How would you like to build a global enterprise-scale data access infrastructure? A daunting prospect, yes? Imagine creating a system that could make any subset of any significant data resource in your organization available where it's needed without incurring insane implementation and maintenance costs ... sounds too good to be true?
Well, that's what Pneuron Corporation promises with its eponymous Pneuron Solution (the "P," which stands for "process" and "predictive," is silent).
It's generally reckoned that about 80% of IT expenditure goes to maintenance (a.k.a. "keeping the lights on"), which means that trying to add major business initiatives such as enterprise analytics or meeting new regulatory reporting requirements can be a huge burden on budgets. Pneuron claims its solution offers a 50% improvement in Time to Value and TCO.
[ PROGRESS REPORT: Building, and managing, the 21st century data center ]
As an example of the kind of problems Pneuron addresses, consider the challenge of building enterprise business analytics to provide a "big picture" view of an operation that's spread across a couple of dozen offices in half a dozen countries. This provides an entire obstacle course of headaches. For example, French law is très particular about certain personal data leaving the country, while data connections to, say, Kenya, might not allow for realtime database access.
To solve this problem enterprises have relied on techniques such as duplicating databases, leaving the source copy where it is and replicating the source to wherever the analytics are to be done. This might work for the inventory database in Kenya as long as you don't need the latest data, but it's not going to do for the human resources database in the French office where the data can't leave the country.
And one of the biggest gotchas with the ways these kinds of problems are usual solved is the maintenance nightmare that results from the "kluge" factor involved.
The Pneuron Solution provides a distributed data access, extraction and transfer solution based on Pneurons, "mini applications" that are deployed to one or more lightweight "Cortex" virtual servers.
The Cortex servers are Java VMs that create a cluster-able, load-balanced, elastic cloud model with fault tolerance and failover. The servers run on Windows (even XP!), various favors of Linux, AIX and Solaris, as well as on VMware vSphere, KVM, Oracle VirtualBox and Amazon EC2, and support MySQL, Postgres, Oracle 10g and 11g, DB2, HSQL and Apache Derby databases.
Many Pneurons are available, including Data Query which extracts data from SQL databases; Service to execute a Web service call; File to access any of a number of file types; Predictive which supports importing and creating PMML models; RulesML for rules-based processing; Matching which compares retrieved data to templates or other sources; and Analytic which provides graphical and tabular displays.
The Pneurons do their work gathering and massaging data and pass that to the Cortex server to be routed to another Pneuron running on the same or a different Cortex server. Data transfers between Pneurons is in an encrypted "universal" self-describing format based on XML.
To create a solution using Pneuron you'd deploy Cortex servers according to local control needs. For example, if French HR data was being used, a Cortex server would be needed in the French office configured so that the non-sharable HR data would not be available. Once the servers are installed and configured then, using the Pneuron Design Studio, you'd configure which Pneurons would run where and do what tasks.
Pneuron is a very serious, enterprise-scale solution for distributed managed data access and manipulation that is quite unlike anything else I've seen. The company is well-funded (it just got $6 million series B financing from Safeguard Scientific) and it has a slew of case studies that articulate the kinds of problems it can address. This is definitely one of the most intriguing enterprise products I've come across in a long time.
Gibbs is not distributed, he's in Ventura, Calif. Connect to gearhead@gibbs.com and follow him on Twitter and App.net (@quistuipater) and on Facebook (quistuipater).
Read more about data center in Network World's Data Center section.