the communities that we can participate in.

You don't build communities by building walls of intellectual property.

Q: What do you see as the biggest technology challenges for enterprises in 2006?

Energy consumption has to top the list.

People around the world continue to join the network at an incredible rate, services are proliferating, and data centers are growing to meet the demand. All of which is great, but I'd hate to see Google's electric bill. Actually, I'd love to see it, but would hate to pay it.

We need the equivalent of the Energy Star label for servers. A good server, after all, is always on -- unlike your dishwasher.

The good news, I'm pleased to say, is that Sun has some very power-efficient servers on the market now, based on both AMD Opteron processors and our own UltraSPARC T1 processors with CoolThreads technology.

Our Sun Ray desktops are also extremely efficient, each using just four watts of electricity -- 13 counting the server it's connected to.

I have also just hired David Douglas as vice president of eco-responsibility, reporting to me directly.

David will oversee environmental initiatives across the company, including enhancements in Sun's products in the areas of energy efficiency, cooling technologies and management of Sun's asset sharing through programs, donations and access to Sun Grid resources.

Q: What keeps you up at night? What worries you about modern IT infrastructure and technologies?

What worries me about modern IT infrastructure is that it is not as "modern" as it should be. There is still too much complexity and issues around the ability to get predictable service level agreements as we see scale of demand.

As companies abstract infrastructure to services, do they take the time and care to ensure replaceable components so that someone else can take over a particular function? Like what Salesforce.com does, for example.

What we are doing at Sun, which helps me sleep better, is the exciting work around Sun Grid and how we're taking the complexity out for customers.

Q: What's next and what are your goals as CTO for 2006?

For the past eight years ... I have been focused on managing Sun's technology direction, architecture and standards.

Our CEO, Jonathan Schwartz, just gave me the additional title of executive vice president of research and development and refers to me as the "chief systems architect."

In this role my goal is to build net-scale systems -- which is the new stack. I will be working directly with all of the chief technology officers within Sun's business units to drive efficiency and commitment to technical innovation.

Together we will build this one "product," i.e. new stack for infrastructure, that allows anyone to create and deliver the next generation of network services.

The pieces also need to be integrated so we get efficient, secure, compliant, predictable service levels that support people building network services.

So, my primary goal is making sure Sun is attacking the complexity of modern IT infrastructure and providing the building blocks of next-generation network services.