Data Mashups vs. Data Warehouses

The DM Radio Webcast, “The Last Mile: Data Visualization in a Mashed-Up” from Information Management continues. The transcript of that Webcast, which was hosted by Eric Kavanagh and included InetSoft's Product Manager Byron Igoe and BI consultants William Laurent and Malcolm Chisholm resumes below:

Byron Igoe (BI): With the typical data warehouse scenario, if one thing goes wrong, the whole system comes down to its knees.

Eric Kavanagh (EK): That’s right.

BI: There’s a huge dependence on the IT side before any other business can get done. Whereas if you flip the situation on its head and focus more on the self-service, you get the users to respond and react quickly in tune to their own needs and then worry about the performance of those queries after the fact with appropriate technologies like column-based databases, temporary caches, even data grids for processing things in parallel, something like a Map/Reduce technique.

EK: Yeah, so let’s walk through this. What would the ideal scenario look like? So for example instead of having your traditional array of OLAP cubes that you have built on top of a data warehouse for example, and maybe even sometimes bringing in additional data through data federation or something like that, would you have essentially instead of that, you would have an array of data marts that would be dimension-specific, or how would that work?

#1 Ranking: Read how InetSoft was rated #1 for user adoption in G2's user survey-based index Read More

Bring in the Other Atomic Data Sources

BI: Sure, that’s an option. But you can also even bring in the other atomic data sources whatever they may be. A lot of people with service-oriented architectures have services providing various data feeds. Of course you have transactional operational databases and of course data marts, warehouses, and the like. Even external feeds from vendors can become one of those sources. So essentially providing just the individual sources to the user in a user-friendly way and really allowing them to drag and drop, combine, and manipulate data sources together on the fly to get exactly the results that they want, I think that is really the idea.

EK: Yeah, and let’s bring in William and Malcolm. William, what do you think about that idea?

William Laurent (WL): Yeah, the federated model is something that I have actually written about, including writing specifically about InetSoft. Yeah, the way we traditionally look at the data warehousing world in which we go through these costly ETL processes to extract, transform, to load the data, and a lot of the logic is quite frankly encapsulated, not transparent, you don’t know the business rules. So with the federated mashup, I see the potential. I am a believer in the potential for the federated mashup and just the sense that we are able to pull data from various systems dynamically and look at the data, have semantic virtualization that makes sense. So you don’t have to mandate consistency because quite frankly people are looking at visual representations now of the data, and they can best understand what it is.

Read what InetSoft customers and partners have said about their selection of Style Scope for their solution for dashboard reporting.

Previous: Implications of Mashup Technology for Enterprises