Changing Role of the CIO

We’ve recently hired our first CIO at MySQL, and it’s interesting to consider how the role of the CIO is different from what it used to be. MySQL has some particular challenges from an information management perspective because we have so many employees working from home around the world. Otherwise, I think the role of CIO at our company is not that different from most other young companies. But what’s really changed is the nature of IT in the past six or seven years.

Back in 1999 and 2000, a lot of IT effort was spent in scaling up. CEOs were asking their IT organizations to get them on the web, rollout ecommerce solutions and get it done fast. Money was no object as long as companies could somehow be part of the Internet boom. Billions were spent on expensive hardware, application servers, content management systems, and complex Enterprise Application Integration (EAI) solutions to tie it all together to the new web infrastructure. Then as we hit the dot com bust and larger telecom bust in 2001, suddenly… everything… slowed… down.

Not only was there an excess capacity in bandwidth, but hardware was going for ten cents on the dollar. Budgets were slashed. Jobs cut. Businesses shut down.

But there were still critical projects to be completed. So a lot of savvy IT folks figured out that they could deploy cheap commodity x86 hardware and LAMP (Linux / Apache / MySQL / PHP) software that didn’t require budget approval. No doubt a lot of open source infrastructure and tools got brought into organizations under the radar.

I also think there’s a fundamental shift in most company’s tolerance for expensive IT projects. As a rule of thumb it used to be that payback periods for IT projects had to be around 12-18 months, about the average tenure of a CIO position. (Remember, CIO stands for “career is over.”) But I think too many people got burned on expensive projects. In my view, the days of multi-million dollar enterprise licenses have gone the way of the dodo.

These days, I expect that even if there’s a positive payback, most companies will avoid “big bang” projects in favor of smaller, incremental solutions. I’ve seen my share of expensive Enterprise ERP and CRM applications that took twice as long as expected to implement and never got widespread adoption. I doubt there was a positive return on investment for many such projects.

Now that many IT staffers have several years of experience with open source infrastructure software, I think a path has been cleared for open source applications, whether it’s for ERP (Compiere, OpenBravo), CRM (SugarCRM, CentricCRM), Content Mangement (Alfresco), eMail (Zimbra), etc.

I don’t think most young IT organizations thinks about the old-school heavy duty client/server enterprise software model much any more. (Or if I’m drinking too much web kool-aid here, let me know.) So for many, the options come down to home grown, open source or “on demand” or hosted services. Given the high quality of open source applications these days and the low upfront costs of on-demand solutions, I think most IT organizations would be better off not writing their own custom solutions. In some cases maybe the needs are very specialized, but otherwise you’re better off getting a working solution and adapting your business.

At MySQL we have a mix of open source, closed sourced, on demand systems and some very old legacy homegrown apps that we’re in the process of replacing. We use JasperSoft for some of our reporting, as well as Pentaho for some data warehouse projects. We have Zimbra collaboration suite for email and calendering running a pilot project. We use Asterisk for VoIP. And of course, we’re running MySQL 5.1 beta in production in about half a dozen internal applications including much of our web site. I’m sure there’s some other open source apps we’re using that I’ve missed.

With the wide range of open source applications out there, from Alfresco to Zmanda, there’s no reason a CIO can’t get what is needed to run the business with a mix of open source and on demand applications rather than expensive closed source solutions.