The Big Bang

As Covisint attests, building the"smarter enterprise" begins by replacing supply chain friction with harmony

By Covisint CTO Kevin Vasconi

Depending on your viewpoint, Covisint represents either the unlikeliest strategic alliance in business management history or the inevitable conclusion to the evolution of the traditional supply chain.

Perhaps both. Formed in early 2000 by the merger of Ford Motor Co., General Motors Corp., and DaimlerChrysler's two previously independent efforts to establish an Internet-based B2B exchange for the automotive industry (Nissan/Renault has since signed on as well), Covisint has a goal of staggering ambition: to serve as the common procurement and collaborative manufacturing and design platform for $300 billion worth of business. In essence, the company views itself as an open integration framework that provides a portal interface into the supply chain for OEMs, their suppliers, and even other exchanges. For that reason, Covisint may signify the "big bang" of business process integration: the first glimpse of what the post-Fordist, agile demand network will look like (see Resources).

Covisint is now an independent company that claims Ford, GM, and Commerce One Inc. (a strategic technology partner) as stakeholders. When it is fully operational - although it claims 250 customers, at this writing it has no permanent offices or CEO - Covisint will theoretically provide the supply chain agility, cost efficiency, and visibility that the industry covets to reduce costs and inventories as well as enable its transition to the build-to-order model pioneered by electronics components companies.

As Covisint's CTO, Kevin Vasconi is responsible for system architecture and technology development. Prior to joining the company, Vasconi was on the senior management team that formed AutoXchange, the Internet B2B joint venture between Ford and Oracle, where he also served as CTO. He amassed more than 10 years of experience in technical management at Ford and has also held technical and engineering positions at Allen-Bradley and GM.

IE: Can you give us a bird's-eye view of Covisint's IT architecture?

Vasconi: Sure. First, let me start by explaining the business problem we're trying to solve, which differentiates us from other B2B trading exchanges: the fact that we have to support extremely diverse customer environments. Most of the people on the Covisint IT team came from the OEM community - Ford, GM, DaimlerChrysler, and so on - which includes some of the largest companies in the world. Therefore, we intuitively recognize that these companies have IT infrastructures that literally include a little bit of everything. Because our business success depends on the ability to connect with such systems securely, we've got a big integration situation on our hands.

Therefore, our first goal at Covisint was to create an open integration framework. Unlike other exchanges, which built their architectures on a single technology provider or application, we determined from the start that our solution has to be big, scalable, secure, and capable of integrating with just about any software release that has ever existed.

Since then, we've proven that our architecture is a good integration framework. By the middle of this year, we'll have a secure, deterministic XML [extensible markup language] dial tone in place between us and all our founding OEM partners upon which we can integrate multiple applications running on disparate operating systems, provide a single view to the ultimate customer, and offer common services such as security registration, monitoring, and transformation routing and workflow in the back end. That's a compelling value proposition to our customers, who don't care what kind of technology we run at the end of the day. What they're trying to do is solve business problems.

IE: How does the architecture hang together?

Vasconi: When you draw it on a piece of paper in a logical diagram, it resembles a big "U"; in fact, we call it the U.

The top leg of that U comprises customer-facing services such as our portal and personalization engine, which includes security and registration components. Most of those components are built on the Oracle Exchange platform, but we've also built some of it ourselves. We're also in the process of integrating Documentum into our portal architecture to provide version-control change management and e-document capabilities. The bottom leg of the U consists of our intra-application messaging capability and our database platform, which is built on top of Oracle8i.

The back part of the U, of course, is where our integration framework lives. Our internal messaging bus is built on XML, and we support multiple flavors. But because the entire industry isn't ready to embrace XML for all their transactions, we also enable EDI [electronic data interchange] as well as very, very lightweight integration via HTTP and FTP. Depending on the application, the sophistication of our customer, and what they're trying to do, we offer an array of connectivity options, including a proprietary product we built ourselves that guarantees message delivery from point A to point B, or industry standards such as IBM MQSeries or Microsoft Message Queue Server.

Those options get us 90 percent of the way there; the last 10 percent involves integrating our message queue into the customer's legacy systems. I define "legacy system" as anything that's up and running; even if it's a state-of-the-art, Web-based app, custom integration is involved.

The center of the U is where the specific applications - for procurement, collaborative product design, e-catalogs, maintenance, repair, and organization (MRO) capability, reverse auctions, and so on - reside. We've built some of our own procurement products to complement those from Commerce One, and we're working on some future lines of businesses as well.

The key is that all those applications, regardless of what operating system they run on or what company developed them, are integrated into our framework, and the consumer sees a common way into them and a common way out of them. Suppliers only have to integrate their systems with one platform, and then we handle the transformation, routing, and workflow in and out of all those applications.

IE: Despite this array of connectivity options, how will you ensure that the Covisint platform "saturates" the supply chain to the second- and third-tier levels? Don't many of those companies still live and die by the fax machine, if not the floppy disk?

Vasconi: That's true. First of all, our applications are browser based. If you're a third-tier supplier that doesn't want to invest in a lot of technology, and your purchasing manager has a browser and a 1995 AOL account, he or she can still connect to the Covisint Exchange and accrue some value back to the organization.

Now, that's not true computer-to-computer integration: The manager obviously can't process 100,000 purchase orders a day. Nonetheless, that type of connectivity does enable smaller companies to get involved, to experiment, or to conduct certain business transactions with state-of-the-art technology that's not going to cost a lot of money. So we've really lowered barriers to entry, which has never happened before in this industry.

Surprisingly enough, that gives the big guys an advantage as well. If they're not sure about making a $100 million dollar investment in software, they can run a pilot in a 10-person department instead. If the value proposition develops, we can then implement a large-scale integration because there's no way the company can pump 100,000 transactions through the exchange without the scalability of computer-to-computer integration.

Basically, if you're a supplier that wants to get involved, you simply have to ensure you have a secure network connection and train the businesspeople who are going to use the system. We'll take care of the integration problem, make sure the software's up and running, and process your transactions or convert your EDI stream into XML.

This approach isn't perfect, but at least it starts to address the challenges we have reaching the "end tier." So far, the limited number of people we've talked to in that tier are pretty enthusiastic.

IE: Based on what you described, Covisint sounds like it's much more than a trading network. It sounds more like a portal into the supply chain, regardless of which procurement processes the consumer is involved in.

Vasconi: That's the vision. Now it's a matter of rolling up our sleeves and making it happen.

But here's the problem: Covisint is just one among many exchanges based on an XML messaging bus. As you know, there are at least five competing standards out there. "XML standard" is definitely an oxymoron.

IE: Let's talk about that challenge for a moment. What is Covisint's position on XML schema?

Vasconi: We don't view ourselves as a standards organization; in fact, we are absolutely committed to not creating an XML standard. We do need a standard for the industry, however, and ideally, I'd like to see an XML standard across industries. So we're having what I consider all the appropriate conversations with RosettaNet and the other standards bodies.

Now the knock on a lot of standards committees is that they're too slow and often impractical. That's what we're trying to add to the equation: We're trying to be the guys out here saying, "Hey, we're trying to implement this standard in production for real customers, and these are the things you need to do to make that a standard across our industry. And by the way, guess what? We only want one of them." In the short term, we've had to pick and choose among the standards that are out there. For now, we think ebXML is the way to go. [See the sidebar, "First Among Equals?"]

IE: Will Covisint XML messages emulate paper documents, or will they reflect a true data modeling implementation?

Vasconi: We're not quite sure. Some of the XML standards are definitely taking the paper-based approach, but that's not really very robust. You're probably going to see a combination of both approaches. Ideally, we would love to have XML schema based on a data model. I just don't know whether we're going to be able to totally achieve that goal.

IE: Do you plan to offer analytic services that would draw on the transaction repository to help customers optimize their supply chain interactions?

Vasconi: Analytic services will be a great line of business for us in the future. That being said, from our standpoint, customers own their own data. We're not going to take a couple of customers' data, aggregate it, and sell it to somebody else. No exchange is going to do that and stay in business.

Nevertheless, we can add a tremendous amount of value. Because again, our value proposition is an integrated application platform where your supply chain management application can talk to your procurement and collaborative product design applications. With that integration, you can do some data mining, in-depth data analysis, or even gaming-type analysis of the data you can pull out of there based on a customer or perhaps a series of suppliers. Perhaps some members that collaborate in the supply chain will all get together and aggregate their data. That's going to be a tremendously powerful tool.

We're not there today, and probably may not get there this year, but I definitely think that that's going to be a huge benefit to the industry.

IE: Covisint is a significant factor in the big automakers' transition toward the build-to-order manufacturing model. What problems will it solve in that challenge?

Vasconi: I'm a huge proponent of build-to-order. I think it's going to revolutionize the auto industry to the same extent the assembly line did.

We are a major enabler of build-to-order because we can bring multiple tiers of the supply chain to the OEMs, and ultimately, to the consumer-dealer relationship. The OEMs have done a lot of great work on their customer-facing Web initiatives, getting consumer demand online. But now they can't get all the way there without us or without somebody like us: Our ultimate role is to connect the consumer to the supply chain through the OEM and the dealer relationship. We can accelerate that process because we can serve as the common unifying force across the auto industry by lowering everybody's costs through economies of scale and the adoption of standards.

IE: But data sharing goes against the grain for most so-called trading partners these days. For example, BMW and Honda have identified security concerns as a reason for nonparticipation in Covisint. How will you break that cultural barrier?

Vasconi: It's all about education: We have to help our customers understand the difference between real risks and perceived risks, and that we have technology and policies in place to address those real risks. We also need to help them understand that some of the perceived risks are either no greater than those implied by the status quo, or that doing business electronically gives you a level of privacy or security that you may think you have today, but really don't.

It's a slow process because we have to win them over one by one. But I absolutely believe that the digital collaborative environment we are creating is more secure than the current environment. In fact, if companies were to closely examine their current business processes, most of them would discover that they're not overly secure.

Technology doesn't solve all those problems, but it can close some of the holes. It can obviously create some problems, too. But as a technologist, I'm confident that we can protect both the security and the privacy of our customers' data. My challenge is to convince them of that fact.

RESOURCES

Related Articles at IntelligentEnterprise.com: