RealTime IT News

Halsey Minor, CEO Grand Central Communications

Halsey MinorGrand Central Communications CEO Halsey Minor has a mind to rewrite the book on application integration.

The executive wants to replace traditional integration methods with network services that pipe software to companies over the Internet, saving programmers from having to rummage through messy strings of code.

That's the message behind Grand Central's latest Business Services Network, a platform that helps applications get to customers over the Web faster. Before Grand Central and with the software market booming around him, Minor created on-demand customer relationship management specialist Salesforce.com in 1999.

While the traditional software market boomed around him, Minor predicted its imminent collapse, believing that customers were better off soliciting software as a service instead of a cumbersome glut of infrastructure that may not work when it's set up.

Though enterprise software sales took a nose dive in 2001 and 2002, emerging technologies like on-demand are fueling a resurgence. So confident is he in the promise of on-demand delivery that Minor recently threw $50 million of his own money into an on-demand venture capital fund.

With a refreshed network services platform in tow, Minor recently discussed his vision for on demand with internetnews.com.

Q: When did you first realize that delivering software and services on demand had such potential?

In 1999. In the middle of the software boom I became utterly convinced that the enterprise software model was about to be broken because I thought people were squandering all of their money buying these $100 million, big bang projects. So I decided to do something about it.

In 1999, I became the founding investor and board member at Salesforce.com. In December 1999, I left to start Grand Central, which delivers integration on demand, recognizing that if my original bet in Salesforce was right, and that applications were going to migrate outside of the enterprise, that there needed to be an on-demand integration offering that tied them all together.

The third step is this on-demand fund to support entrepreneurs who really want to make their companies and industries more efficient by building a set of processes on top of Grand Central.

Q: How has the software industry changed to the point where you see a new paradigm on the horizon?

It used to be that you could buy a capability in only one flavor called software. Now you can buy it in two flavors -- software or service.

I think the big question that every CEO or CIO is asking is -- with this very limited budget and with about 63 percent already fixed -- the question is: How do we get maximum business value from a very finite budget? What people have started to say to themselves is, for a commodity infrastructure, "we shouldn't be building it ourselves."

The thing that became really clear to me -- and the reason I got involved -- was that the whole "pay and pray" model that I was watching in 1999 is not sustainable; it is simply coming to an end.

Q: What is the "pay and pray" model?

We're talking about companies that want to do an integration project. So you take your sales and put them in your accounting practice so every time something is sold it shows up in accounting. You've got to buy an application server; hardware from HP; a database; Sun Solaris to run on top of that; Netegrity to come in to provide some sort of directory or connectors.

You're going to spend all of this money and you're going to turn it on and hope it works. And I started going against this model because I knew a lot of people would turn it on and it wasn't going to work. And guess what? They've lost huge amounts of money. I'm a business person first, so I asked myself: How many things can I buy where I can go spend $50 million if I'm a big company or $2 million if I'm a small company and I get zero back? It's enterprise software.

If I go out and spend $100 million and start building and building and it starts to be the leaning tower of Pisa and falls over, at least my insurance company will pay me back. There's no insurance in the technology industry. If I were an insurance company, I wouldn't insure these projects either. The whole "pray for success" model is over.

Q: How does Grand Central's model for on-demand integration work?

What we're challenging the software industry to do is move to a "pay for success" model. The way it works is all of our customers come in and build a set of processes and connect the endpoints together. It's not until they start sending more than 25 megabytes per month of data across the network do they start paying. Our cost scales proportionately to their success. The more phones and computers you plug in, the more you pay. Pay as you go. We take all of that infrastructure, we run it for you and you pay for what you consume.

My metaphor is that of a shared infrastructure, whether it's the telephone grid or the mail service. I named the company Grand Central because the railroads seem like the first modern example of shared infrastructure. You can walk into GC and go anywhere within the 48 contiguous states. You just pay for where you're going. You ride on a train you didn't build or buy on the tracks you didn't lay with an engineer you didn't hire. GC is the core for running these services. Our customers add more and more endpoints to our network.

Q: You just refurbished your core network services platform. What's different about the '05 version of Business Services Network?

'05 has a directory and facilitates policy and enforcement -- with reporting APIs based on Web services -- and those connected can consume services. The enablement to have a shared infrastructure, developed communally using an ID system as a key, is completely novel in the integration business. We're building a network with more and more companies to add to it.

Q: How are Web services impacting traditional integration rules?

Web services is commoditizing integration. Every level of the stack is based on a standard: WS-Security, BPEL, SOAP, XML.

The question is: If we're delivering a standards-based solution, why would you be in the business of running commodity infrastructure? Why would anybody in the business of buying an Oracle database and running it, or buying hardware and software which goes bad all the time, run commodity infrastructure? Do you think anybody gets any competitive advantage by buying and running enterprise CRM internally? Can you create a use case where somebody should spend 10 times the money and take on way more risk to do it because they're going to get a competitive advantage? No. You take on so much more risk to take on these enterprise projects.

Q: So even the newer, on-demand model is a commodity business?

We're in the commodity business. Every day we think about how some brilliant engineer at Microsoft or IBM can give us a new standard that we can implement that solves more customer problems. At the end of the day, what people are really going to want is if there is no differentiation between software and service; service wins because you don't have to deal with it. We operationalize it. If it all goes to zero in value, there is still value in managing and running it. That's what we're driving toward: the commoditization of something that is being commoditized anyway.

We're living in a world where more and more companies are waking up and saying: "Do we really have to build expertise in integration?" Because if the answer is yes, the failure rate is so high. A couple of years ago, some guy from CitiBank was telling me that he found that 21,000 endpoints needed to be integrated in his system, but so many of these projects have failed that they just gave up.

The analogy I use is: it's like unemployment when unemployment gets really bad. You have a bunch of people that just quit looking, so the unemployment number is artificially low because there are those people who just don't look anymore. I think the number of endpoints out there is staggeringly larger than people can really describe.

Q: What is the biggest challenge Grand Central faces, considering the new model?

Security is the big question everyone asks because everyone adopting the model doesn't want to be embarrassed. We're still, I believe, at a very early phase, and so we have to work a lot harder to overcome what is an inherent bias in our model because it's new. That's one of the questions that comes up over and over again so we have to be substantially better than our best customer on it.