No one believes Patrick Grady at first. Why should they? He comes out of nowhere, radiating confidence, claiming his company, Rearden Commerce, has pulled off the IT coup of the new century. His triumph: a working b-to-b marketplace, fronted by an ultracustomizable application and based on SOA (service-oriented architecture). Delivered through the browser, Rearden's EBS (Employee Business Services) is capable of automating the purchase of many everyday services, including shipping, conferencing, meals, entertainment, and even travel.
Yet Rearden's lineup of corporate customers is enough to make a believer out of anyone. Officially launching today after five years in stealth mode, Rearden has already snagged Cingular Wireless, Genesys, JDS Uniphase, Motorola, Warner Home Video, and Whirlpool, all of which have signed up for enterprisewide deployments of EBS. This SaaS (software as a service) application goes where no enterprise software has gone before: to control spending on non-PO (purchase order) services, all according to identity-based business rules. Moreover, Rearden has partnered with American Express and Hewlett-Packard, which will resell EBS worldwide.
The roster of Rearden vice presidents and technical advisers is stunning. During the past year, senior executives from Hewlett-Packard, Sabre Holdings, Salesforce.com, and Siebel Systems have joined the company. Advisers include Jon Bosak, one of the creators of XML, and Adam Bosworth, vice president of engineering at Google and former chief architect of Microsoft .Net. "I think this is new," Bosworth tells me. "I'm a fan."
How did Grady get such prestigious friends? After all, he's not the first to propose a platform for interenterprise Web services connections. What's special, in Bosworth's view, is that the company designed EBS from the top down, by creating a new class of a service-purchasing application first, and then building an identity-based Web services platform to support it. Essentially, EBS is a container of application services with no vanilla version. By default, customers use EBS' application framework to whip together -- without coding -- browser-based purchasing applications tailored to their company's business rules, processes, and employee roles.
The result is a services marketplace that wraps itself around the identity and permissions of individual business users, automating, centralizing, and controlling purchases that in most organizations remain slapdash. Carriers' shipping prices can be compared side by side, air travel rules can be enforced at the user level, audio- and videoconferencers can get the best rates, and so on -- all through a unified Web app that runs on anything from a desktop to a smartphone.
Grady claims companies can expect a 20 percent reduction in hard costs and much greater savings in process overhead. And because EBS is a multi-tenant SaaS application, every time a new service provider plugs in to Rearden's services grid, it becomes instantly available to all Rearden customers. Theoretically, the platform should be extensible to a vast array of services.
Is this Kool-Aid for real?
To jaded IT types, the Rearden story sounds too good to be true. Toby Redshaw's first take was typical. As vice president of IT strategy at Motorola and an SOA evangelist, he thought, "If this existed I'd know about it, but on the off chance it does exist, there's no way they got the architecture right." So he sent a trusted colleague, whom he calls the "Mikey" of his group, to investigate. Mikey liked it. Today, Redshaw is both a customer and booster.
"They've built an architecture which is maybe the purest SOA Web services platform I've ever seen," Redshaw says, extolling the platform's inherent extensibility. "Think of any service that I can procure today through (EBS) and just change the noun. There are conference services, consulting services, real estate, fix-and-repair services, warranty services -- all of that stuff. We spend several billion dollars a year on services."
Deb Stanton, general manager of global procurement at Whirlpool, had reservations about going with an unknown company. "You have to think about it hard," she says. But she wanted to bring off-PO spending under control and felt the business case was very strong. The clincher was what she calls Grady's "passion and vision." Stanton is rolling out EBS across her organization.
The most important deals, however, are the ones Rearden has with AmEx and HP, both of which will resell EBS worldwide. The deal with HP goes a step further to support HP's BPO (business process outsourcing) initiative, where HP shoulders nonstrategic business processes for its customers. "HP plans to extend the Rearden Commerce platform to create some uniquely focused BPO-based solutions," says Bob Schultz, HP's vice president of BPO.
To better understand what Rearden has wrought, consider that the original Web services concept descended from two very different schools of thought. To some, Web services were primarily an Internet extension of the component development model; to others, they enhanced XML's promise to link businesses dynamically, allowing them to swap XML messages over the Internet instead of using dedicated EDI links.
Rearden's platform and its EBS application blend both lines of thinking. On the one hand, EBS' application framework is entirely component-based, so nontechnical customers can drag and drop services together to create end-user Web apps. On the other hand, Web services provide the basis of Rearden's services grid, which is essentially the supply side of a b-to-b marketplace.
To top it off, Rearden takes a page from another early Web services scheme: Microsoft's ill-fated HailStorm initiative. That was the code name for Microsoft's extension of Passport, intended to hold profiles, preferences, credit card numbers, and more. HailStorm may have met a grisly fate, but the generic notion of a secure identity whose home is the Internet -- one independent of device, that would be a hub for opening a widely distributed array of Web services unique to the user's identity and to the transaction -- that was a neat idea.
Rearden has rolled the .Net-dependent HailStorm idea to a Java platform and has adapted it for business users, whose roles and identities are uploaded from customer LDAP directories, along with any policy information. When all that has been uploaded to the system, an individual's identity determines the scope of business functions he or she is authorized to carry out.
To pull this off, Rearden Chief Architect Satnam Alag created an XML schema for services to provide a common denominator for service providers as well as the service components inside the application framework. To coordinate those components, Alag also created SBL (Services Business Language) -- a homegrown version of BPEL (Business Process Execution Language) -- to support event-driven composite applications. All that tough development work was necessary for Rearden to create EBS' console, which is one of the first graphical, nontechnical development tools for building useful composite enterprise applications.
Add it all up and you have a new type of user-centric application for delivering services, big corporate customers ponying up user identities, an SOA application framework tuned for services delivery, and dozens of service providers -- from Airborne to Zagat -- already plugged in to the grid.
Still, any new application category is risky business. Grady could fail to meet his stated goal of achieving the network effect this year -- where EBS has so many users it becomes irresistible to service vendors -- or there could be a backlash among business users who balk at having their expenses so closely monitored and controlled. But if successful, Rearden's platform and EBS application will deliver tremendous value to the enterprise and maybe just provide a blueprint for the future of Web services and SOA.
Patrick Grady's calculated debut
How did Patrick Grady manage to build his service when others have failed? How did he draw in big-name customers? In addition to his forceful personality, 10 years in high-tech venture capital gave him extraordinary access. In the early development phase, for example, senior technologists from Ariba, BEA Systems, BellSouth, CommerceOne, Genesys Labs, Palm, and Sun Microsystems got together once a week to advise him on architecture. That lends some credibility to Grady's claim that his platform will become "the global de facto standard for how you describe and discover and deliver and transact for services."
Why are you coming out of stealth now?
PG: The only way I would come public with this in detail is when I could establish proof across three areas. One, that I could build the entire stack: the integration layer for suppliers, the orchestration layer, the application layer, a real platform that could scale -- which we've done. Two, the initial application alone would yield very large subscription dollars in the enterprise and midmarket and would allow me to build a very large business on its own -- which we've done. Three, this ecosystem that I've been positioning from the onset would actually begin to take root. I'm launching with two partners, AmEx and HP. Now, I'm a pretty calculating guy; if I'm launching now, it's probably not much of a leap that there will be other partnerships in the not-too-distant future.
I can see that the cost-control potential of your service would make it attractive to the business side. What's the pitch to IT?
PG: Take a company like Cingular. It had 15 disparate applications before we walked in. They just bought AT&T Wireless. Talk about a heterogeneous hair ball of applications! All of those can be sunsetted now and migrated to one on-demand application. They don't buy it, they don't maintain it, they don't support it, and because it's all Web-services-based, they have forward migration.