InformationWeek, July 17, 2008, written by John Foley
As a former database reporter, I take notice when a startup comes out with a new database platform. The 30-year-old database market is dominated by software behemoths likeOracle (NSDQ: ORCL), IBM (NYSE: IBM), and Microsoft (NSDQ: MSFT) with entrenched customer bases. Is there room for another player? TrackVia, which just closed Series A funding, believes so.
TrackVia’s database is available online as a service, not as PC or server software. It's aimed at small and midsized businesses, and it's designed for fast and easy implementation. The company's Web site shows a variety of sample applications -- spreadsheet, contact management, real estate listings -- all variations of the row-and-column format of a relational database.
Relational databases are more or less alike, and if you peel back TrackVia's service, you'll find MySQL. The challenge has always been that databases are notoriously difficult to set up and maintain, especially for companies that don't employ SQL programmers and database administrators. But businesses have no choice -- data volumes keep growing, and they need some way of managing it all. So new market entrants tend to focus on usability, manageability, and performance, not rewriting databases from scratch.
I asked TrackVia for a better understanding of the software that drives its service. The following explanation comes from CEO Matt McAdams.
"We do use MySQL underneath TrackVia, but only for the physical storage and retrieval of raw data values. The management of all metadata -- that is, the fields, data types, and records that define an individual customer's database -- is done in a proprietary data engine that sits above MySQL. So when a user converts a TrackVia field from a number to a short answer, for example, or from a drop-down list into a paragraph box, all of that is handled by our own data engine. As another example, when a user creates a "checkbox" field type, with 5 options that can be checked or unchecked independently, that's entirely a TrackVia construct, not a MySQL construct; and TrackVia's metadata engine is responsible for searching, filtering, updating, converting, and other manipulation of that virtual data type."
"When we built the data engine, we consciously disregarded a lot of conventional wisdom in the database industry. For example, our core data model is not normalized -- we use a very flat data structure for speed. We provide the illusion of normalization in code above the data engine, for example, when implementing a drop-down list or pick-list field, or when linking data together with TrackVia's relational fields feature. As another example, we do very few joins in TrackVia -- we slurp large, flat lists from MySQL and post-process them to achieve the same result as a join. The result is that TrackVia is fast, much faster than any other commercial SaaS software (or any with thousands of customers). And although we've intentionally reinvented the wheel by building many database features in our own engine, our customers have the benefit of using a database that was designed for businesses, not programmers."
I haven't tested TrackVia, but it appears to be a viable option for workgroups and small companies that need to build a database app with minimal fuss. Pricing starts at $30 per month for up to 3 users.
A few years ago, it looked as though the relational database market was consolidating into just a few big players, with IBM acquiring Informix and CA spinning off its CA Ingres database. But new products from new companies keep springing up. In addition to TrackVia, founded in 2005, here are some of the other relative newcomers.
