e-speak talks the talk: Doing for Services What the Web Has Done for Data

Underlying HP's e-services strategy is a vision of a transformed Internet. At the core of this vision is e-speak, which acts as a dynamic interface, allowing services to come together without human intervention.

Underlying HP’s e-services strategy is a vision of a transformed Internet – one that is no longer a collection of massive Web sites accessed with a PC. Rather, the Internet has evolved into a network of interconnected services.

E-services combine "on the fly" to solve a problem, meet a need or complete a task. Computing has also been transformed. It is no longer simply a process of giving instructions to your computer, but has become a means of enlisting services as you need them. In other words, you tell your computer what you want done and let a dynamic federation of services figure out how to do it.

At the core of this vision, and HP’s e-services strategy, is e-speak, the company’s Java-based technology. e-speak is an open software platform designed specifically for the development and intelligent interaction of e-services. e-speak acts as a dynamic application interface, allowing services to come together on the fly to achieve the end a user desires, without human intervention.

e-speak operates independently of programming language and operating system platform, and enables spontaneous and secure transactions across firewalls. It also scales to millions of devices. HP believes that the technology does for e-services what the Web has done for data. Web sites share a common language – HTML – for creating and describing content, but until recently, there has been no lingua franca for services. e-speak, HP claims, makes it just as simple, or simpler and safer, to create, deploy, manage, personalize and access services as it is to publish and access data on the Web.

There’s nothing quite like e-speak. Not Sun Microsystems’ Jini, which enables device-to-device communication within workgroups, where scalability and security are not much of an issue. Not Microsoft’s BizTalk, with its low-level messaging server. And not Chai, HP’s own technology for device-to-device communication. Each of these technologies serves a purpose, but one complementary to that served by e-speak.

Creating No-Cost "Killer" Apps

When Rajiv Gupta and his colleagues in HP Labs developed e-speak, they had a clear goal in mind. The engineers wanted to help developers "create no-cost killer applications to create services rapidly," says Gupta, now General Manager of HP’s e-speak Operation. The goal was ambitious. It’s fairly easy to provide a service directly to a user, but it’s an entirely different matter to provide a service that can be used by another service, or to link services. Web standards, like HTTP, HTML and XML, are complementary, but operate at the protocol or language level that is a level below service interactions.

If Gupta and his co-workers were to develop a successful solution, it had to meet a number of criteria. First, the solution had to operate in a heterogeneous environment, one embracing a diverse collection of platforms and standards, from computers and PDAs, to cell phones and soda machines. Second, the service architecture had to encompass the entire range of functions and interactions required by e-services, along with a level of service metadata that would allow for the brokering of Internet transactions, composition of complex services, and autonomous application-to-application interactions. Finally, it had to resolve issues of naming, security and scalability arising from the shift from the workgroup level to the Internet, where service providers offer constantly changing services.

The modular e-speak architecture meets all of these needs. First, it supports a heterogeneous environment because it sits on top of standard protocols, like UDP, WAP and IR, and it allows services to define their own interaction semantics. Second, the services architecture allows e-speak to act as a service mediator, facilitating the brokering of services and managing the interaction among them. Finally, e-speak was architected with naming and security models specifically created for the Internet and peer-to-peer operation that allows it to scale to extreme sizes.

When developers program to the e-speak services specification, they don’t have to worry about the complexity of the Internet. A developer can write an application and invoke the services needed in terms of e-speak abstractions. When the developer is ready to open it up to the Internet, he doesn’t have to modify the application to allow others to use the service.

Speaking the Language

e-speak is made up of four main components: a universal services interface (APIs that provide a standard way of transacting with services across multiple application areas), a runtime deployment engine (the e-speak engine), pre-made components (building blocks for creating new services), and a services development kit.

The e-speak technology stack rides on any transport, while the e-speak engine provides the basic functionality. On top are the programming models, document exchange, and direct messaging. The e-speak framework is the basis for the exchange of application-specific data. The e-speak service bus is critical, defining a set of interfaces that allows independently written services to be composed into higher level e-services. The e-speak bus architecture also defines a set of infrastructure services, such as authentication, authorization and others.

e-speak differs from traditional middleware. With middleware, the client goes to a directory service and obtains contact information for the service provider, which it then uses to invoke operations on the service. This is how Jini works. But, there’s a problem here. If the owner of the service wants access control, billing, management, and so on, the service provider must deliver it. What happens, then, if the service provider is a printer? With e-speak, the client obtains a name binding and the request is mediated by one or more e-speak engines. The engines provide the access control and generate the management and billing events, and there’s no need to place the code into every service provider.

The Opportunities for e-speak

HP is not selling e-speak, so it will not make money directly from the technology. That leads to the obvious question: Why did HP, which has traditionally kept its innovations close to the vest, open source e-speak? There’s a simple answer. HP will exploit the opportunities opened up by e-speak by, for example, selling computers and services to companies that use e-speak in their own e-services initiatives. But, the opportunities extend far beyond this.

Gupta compares HP’s position with e-speak to that of a mythical company that has developed asphalt. If the company gives away the formula for making asphalt, roads begin to sprout up everywhere. New business opportunities are created, because there must be companies to maintain the roads, lay new ones, and so on. Moreover, cars begin to travel the roads, and that creates a need for gas stations and motels. The inventor of asphalt can exploit all these new business opportunities – which is precisely what HP hopes to do with e-speak.

Of course, HP is not planning to go it alone. Partnerships are core to the company’s strategy for e-speak, just as they are to HP’s e-services strategy as a whole. Moreover, some of HP’s e-speak partners are heavy-hitters in the industry.

One is Oracle, which is collaborating with HP to integrate the Oracle 8i database with e-speak. "The database is part of the e-speak infrastructure," Gupta says. The e-speak repository uses a database to store service and user information. Users do not access the database directly but access its functionality through e-speak interfaces. One reason HP went with Oracle, Gupta says, is that e-speak can operate with any database built to the Java Database Connectivity (JDBC) standard. "Another is that Oracle is evangelizing e-speak to developers," Gupta points out. With the partnership, HP and Oracle are hoping to attract a broad base of developers, including the 500,000 already connected to Oracle.

HP also has attracted Ericsson and Telia as partners in a project to run Wireless Access Protocol (WAP)-based mobile Internet e-services on the e-speak platform. Telia is deploying the platform in its mobile network, and Ericsson is contributing its WAP solutions, application development, and WAP-enabled terminals. The first of a series of e-speak-enabled wireless e-services offered by the three companies will be an automated scheduling service and a corporate directory service for a select group of medium-sized engineering firms that maintain fleets of field engineers. Using WAP phones and GSM-based positioning, the field engineers will be able to receive and update customer job information while on the road, as well as access corporate phone and database directories. Corporate databases will still remain behind firewalls.

e-speak received a big boost through another partnership. ERP giant SAP has promised to include e-speak in mySAP.com offerings. HP is also planning to leverage e-speak in its HP Rapid/WEB product that provides integration for linking SAP/R3 with the Internet.

In another collaborative project, HP is working with Helsinki Telephone to develop a multimedia brokering solution for corporate training services. The solution will use of e-speak to facilitate discovery of new training courses, match customer needs with training courses, monitor course utilization, and provide a billing and payment infrastructure.

e-speak is also being integrated in a number of HP technologies. It will play a key role in a variety of specialized Internet portals HP is developing in such areas as procurement and supply chain management. HP will first develop the portals, then add e-speak to all the trading communities and incorporate it into each portal. HP is also integrating OpenView, its enterprise-management solution, with e-speak to provide an out-of-the-box manageability platform and make OpenView the manageability backbone for customer implementations of e-speak.

Last September, HP announced the integration of e-speak with Chai, its technology for device-to-device communication. This combination of Chai and e-speak allows intelligent appliances to give users instant access to personalized e-services.

Clearly, HP is banking on e-speak to drive its e-services strategy. And while e-speak has commercial value, Gupta hopes its implications for HP reach far beyond that. "With e-speak," he says, "HP – a company that invents – is back in the saddle, back in a leadership role. That gives us mind share."

-- Jean Nattkemper is the Editor at Large for HP Professional, and has been reporting on HP for the past several years, writing extensively on HP’s Internet-enabling technologies. She can be reached at jnattkemper@101com.com.

HP’s Open Source Strategy for e-speak

HP’s e-speak platform is available on a site hosted by Collab.Net (www.e-speak.net). Offerings include a universal services interface (a set of APIs) and a runtime deployment engine (e-speak engine). The e-speak engine software was released under a combination of the GNU General Public License (GPL) and the GNU Lesser General Public License (LGPL) on December 8, 1999.

Collab.Net’s site also provides several well-known packages of open source software: bug tracking using the BugZilla software from the Mozilla Project; mailing lists, archived and searchable; source code browsing tools; and an administrative interface for "admins" from the e-speak community.

Those interested in seeing e-speak put into practice can download offline demonstrations of programs using e-speak. One such program, for network document printing, uses e-speak to locate the nearest printer and automatically convert documents in Word, PDF, and other formats to printable form, eliminating the need to install drivers on every computer. A virtual file system, allows for transparent storage and retrieval of files from multiple network servers. The program provides update notifications when a file has been added by someone else and allows users to arrange and name files independently of other users. From Windows, users can employ the visual browser to automatically launch the appropriate program to edit or view each type of file.

An HP-managed site, www.hp.com/e-speak, provides access to the e-speak Developer’s Program and registration information for it. At this site, you can also find information on HP’s Developer’s Conferences, intensive e-services conferences for Internet developers.