Cover, Irish Computer, June 2001 Download the PDF version of this article (670kb)

Epionet: The Web Servants
By Laura Coates

Amid the dot bombs and software slip-ups, one Irish company has kept its head down and worked away on development. Now Epionet believes it has a world leader in web services management technology.

An Irish company has recently started moving at full speed into the marketplace with its web services development and deployment product.

For the past four years, while the country has been awash with dot com silly money, this company's three founders and their 21-member team have been keeping their heads down and avoiding publicity, developing a technology that other big players are now just starting to think about.

The company's small offices above an off-licence in Marino Mart in Fairview are crammed with developers working on the product. The atmosphere is caffeine-fuelled and energetic, but friendly and the walls in the entrance hall are decorated with snaps of the last staff outing to Amsterdam - a raucous affair, judging by the pictures. Epionet as a company doesn't stand on ceremony. Yet the technology that is coming out of this small Dublin headquarters is impressing some very big players.

In a recent review of Epionet's web services development platform, commissioned by a potential investor, the Gartner Group said "there is no competitor with a similar offering on the market".

What Epionet has developed, and what Gartner called a "truly remarkable achievement", is essentially an open-environment based on XML/XSL (extensible markup language/extensible stylesheet language - the latter a new technology influencing the presentation of documents) standards. EpioRADD (rapid application development and deployment), as this is called, lets businesses quickly build and easily maintain component-based systems using web services. These pieces build up into an e-business system, accessed by the user through a web-interface on the enterprise portal. It makes it quicker and cheaper to get a b2b site, or an internal system, up and running, and in-house technologists are not bogged-down with mundane web development problems. The environment comes with its own raft of application services and builders, context analysis methodology, system services, a deployment application and an implementation framework.

Web services are currently being predicted as the way all system deployment will be achieved in the future. Small 'lego bricks' of functionality are loaded across the web as a result of a user request. This functionality could be something as simple as 'open new mail message' in an e-mail program or it could be requests to retrieve project data or company information. The service performs the specific task requested of it and disappears. EpioRADD leverages off the Microsoft .NET initiative, which provides the fundamental infrastructural technologies that clear the path for web service implementations. The key to the system is an efficient web services management system, which can control the development, deployment and management of thousands of these services. This, the small Dublin company believes it has perfected.

Epionet differs from the other software start-ups in that it has had from the start an experienced management team who Gartner praised for their realistic commercial expectations. Founders Stephen Brogan, CEO; James Brennan, vice-president, strategy and finance, and Ambrose Curry, vice-president and chief technical officer, met through a mutual business acquaintance and formed the company in 1997 with the intention of being involved in the business-to-business space.

Brogan, who has a science, engineering and MBA background, has over 20 years experience in IT management and business development. A Siemens Medical veteran, he also worked with Trinity Institute (TCD's executive management institute) and Technico Communications. Brennan is a business graduate and a certified management consultant and was a partner in the corporate finance department of BDO Simpson Xavier. Curry, a DCU engineering graduate, has worked for Siemens in software development and also delivered projects in the pharmaceutical and airline areas. Liam McMahon, who joined the company last year as vice-president of business development and strategic alliances, also has an MBA and engineering background. He was formerly vice-president of industrial solutions for CNH Global with responsibility for the global supply chain, and was also managing director of the Irish operation. He has also worked for Perot Systems and Siemens.

Early progress

Epionet has had two development stages so far. In 1997 it started up by taking on individual business-to-business implementation projects. "Our goal during that phase was to build up the company in terms of its technical capability and capacity with the skills that we needed for business-to-business projects," explained Brennan. "We built up a database section, a business process section, a multimedia section and a web technology section."

Epionet did projects for Bertelsmann, AIB Bank, Modus Media and Buildonline to fund the research and development the founders realised they wanted to undertake.

"What we found when we started doing business-to-business projects was the hardest thing about implementing them was not making the business application work but making the web technologies work," said Brennan. "So many web technologies have to work to allow even the simplest of applications to run, and you need specialists in every single one of them, and those specialists are very expensive and very scarce. And we found we were solving the same internet technology problems over and over and over again and that the applications were quite straightforward.

"So we sat down with those principles of simplicity, lack of complexity, quicker development in mind, and we came up with the product which we then spent three years developing. It is basically an architecture that has embedded within it all of the internet technologies - HTML, SOAP, Java - everything is built into it. So that if a programmer concentrates on building the business logic of his applicational system in any language he likes, as long as he does it within the architecture it will run on internet even if he knows nothing about the internet. And that's at the simplest level."

"I've been involved in IT for twenty years in designing and implementing systems," said Brogan. "James has been using the technology for nigh on twenty years and the other guys have been developing the technology for twenty years. And I believe as an IT strategist that in all the technology being delivered to clients, it could be delivered better and we set out to achieve that goal."

The company recently secured £2 million in first-round funding (£1.5 million from the Alliance Fund and the rest from private investors) in return for a ten per cent stake, valuing it at about £20 million although it has no confirmed customers yet for this product. Epionet had been in conversation with potential investors from the early stages of EpioRADD development, which was finished at the end of last year. The initial development was self-funded from early customers. "When we started developing the technology, we deployed half our staff on projects and half on development," said Brennan. "But in 1998, we had a lot of cash reserves at that stage to take us quite a long way, and in fact they took us all the way through to end 2000, and we put everybody from that point on development."

This latest round of funding should take them through to early next year, and will be used primarily to push the technology out to the marketplace.

"We believe we've got the best technology available, now we've got to prove that the technology delivers what the market actually wants," said Brennan. "There's two ways we can do that: one is convince one of the major players - the Microsofts, Oracles etc - that this is the best technology and to partner with us in delivering it. We believe that the business to business model has not taken off to date, because of the complexity of the technologies, the scarcity of the specialists in the area, the long lead times and the failure of existing technology to deliver the goods. We believe the market is still in the early adopter phase where people will not go and webify their whole systems. They'll say 'lets webify that bit and see how it goes'. We believe that in tandem with talking to big players we are going to go out and win some of those pilot projects, which will prove to the marketplace that this system actually works and delivers the goods."

The company will be trying to crack American markets as quickly as possible, as it believes this is the route to quickest success, even if pilot projects are European-based. It is currently holding board-level talks with its target partners, appearing at exhibitions and conferences and generally trying to get a buzz going around its technology. But why keep its development nearly secret for three years, when every other tech start-up with a fresh or half-baked idea was screaming from the roof-tops? The company founders say they did not want to sell the promise of a product, but a system that worked.

"We were talking to one of the major players three years ago and said what we were going to do," said Brennan. "They said that this could not be achieved technically. And we said: 'we'll see you when we have it done'. And it took three years of development work by 24 people. So there's an enormous amount of design and intellectual capital gone into this."

"We come out into the marketplace, we're a small company, we've got something like this - people will be sceptical about it," said Brogan. "They are sceptical. In one of the companies we attended the guy said: 'you don't mind if I say this, how does a company with 24 people create this' - at the time, actually we had 14 - 'how does 14 people come up with an idea that everyone else is hunting after?' We didn't enter the market earlier because we knew people wouldn't believe we could do it. Today, we don't need to tell them, we can show them. Today we can go in with the best tool a salesman can ever have - we have done it."

New approach

In traditional web-based systems, the business logic functions, the presentation layer and the data layer are all interconnected. This makes it work seamlessly, yet conversely, makes it less scalable. "The bigger the system becomes, the more difficult it becomes to maintain and upgrade. You can't re-use any components, it's very slow to develop for and you need teams of specialists," said Brogan. "We split the business logic layer, the presentation layer and the data layer absolutely and completely.

"When you develop a web application, there are certain things that you need that we believe are common to all deployments. By creating this environment here we are providing all the web technology, the security and the enterprise portal so when you create something you can plug it into the environment and not worry about the web technology. We've talked to a lot of technologists over the past couple of years and some of the big players told us what we were trying to do could not be technically achieved."

The EpioRADD environment has benefits for both developers and users. Its creators claim that it reduces demands on a company's skills and resources by up to 50 per cent. It also vastly reduces deployment time. Web developers using the environment can start with a blank application, build up some core functionality and deploy it on their system straight away. These applications are 'plug and play' in that the underlying web technology does not have to be designed or tested, as it is already present in the environment.

"That way you build with the business leading the technology, and not the technology building the business," said Brogan. "From our own experience, one of the systems that we created for a client [pre EpioRADD] took us eight and a half months with a team of developers, in the old days. Today we can deliver this system in less than a week. It's chalk and cheese."

"One of our senior XML programmers developed a quite simple program - to look up your e-mail server, see what e-mails are there and to deliver them back to you," said Brennan. "Written in XML that was six or seven pages of code. It took a week to write and test and implement and when we stored it the computer stored it at 14K. One of our junior programmers could do the exact same thing with no loss of functionality with EpioRADD. It took six lines of code, took about 10 minutes and was stored at 400 bytes."

The system is XML/XSL-based, meaning that it is totally open. It can link into any legacy system - even older systems that have a simple XML-interface written for them.

Epionet made a judgement call to develop on XML. "We didn't pick XML because we thought it was going to be the standard for the future," said Brogan. "We picked XML because we believe it delivered a business advantage to the companies that would use our systems to create the environments they needed to sustain their competitive edge.

"We believe that if it doesn't fit into their current infrastructure, then people will not buy it," he added. "Ultimately all companies put investment into their history, into the systems they've developed, and when it comes to a certain stage what you're going to do in the future inevitably has to link back into the past." The system can sit on top of any ERP system or database, and customers can even switch database types without users noticing any operating differences. The web-based environment means user-training is minimal.

The web services architecture also means that users don't have to know where internal documents are coming from in order to see them, and applications are seamlessly and intuitively linked. The system is scalable infinitely, according to Epionet.

"We believe today that we have a product that is of significant value to anybody involved in web development," said Brogan. "And we've been told we're ahead of all the knowledge space." This first mover advantage puts Epionet two years ahead of the major software corporations, and ahead of its nearest direct competitors such as US-based BEA, Bowstreet and Web Methods. It has a number of prototype sites with pilot customers on the go at the moment, and is actively looking for more.

"At the end of the day, the ultimate vision is that this becomes the development platform of choice for systems on the internet," said Brennan. "And that we have a global network of partners who are developing systems for their customers using our technology."