Lavastorm: An Analytic Power Tool for All Seasons
November 8, 2011 Alex Woodie
The folks at MDS Lavastorm Analytics have created an analytic tool that defies easy description. On the one hand, Lavastorm is a lightweight ETL tool for cleaning and validating data. But it’s also a graphical, drag-and-drop SQL editor. It provides classic reporting and dashboard functionality, but also detects fraud and other irregularities in billing systems. It’s a cross-reference tool, and a business rules engine too. It slices and dices–and does just about everything but make julienne fries. But perhaps most surprising of all is it’s commonly used in IBM i environments.
Traditionally, business intelligence tools are used by non-technical business users to generate reports and dashboards that tell them how the business is doing. On the other side of the spectrum are IT power tools that technology professionals use to keep information systems running smoothly. “What we’ve tried to do [with Lavastorm] is connect those two products together, and cross that gap,” says Tim Dunne, the managing director for MDS Lavastorm Analytics’ North America and EMEA regions.
In many respects, Lavastorm is all about validating the dynamic flow of data among applications, as opposed to taking snapshots of data sets at any point in time, which is the classic use case of BI tools. Call it a business performance analytics tool, with a heavy emphasis on addressing an organization’s tactical operational issues, as opposed to crafting big strategic plans.
“It’s not really about looking at static views of process, but understanding how do you connect data processes across the business, how does data flow through the business, as well as understanding anomalies of the data and validating business processes as the data flows through business,” Dunne says.
Lavastorm was founded by a group of MIT engineers in 1993, and has been used heavily in the telecom space. In fact, the tool is a regular in the hands of the IT professionals of MDS Lavastorm Analytics’ sister company, Martin Dawes Systems (MDS), which has a very successful business selling access to its home-grown IBM i-based customer care and billing system to communication service providers.
As Dunne explains, Lavastorm’s capability to discover how data flows between systems has been especially useful for tracking down problems that crop up in customer care and billing systems. For example, the product’s cross-referencing feature has been used to match data between customer care systems and order-entry systems, where it can easily spot what orders have been placed but haven’t yet been provisioned.
“With one action, I can do a cross reference, which would be 20 lines of code in SQL. I can drill down and open up a data viewer, and look at what data is in there, and understand, why are these data records in provisioning but not in the order entry system?” Dunne says. “You can look at the data and understand what’s happing and create ongoing controls to prevent that from happening again.”
One of the largest cable companies in the U.S. saved $17 million by using Lavastorm to get a better handle on work orders, according to Dunne. “They were having a lot of trouble with orders, rolling trucks before the actual order was validated,” he says. “So they did an analysis and created a control, where we actually audit through the tool every order they do in the consumer segment, every night in batch environment. Then it auto-corrects the order problems, and passes the orders on to the provisioning system.”
The software is also used as a front-end to more traditional business intelligence tools. “One way to think about it is we’re a lightweight ETL tool, but we can manipulate and transform and enrich data pretty easily,” he says. “We have many customers that use our tool to clean data, create understanding of what they want to get out of the data, and then feed Tableu or Qliktech to do the presentation layer or dashboard.”
Lavastorm’s heritage is in the telecom space, but the product is finding more uses among customers in the healthcare, utility, financial services, and government arenas. One use case in the electric utility industry involves applying Lavastorm against information that’s coming out of smart meters. In this case, the smart meters are so new that a utility may not fully understand how they can make use of the data.
As Dunne explains Lavastorm can work with the data from the smart meter, along with other data about network capacity on the network. This would enable the utility to perform simulations “to decide if they should buy more energy off spot markets for the day, try to store it, or try to stimulate a marketing program to try to have consumers not use dishwasher during the day,” he says.
“By building analytic and simulation models, where they’re bringing multiple data sources in, correlating that data, and building rules in a structured format, it can help them to understand what’s happening with their system and help them make business decisions.”
It all sounds quite complex, but Dunne insists that Lavastorm’s GUI drag-and-drop interface and use of “nodes” with pre-defined functions enables non-technical users to make good use of the software. A power user who is proficient at using macros and pivot tables in Excel could learn the software. For more advanced functions, more experience with SQL is beneficial.
Lavastorm ranges in price from about $3,500 for a desktop version to more than $150,000 for a server implementation. The software runs on Windows, Linux, Solaris, and HPUX. For more information, see www.lavastorm.com.