Analytics Vendors Show There's Still Innovation in BI

Managing data isn't as glamorous as many of the glitzy Web 2.0 apps on display at the Demo Fall 2008 conference in San Diego, but it's what keeps businesses running. Enterprise applications to speed databases, manage unstructured data, and handle BI queries on a massive scale made their debut this week, and at first blush they look positive. Of course, all will have to stand the rigor of real-world use and testing.
Business Intelligence Meets the Cloud
Lately, BI vendors have been more likely to make news by being acquired than for technical innovation, so it's refreshing to come across a BI vendor that might actually be doing something different.
Quantivo combines what it calls an "affinity" database with a cloud-based infrastructure hosted by ... well, the company won't say, but Amazon is a reasonable guess. CEO Brian Kelly, an alum of Oracle and Sybase, says Quantivo is designed to recognize patterns and index relationships, rather than looking at data in a conventional row and column configuration.
One early customer, he says, was a retailer that exported some 2 billion records to Quantivo, which was able to get the customer up and running in less than a week. Kelly says a significant benefit of the product is its ability to recognize relationships within structured data. One example: drilling down into retail data to get a sense of which customers that have bought relatively inexpensive plumbing equipment are likely to purchase pricey items like counters or cabinets.
It's easy to make claims about business intelligence, and Quantivo is making some big ones. If it lives up to its founders' hopes, the young company has a chance to make a mark in a market that badly needs some excitement.
Structured Data Search Meets the Search Box
Much effort has been spent on making unstructured data -- such as the content of PowerPoints, call center logs, and PDFs -- available for analysis in structured, database-oriented BI tools. Radiant Logic has inverted that approach to make structured data available through the ultimate unstructured query tool: a search box.
Its Virtual Context Server analyzes various databases, applications, and other structured data sources and generates the relationships of meaning among the data -- that is, its semantics. IT analysts can modify and create their own semantics for data as well. The result: natural language text that users can search for, bringing that structured data to the fore in an unstructured search. Thus, an intranet search for a company will not only bring back the company's contact information from a CRM database, but also the current orders and payment history for that company from the ERP system, notes Michel Prompt, Radiant's CEO.
IT can also set access rules to the data, so database information such as pricing isn't available to just anyone who does a search। VCS also can be set up to let users correct information they get in their search results, and either update the source database or have the change stored in VCS so it is applied to future searches as well.


Unclogging the Database
Here's a situation that should resonate with experienced database users: Every morning a cadre of dispatchers log in to a company database and prepare to dispatch trucks and drivers on their way. The problem: With a large group of users making calls to similar tables or objects, the database bogs down at the worst possible time.
Eric Ruck, a long-time consultant and now CTO of Tetrabase, says his young company has a solution. Tetrabase SDT is an accelerator that improves the performance of installed databases. Although the product is compatible with all major flavors of databases, at this early point it works best with Microsoft's SQL server, says Ruck.
Simply put, Tetrabase adds a layer of management intelligence। Once administrators spot a problem (like the database bogging down every morning) they can use Tetrabase to find a solution. To make a complicated example simple, the trucking company's database needed to identify the data dispatchers were calling up each morning (usually orders tagged for that day) and load it before the dispatchers punched in.

Reference : http://www.pcworld.com/article/150907/bi_analytics_demo_2008.html?tk=rss_news