NEW YORK—What comes after open-source software? Open services, said Sun Microsystems Executive Vice President and Chief Technology Officer Greg Papadopoulos.
"After open-source software is open services," Papadopoulos said at a Sun Worldwide Education and Research Conference here March 9.
Open services provide freedom for service developers and represent a place where open-source-based services can be shared and executed, Papadopoulos said. In an open services development scenario, developers work on common patterns for service levels and management, he said. And developers will "not just publish code but prototypical services, too," he added.
Indeed, open-source development will move from participative communities of code developers to participative communities of services, Papadopoulos said, noting that the University of California at Berkeleys Reliable, Adaptive and Distributed systems laboratory, or RAD Lab, in which Sun, Microsoft and Google each invested $500,000 to launch in December, is an example of an environment where open services can thrive. The three companies plan to put $7.5 million into the RAD Lab over five years.
RAD Lab officials said the Labs researchers will focus on developing alternatives to traditional software engineering, which follows a "waterfall" model of development. In such a traditional system, work is completed in orderly stages starting from system concept to development, assessment or testing, deployment and operation.
Papadopoulos said as far as timing goes, he expects that networks of open services will be prevalent in three to five years.
For instance, five to 10 years ago, "if you walked down Sand Hill Road—where the VCs [venture capitalists] are—everybody was making pitches saying Im going to be the Microsoft of this or that area," Papadopoulos said. "Today theyre all saying Im going to be the Google of this or that area. Very few business plans are coming in saying Im going to do the bits."
Meanwhile, Suns role in open services is to drive open communities for what Papadopoulos calls the New Stack. Part of Suns contribution to drive adoption of open services is its Open Solaris, Solaris Enterprise System and Open Sparc, he said. And the Sun Grid offering can deliver scalable infrastructure.
The New Stack is the environment upon which open service development should take place. The stack consists of scalable hardware—storage, servers and network—at the lowest level of the stack. Above that is a level of virtualization. Above the virtualization layer is a set of containers or core services, such as database, network area storage, directory, process grids, application servers, messaging, Web servers and streaming, Papadopoulos said. Above the containers will be thousands of services, including CRM (customer relationship management), ERP (enterprise resource planning), CAD, auction, gaming, desktop and others. Finally, above that layer will be the layer of service networks, he said.
Papadopoulos said innovation above the virtualization layer will occur separately from that below it and will focus on improving the predictability of service levels; simplifying distribution, integration and management; and improving developer productivity through integrated tools and service orchestration.
Winston Damarillo, executive chairman of LogicBlaze, of Marina del Rey, Calif., said open services development requires an assembly of multiple development teams. "We think the services development model is very different from the traditional model," Damarillo said. "It requires a development model that accommodates multiple contributors … and supports iterative building and a constant cycle of developing and testing. Its also designed so the integration of software shows how it is used and whats in it."
RAD Lab officials noted that critics say the traditional waterfall model is often too slow and therefore obsolete for the high-paced Internet era. Instead of infrequent, well-tested upgrades, code for Internet services is continually being modified on the fly as the product is scaled up to accommodate millions of users, RAD Lab officials said. Indeed, this fix-it-as-you-go feedback loop enables speedier deployment, but it also requires a large technical support staff to make sure operations are not disrupted as bugs are resolved, officials at the Berkeley lab said.