Web 2.0 has evolved into an indispensable tool for millions of daily Web users. But who are the people driving this technology into the Enterprise? Steve Perry is not your typical Web 2.0 visionary. He is a veteran of the database industry. He installed his first database in 1996; an Oracle 7.3 database, on an NT system, for a Great Plains warehouse package.
The data sharing platform he has developed, called Jumper 2.0, promises to change much of the industry he has worked in for so long.
The idea behind Jumper is simple.
"Search any data, in any format, in any location was the very basic premise," he says.
"The general idea was how to search the increasingly growing volumes of structured data without having to manually query each local system, assuming you even knew where all your data stores were located."
"At the time, I was engaged in a large project at Genzyme. The volumes of data they were generating were stunning and they faced huge challenges managing this deluge. It was this experience that catalyzed the ideas for Jumper."
At the Beginning
"Jumper originally started as a project on the Jini Community backed by Sun Microsystems. Project Jump was a name server for storing persistent names and descriptive tags assigned to data objects."
"We created the project to run as a super-peer in the JXTA environment. When they couldn’t come to a consensus we formed Jumper Networks to build Jumper as an Enterprise server. JXTA has since fully adopted our core ideas into their platform."
"As data integration consultants one of the first things we always do in a project is schedule meetings with the DBAs and System Administrators. Only they can typically tell us what is actually stored in a legacy database."
"They knew all the detail; what the arcane column names meant, how the data was formatted, what the constraints were, etc. From this we would create a Data Architecture report."
"Early releases of Jumper initially only allowed the DBAs to tag metadata with information. It was fairly obvious that opening this console up to all users would allow us to capture far more than just the metadata," says Perry.
In Need Of Change
"This was a big driver for Life Sciences because the assays (or data output from experiments) are created at the R&D project level and sharing this data was difficult. This information is called 'Provenance' information. It is the who, what, when, where, and how details about the data."
"This matters in Life Sciences because if the experiment is successful and eventually becomes a drug they need a data audit trail. It also mattered because the R&D pipeline was getting longer and more expensive to develop new drugs. They desperately needed a new way to share research results across projects and disciplines."
"Jumper met this need by allowing the scientists themselves to bookmark and tag the data, describe what was stored, and give the thumbs up or down on data they find searching Jumper."
What makes Jumper go for researches sounds pretty simple.
"With hundreds of millions of data sources available across your site, partner sites, and the public Internet, how do you find what you need?"
"Relational tables and flat-files can be pretty arcane, they typically do not provide good annotations (documentation about the data)."
"A users visibility of the real data horizon is often very limited and manually searching the resources you know about is very time consuming."
"Jumper expands your visibility and eases the discovery process enormously."
