Thomson Financial invests in Web services

THE FINANCIAL INDUSTRY is a traditional early adopter of new technologies, and Thomson Financial, a global provider of integrated information solutions, has been an early investor in the area of Web services. At a recent VeriSign news event, Mark Barbagallo, vice president of Distributive Technologies at Thomson Financial, met with InfoWorld News Editor Mark Jones to discuss his company’s commitment to Web services, the importance of building a business-client trust relationship, and the role of vendors such as VeriSign and Grand Central in helping Thomson deliver Web services to its clients.

InfoWorld: Give us a brief overview of what Thomson does and its involvement with Web services.

InfoWorld: Can you explain what the architecture stack is?

Barbagallo: We have an architecture council across all of our companies, and the architecture council’s mission is to set not necessarily architecture specifically, but policies and practices around architecture. One of the things that we do is get together and discuss standards and best practices and so on. Recently — this is [further] evidence of the [company’s] commitment on Web services — we’ve been told by our CTO council and our executive committees that Web services are a big deal for the whole corporation. So the technology leaders have gotten together and decided where the common ground is that we can all agree on and effectively published that to all the technology organizations. So within the bounds of a particular standard and particular protocols we’re all in agreement, and those are the types of services that we’re going to offer at this point. As more standards become agreed upon in the industry and across our businesses, we’ll continue to build on that stack. That allows us to drive leads within our business, but it also drives reuse in a solutions aspect, because now we’re starting to see cross-utilization of services between, say, Thomson Learning and Thomson Financial.

InfoWorld: Where does the committee stand on UDDI?

Barbagallo: That’s one of the standards we haven’t agreed on. In fact, this is a good example where the external standards-setting committee may be a little bit ahead of where the corporate mindset is. We haven’t really bought into the notion that we want to publicly publish all of our services. We’re not [even] sure that we want to publish them internally. We think that we can build practices to make people aware of [our services] that don’t necessarily have to revolve around those standards.

InfoWorld: What are your hesitations?

Barbagallo: There’s a lot of discussion about whether those approaches could be privatized, or federated, Web [services] vs. globalized. We have services that we probably would not want as a public domain but we might need to publish internally. The other thing is, in a business sense, we’re not ready to give up that direct relationship with the client. We’re afraid that in some respects if you just publish it out there and you completely enable this total e-commerce business-to-business, you take the salespeople and that customer relationship a little bit out of the business. We really value our customer relationships. I don’t say that those relationships won’t ever go away. What I’m saying is that we don’t recognize in the standards organization that they’re thinking about that part of our relationships with a client. The standards are pretty much [concerned with] nuts and bolts — discover it, acquire it, utilize it, pay for it — [but] there’s tremendous value in that relationship between our business and the client. The [business-client] relationship is all part of that trust relationship, and all part of that doing business in the relationship. We haven’t seen the standard yet that really aligns with our business model, so we’re not jumping on board with any of them yet. That’s not to say that we won’t. And in terms of directories, a directory is a directory. It could be something that you just look up on a database or it could be something that’s published in the order of UDDI or one of the others.

InfoWorld: Do you agree that standards and the automation of processes for exchanging certificates and so on can help drive costs out of the organization?

Barbagallo: I agree that those things will drive costs out of the organization. But I also agree that we ought to be able to choose which of those costs we want to drive out. One of the costs I’m not sure that Thomson is prepared to drive out of the organization is that relationship cost between their clients. Obviously there’s manpower involved there, there’s people that maintain those relationships and maintain the interface between the client and our businesses, [but] I’m not sure that we’re prepared to give up on that yet. I don’t say that we won’t get more efficient about it. I just don’t think we’re prepared to move that quickly in that part of it. But we think there are significant cost savings in doing services business directly with our clients. It provides us with opportunities to more intimately integrate with those clients and it’s a great value driver for us. It’s a value driver for the clients, it’s better for time to market, and it’s better for our technology relationship.

InfoWorld: Are you looking to the VeriSigns and Grand Centrals to deliver packaged Web services?

Barbagallo: No. We’re looking for VeriSign and Microsoft and Grand Central and BEA and those vendors to provide us with technical capabilities that allow us to deliver packaged Web services. In the end, it’s not going to be BEA and Microsoft and Grand Central who are the domain experts in our business. This goes right back to that personal relationship with our client. Thomson Financial is the domain expert, we are the people who know the financial business. So it doesn’t make sense for Microsoft to be building a bunch of financial-related Web services. It does make sense for Microsoft to be building technology, because that’s where their expertise is. We like to have ourselves viewed as a technology and information company, but we recognize that we’re not a technology creator, we’re a technology consumer and utilizer. VeriSign’s probably a better example than, say, Microsoft or BEA, because they’re more pure-platform providers. VeriSign has long delivered services and they’ve delivered complex services that are hard for businesses to engineer. So payment services, gateway services for recognition and identification — those are all valuable services that I think that we could use vendors to provide.

InfoWorld: What are you thoughts on VeriSign’s drive to build out the XKMS (XML Key Management Specification) standard?

Barbagallo: VeriSign’s got ideas around creating tighter identity services or the interactive identity services and things like that. Those are the types of services that I can see us going to outside vendors for because the cost to develop those things would be tremendous.

InfoWorld: How much of an issue is latency in transactions?

Barbagallo: Latency is an issue for some of our data, primarily real-time market data. The latency is prohibitive for us to deliver that [data] through a Web service, other than to maybe a financial portal like Yahoo where it’s 20 minutes delayed anyway and so the latency isn’t as much an issue. [But] for trading floors and things like that, the latency currently is an issue. But realistically, I don’t know that latency and performance issues are, over time, solved by efficient software as much as they are by improvements in technology. Bandwidth gets more available, data rates get higher, and I think that’s where the efficiency will come from. Cost is probably the biggest driver for all of these things. Eventually there becomes a point of diminishing returns where the necessity for the trader on the floor to have mini-microsecond access to market data is outweighed by the cost [to] the company to deliver that information to the trader. There’s also the government aspect: There’s been a gradual trend over the last 10 to 20 years to level the playing field for individual investors vs. institutional investors. So some of those advantages that the institutional investors and actual traders have had may over time [be lessened], because it’s seen as an unfair advantage over [individual] investors. If you buy a stock, you get a stock quote and your only access to those quotes is through a Web service that has some latency. [Institutional investors] have direct access and they can close or capitalize on a buy that you’re doing. That’s kind of unfair.

InfoWorld: Are you suggesting more regulation is needed?

Barbagallo: I’m suggesting that over time the government and the industry will flatten that [disadvantage] out, so some of those traditional aspects of the real-time market may lessen. Eventually the technology and the need for those things will resolve themselves, and then the latency issue will probably go away. There are plenty of technological approaches to eliminating latency. A lot of companies are starting to envision services as federated services. Maybe VeriSign builds a gateway service that does very sophisticated identification, but there’s no reason that we should necessarily have to build completely across the Internet to get that service. It might become a federated service where we have a private copy of it inside our firewall that does slower level transactions with a more broadly based service. But that could eliminate a large degree of latency in those services.

InfoWorld: Are we getting close to a federation model that will work?

Barbagallo: Everyone is different. There [are] certain services that will work in a federated model and there are other services that won’t.

InfoWorld: Given that we’re at such an early stage, how is the cost of delivering these services changing?

Barbagallo: We’re seeing the cost in delivery of services go down a little bit. The amount of technology in code that’s necessary to enable existing assets is getting less and less. So that’s positive. The cost of adoption for the client is also going down. As they become more and more acclimated to using these technologies, they’ve seen cost advantages as well. The only concern I have about the evolution of standards is that we do run the risk of having to do things twice or sometimes three times. But my experience in the financial community is that time to market trumps that cost savings. Being ready to do the business at the point where the client is ready to do the business far outweighs the additional cost of having to come back to maybe reapply a security standard later as the security gets bigger. And since the cost of implementing those technologies seems to be going down, the cost of doing it the second time or the third time is somewhat less anyway.

InfoWorld: Do the major vendors understand your performance metrics?

Barbagallo: We are making them understand. One of the things that we’ve discussed in the architecture council is the opportunity to use [our] leverage [as] a multibillion dollar company to influence the vendors. We’re not going to go to them and say, “We want you to implement this standard.” What we’re going to do is go to BEA and Microsoft and IBM and others and say, “We want you guys to get on the same page of standards. We think that’s very important. And if three or four of you decide to get on the same page and one doesn’t, we’re not going to view that very well.” I think that there is some subtle pressure that we can bring against vendors to make them interoperate a little bit better.

InfoWorld: How do you see that influence playing out in the political debates around Liberty Alliance and the WS-I?

Barbagallo: I’m not sure that the influence of technology users like Thomson and some of those companies really come to the floor there yet. I don’t think it will until we have some major deal someplace and we’ve got two or three of the vendors sitting at the table and we look at one of them and say, “We’d like to use you, but unfortunately you can’t get on board with these other two guys.” I won’t say that we’ve had any negative business dealings because of this, but I’ve been a little bit disappointed in Sun’s reluctance to get on the same page as some of the other vendors. It’s changed my view of Sun. It doesn’t mean that we won’t go out and buy their hardware and that we won’t utilize their technologies, but we do think a lot longer about it now. I think we’re probably not the only ones who would do that.

InfoWorld: How would you describe Sun’s approach?

Barbagallo: I don’t want to talk negatively about a vendor, but I think that their approach has been a bit clunky at this point. I think that they could have done a better job in their relationships with Microsoft and IBM and other competitors. It’s interesting that in the rest of the business world we constantly do business with our competitors. We don’t like it, but we do it because that’s what our clients want us to do.

InfoWorld: Sun would probably say in response that Liberty Alliance is their open approach to that issue.

Barbagallo: My response would be, “Why do I have to [join] another [alliance] just to influence [these issues]?” They could easily bring those issues into one of the other standards organizations and have them addressed in a slightly broader form and probably get their way on a number of those things. All of the vendors have gotten their way on some of the aspects of it. I suspect there aren’t any major board members in WS-I or W3C that are thoroughly disgusted about the way it’s going. If they were, they probably wouldn’t be board members.

InfoWorld: How is Thomson gearing itself to benefit from the growth phase that will inevitably follow this down economy?

Barbagallo: I think that the economic downturn has had some positive effects — it’s hard to say that, but it’s had some positive effects. People have had to focus more on cost, and since Web services truly do offer an opportunity for cost savings, it’s an area that consumers, as well as service providers, are now going to focus on. Thomson Corp. has a huge commitment [to] Web services. I would expect us to have hundreds, if not thousands, in the next few years. And obviously all of our services are around a variety of different businesses, a variety of different industries. But the key is that we’re positioned at a cusp now where if the economy does get better, we want to be able to rapidly take that up without having to add a tremendous amount of cost back into business. And I think our clients would say the same thing. So Web services are positioned now. With even minimal agreements on security, there are things that we can do today that we couldn’t do a year ago. And with higher levels of compliance and agreement and some standards in security and so on, there’ll be other things that we could do. So we’re positioned so that [when] there’s an upturn in the economy, there could be tremendous adoption of these things.

If you look at what we’ve done, 20 of our more popular data sets are already available through Web services. Most of these customers are already taking these data sets. They have nothing more to do than call us up and say, “We don’t want to take it through an FTP feed anymore or through a leased line. We want to do this through a Web service. Can you do it?” And we say yes. There are other products that we’re not quite there on yet. Some of it is because of legacy issues where they’re not easy to enable. Others are because there just really isn’t a demand for it, but we would expect to see a demand. I made a trip to Seattle to talk to a financial strategy consultant recently, and I was actually surprised because we went to talk about a new delivery alternative for a particular set of data that they’re getting, and they were very aggressive about talking about Web services vs. our traditional delivery mechanism. They wanted somebody to talk to them about potentials and possibilities in our Web services.

InfoWorld: Aren’t e-commerce enterprises, such as Amazon.com, looking to the financial community to start a more pervasive use of these types of services?

Barbagallo: Yes, the financial community is the traditional early adopter of these things. We delivered our first commercial Web service a year ago, so we’re actually making money doing this. And we expect to continue making money doing it.

InfoWorld: How much?

Barbagallo: Lots.