Adesso Looks to Mobilize Application Development

In 1995, shortly after IBM Corp. purchased Lotus Development Corp., IBM chairman Lou Gerstner appointed John Landry vice president, technology strategy. His mission: evaluate disruptive technologies from edge companies so IBM could fit them into some of its more prominent corporate strategies. But that wasn't enough disruption for Landry. That same year, during the first joint IBM-Lotus sales and marketing meeting, Landry unexpectedly came on stage dressed as Eva Peron to entertain the group while Gerstner, his new boss sitting in the front row, perhaps wondered what he had got himself into in entrusting so much responsibility to Landry.

Over the past several years Landry has helped launch a number of small software companies, most touting products that threatened to derail IT law and order. His latest venture, Adesso Systems, is no exception. The technology allows IT developers to create, customize, and deploy distributed and mobile applications as quickly as low-level Web applications. Landry sat down with Ed Scannell, Redmond editor, this time in men's clothes, to discuss his venture.

Redmond: Where did you get the idea for Adesso?

Landry: The initial idea was to build an architecture that exploits some of the big economic trends such as cheap storage and connectivity, although sporadic connectivity. We already did that in Notes, which was based on a replication technology. So given that I know where all the bodies are buried in replication architectures in general, we decided the right architecture for this is a replication model and to build it on a fully distributed database architecture that sits on top of existing database management systems. The distributed logic can run on top of existing platforms but not be part of the existing platform, allowing us to move information around in ways not possible before.

What is the secret sauce involved in the distributed logic?

We have essentially mapped the global unique identifiers [GUIDs] into those [underlying] databases so we can identify what data is underneath. We can't depend on the naming the database gives us. So hooking up to something like Oracle and SQL [Server] Express is fairly trivial for us. With this distributed database layer you are able to deal with phones and PDA databases, laptops and desktops, all of which can participate equally in this distributed architecture.

Do you find that many users and developers are ready to implement this sort of distributed approach?

Given distributed is a tough concept for people to think about, we tried to extract out just the things they really need to build apps. In our definition that includes tables, forms, views, filters and components. The idea is to provide a design capability that allows you to very rapidly build apps that take advantage of a distributed architecture right out of the box. If I build an expense report system on Saturday, I can replicate it out to all users on Sunday.

Others claim they can do something similar to this. What is unique about Adesso?

For us, doing distributed is not just a matter of moving the data, it is a matter of also synching and replicating the design, access control rules, schema, and content control rules of the database. And this has to happen just moments before the data is moved. That is where it gets really hard.

What tools are you making available for developers?

We have gone extreme on this. The tool is also distributable, it is part of the product. So if the administrator gives you rights to design an application, or just part of an application, with the tool built in you can create or change the design of something quickly and synchronize those changes to all users of that application. They receive those changes the next time they synch in. One of the goals is to make distributed mobile architectures as easy to build as an Excel spreadsheet. Developers can fly at any level [of the architecture] they need to so a rookie user can build a simple app and experienced developers can build and extend it out with C# DLLs or whatever they are using.

This product appears to do some of the things that Microsoft's WinFS file system was intended do?

It does. Think of this as WinFS today. At its core [WinFS] is a hybridization of database technology with files. The idea [with Adesso] is to let the distributed data base functions be attached to the file system meaning I can correlate a file with a record in the database that Adesso is managing.

Have you shown this to Ray Ozzie at Microsoft?

Ray likes it a lot. We have a good relationship with Microsoft and it's getting even better.

You're essentially talking about marrying structured and unstructured data with this product?

Yes. Using what we call crackers, we can "crack" the metadata out of files and map it to the relational database. We can then use the Views capability in the product on that data to organize, sort and sequence it, as well as use that data as a vehicle to synchronize intelligently. The idea is to have an intelligent distributed file system hooked into distributed database functions.

Will you be positioning this as something that can compete in the composite apps market?

Yes. Any app you build in Adesso, you can then take pieces of that application like a form, view, or plug-in and turn them into components and export them into any other application. The more apps you make the more components you have and the easier it is to assemble applications. In some cases you can put together a sophisticated business app in 30 minutes.

What is Adesso's business model? How will you sell this, direct or through partners?

This will be a big Web play. The development platform we are giving away. You can download it off the Web and start developing apps right away. If you want to put that application back up on the Web where other developers can access it and blend it in with theirs. They can go through our Web site [Adesso Now Environment] to do that. You can also host the application you created there. We will charge you to do that. That is how we will monetize our intellectual property to make money for ourselves and the developers. The idea is to offer the development environment, hosting environment, and billing environment all in one. We'll even send you the check every month. The idea is to let all flowers bloom.

About the Author

Ed Scannell is the editor of Redmond magazine.


comments powered by Disqus

Subscribe on YouTube