Data-driven businesses are facing some tough challenges in today's rapidly changing information landscape. As decision cycles continue to shrink, companies need to act on information within hours and minutes rather than weeks and days. At the same time, the volume of data that needs to be analyzed is growing exponentially. Business intelligence (BI) approaches that might have made sense a decade or even five years ago may no longer be the best fit for organizations that must quickly and affordably make sense of terabytes of incoming data that shows no sign of slowing down.

For my company, MX Force, speedy data analysis is not simply a "nice to have," it's critical to our business. As a cloud-based provider of email security for organizations of all sizes, we need to identify the origins of spam, viruses and other potential threats for our clients, fast. But as our business has grown, so has the volume of email log data that we must store, filter, search, analyze and report on. Recently, we were challenged to find a database that could reliably enable quick and efficient ad-hoc queries on up to a year's worth of email log data. Our staff uses this data to analyze and report on statistical information, and we also give our clients the ability to query their own logs to diagnose mail delivery issues. It was important to find a database that could deliver the high performance we required, but affordability and ease of administration were also of vital concern. These considerations prompted us to seek an open source solution.

Open Source Meets Business Intelligence
MX Force uses a number of open source tools within our organization. The low cost of open source is one reason for this, but flexibility is another important driver. Because open source projects are community-driven, users can tweak, customize and tinker with the software as much as they like. This is a big advantage when it comes to business intelligence, as data analysis requirements can change quickly, and you don't want to have to wait weeks or months to get a new query set up or to change the parameters of those that are already running. MX Force was already using MySQL in our business, so we decided to try Infobright's open source analytic database, ICE (Infobright Community Edition.) ICE combines a columnar database with innovative compression and self-tuning capabilities that eliminate the need to create indexes, partition data or do any manual intervention to achieve fast response for queries and reports. The software is built on MySQL, so for us there was a very small implementation and training curve - ICE uses the same familiar MySQL interface. The fact that ICE is an open source analytic solution presented us with several key benefits:

1. Deployment speed: The time from download and installation to first production use was just three weeks.
2. Affordability: Many of the proprietary commercial BI solutions available today require custom configuration, expensive licensing agreements and equally expensive hardware to support and run it. Not only was ICE free to install, we could also run the software on inexpensive commodity servers, eliminating the need to invest in high performance servers and storage arrays. (Our entire workload is supported by a single quad-core server.)
3. Simplicity and flexibility: Because ICE is open and standards-based, we can quickly make changes as needed without requiring extensive IT assistance. In addition, it's often a lot simpler to make fixes or upgrade an open source solution because an entire community contributes their expertise to fixing bugs and making improvements. With proprietary software, users have to wait for issues to be addressed by the vendor, which can take much longer.

MX Force is currently using ICE to quickly isolate mail flow problems and trends. In our experience, using a free, open source product has not in any way involved a compromise on performance or capabilities. We are achieving 10:1 data compression, which saves on storage costs and boosts performance. Most statistical queries render results in less than five seconds. Ongoing administration is simple. The net result is that the product delivers the fast query performance and reporting functionality we needed, at an incredibly low cost for hardware and ongoing maintenance.

Look, then Leap
Interested in giving open source a try for your BI and analytic efforts? There are a number of compelling benefits to doing so, but as with any type of software, it's also important to look before you leap. Evaluation and testing considerations are no different than they would be for licensed software - you want to make sure the solution has the features and capabilities most relevant to your business. Also, there's a difference between open source projects that are at a very early and experimental stage and software that is well established and has a vibrant and involved community behind it, strong vendor support, or both. Investigate the support offered for the solution under consideration. How often are new features added? Are bug fixes made in a timely manner? Is there useful and accurate supporting documentation?

With ICE, we were certainly attracted by the many resources and significant participation of both Infobright and the user community. We also knew there was a commercial version available if we decided we needed the additional functionality it offered or a formal support contract. For companies just jumping in to the open source arena, it's best to avoid tools that haven't yet cultivated a strong following. But even if you do make a mistake, the low (and usually free) cost of open source means that there's minimal risk.

The BI requirements of today's data-driven businesses demand speed, simplicity and affordability. As open source solutions continue to mature, it's worth looking at projects that are focused on analytics, BI and other data management activities. The more nimble and flexible approach embodied by open source may just be the best fit for addressing the many information management challenges driven by data growth and complexity.