Eric Miller of Zepheira, who is working on a new open source semantic data integration tool presented in the SOA track. Rob Drysdale has provided this blog entry.
Eric covered the synergy between web community and how they have been dealing with data, searching, combining and how it should be done in the enterprise. He also touched on the lessons of the we are applicable to managing enterprise information.
Most of the web is pages and links, designed for direct human consumption, limited linking (pointed at through URLs), data is hidden and siloed (similar to the enterprise).
The web is becoming one of the most pervasive data management and integration platforms ever imagined...(Think of Google and mashups).
The Semantic Web is a technological way to support data reuse and knowledge sharing. It enables you to combine data and use it to make intelligent decisions. In order to do this you need a common model for representing the data, expressing constraints and using common query mechanisms to come up with a means to expose traditional data.
We need to take the lessons from the web and use architecture that is consistent. The focus needs to be on information and workflow and not on the implementation. These building blocks will allow us to take things and combine them in a way that we want that is specific to what we need or want. Eric used an analogy about Lego to illustrate this point that Legos are consistent building blocks that can be put together in any way based on the person.
Overall this will allow us to continually change how we look at the data and not have to continually version software to make all these changes. This will reduce cost and allow the data to serve the business and not just be part of the vendor solution.
Currently we all have many systems that track our data but it isn't integrated and can make it difficult to do real information exchange. Trying to put the data all in one big system (Enterprise Data Systems and Models) can be troublesome and cumbersome. This restricts agility.
If you look at the web it does not use any supermodel or consistent architecture, but it all works really well together. So you need to look at it as if you have a web of data in your enterprise. So don't look at it as a supermodel, but as a need for a consistent way to link it find it and represent it.
Re-use, Re-purpose, Remix
They have used open source tools to integrate data and be able to get data integrated in organizations.
Exhibit allows you to map out information and present it (think Google maps) in a way that people can interpret easier. Once the data is all in Exhibit you can actually start to put it together by pointing at multiple data sets. It doesn't organize it specifically, but allows the user to quickly do it and try different things.
Simultaneous Editing (tool) allows you to edit data quickly.
Eric showed an example of two Excel spreadsheets being combined using something like Exhibit and combining it to allow the user to manipulate it and look for patterns and overlaps. You can use a Excel spreadsheet to drive a web page type of view.
Overall you don't change the data, but you stitch it together using these tools so you can use it to analyze it and make decisions about it. Many of the tools that Eric talked about are open source, but used in a different way. Basically you're mining the data and putting it together to do other things with it. Once you connect it you still may want other tools to do some of the final layer of consistent reporting and management of the data, but the tools allow you to get the data together.
Eric gave suggestions on how we can take this back to our own organizations and try this. Basically we need to analyze our problems and see how we could do it and start with something small and try it and see if we can build up momentum and support.
_________________________________________________________
The Enterprise Web of Data
Eric Miller
President
Zepheira
Wednesday, March 19, 2008
08:30 AM - 09:30 AM
Level:
Intermediate
What if you suddenly had the ability to integrate large data sets…
- while your business sponsor is standing over your shoulder?
- at one tenth the cost of traditional mapping methods?
- without requiring widespread agreement on naming conventions and data definitions?
- and you suddenly had new insights into data relationships and patterns that nobody had been able to see before?
Yes, it all sounds too much like Silver Bullets falling from the sky, but in fact there is a new class of tools and technologies that changes the old rules forever. It’s based on semantic web technologies, and it allows you to do all of this. The paradigm shift at work here is that we need to starting thinking about Enterprise Data in the same way that we think about the world wide web. This session is a glimpse into the future - right now!
Eric Miller is the President of Zepheira. Prior to founding Zepheira, Eric led the Semantic Web Initiative for the World Wide Web Consortium (W3C) at MIT where he led the architectural and technical leadership in the design and evolution of the Semantic Web. Eric is a frequent and sought after international speaker in areas of International Web standards, knowledge management, collaboration, development and deployment.