What SaaS can teach us about good software design

It may not be a man-bites-dog situation, but it’s still surprising to hear about a software-as-a-service vendor moving toward an on-premises deployment model. And yet that is just what Birst is doing, first with its business intelligence service in September and now with its homegrown in-memory database.

Conventional wisdom has it that pure-play SaaS suppliers are in the cat-bird seat now, watching their valuations skyrocket while the value of legacy enterprise software vendors merely rises and those legacy players try to freshen up by buying SaaS expertise or entire SaaS companies. Birst CEO Brad Peters’ conventional wisdom is that vendors should sell their software they way people want to buy it.

To be clear, Birst isn’t giving up on SaaS, it’s just being flexible, he says. And the on-premises software comes in the form of a software appliance that is easily updated over the web.

“Some people want their data near their BI and some want their BI near their data. We just want to be flexible,” he said in an interview Wednesday. Updates and upgrades happen essentially the same way for both models, which is to say easily and behind the scenes.

The move seems to be paying off. In the first quarter since the BI service was made available in both formats, the appliance accounted for 40 percent of deal volume, he said.

“The issue is how you support your software, and if you can support it like you do in the SaaS model, it really doesn’t matter how it’s deployed. On-premises software doesnt’ have to be crappy just because it’s on premise.”

In-memory databases are hot now as companies want to quickly tap into, parse and visualize their enterprise data, and Birst’s decision to build versus buy such a database was not taken lightly.

“We traditionally used third party databases — which was the one part of our stack we didn’t provide. But when we searched for an analytics database to embed from a third party, the only options were Oracle’sTimesTen and SAP HANA, which cost a million bucks and come with proprietary hardware. That is not viable in our market,” Peters said.

The company says its new SQL-based columnar database, due in January, can be tied into outside systems via the Java Database Connectivity (JDBC) API. Single nodes of the Birst In-Memory Analytics Database come with the company’s analytics service so customers can try it out.

As noted, though, Birst is just the latest entrant into this hot market, where SAP’s HANA won rave reviews since it debuted earlier this year and Oracle’s Exalytics data analytics appliance is due by year’s end. Given the sheer size of this addressable market — businesses of all sizes are clamoring for real-time analytics — it looks like there’s room for at least one more vendor.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

  • NewNet Q2: Google closes the quarter with a bang
  • From car to cloud: the future of the in-vehicle app landscape
  • Finding the Value in Social Media Data



GigaOM