(Photo: Alex Williams of TechCrunch, Jay Ferro of the American Cancer Society, and Sarah Novotny of NGINX. Didn’t catch the cartoonist’s name, sorry.)
Technology is pushing change onto IT departments quickly, but a department’s culture is the aspect that needs a more severe change, one that ties IT more closely to the core of the business.
That was the consensus reached during a panel discussion last week, an informal affair called “The App Gap” arranged by Dell and hosted at the headquarters of venture firm New Enterprise Associates. (Replay available here.) Local press got invitations to hang out during the discussion, which was broadcast over the web and is archived here.
Asked for four big bets they’d make in a greenfield scenario, the CIOs picked three directly related to people: creating a culture where people share knowledge (versus senior IT people dictating things to junior ones), encouraging experimentation and the inevitable failure that comes of it, and thinking of IT as providing services rather than tools. I’d argue the fourth is a cultural choice too: taking full advantage of open-source software.
These overlap quite a bit, especially the culture of learning and the culture of failure. The point is that what was on panelists’ minds was not technology, but the changing nature of IT jobs, a factor that’s lurking behind most of the discussions we’ve been having on SDNCentral.
IT is becoming a competitive advantage for some companies, but it’s still a cost center. Safeway, represented on the panel by CIO Barry Libenson, doesn’t make extra revenues directly because its networks are awesome or its user authentication is slick. So, IT has to convince the rest of the company that these things matter. To that end, participants such as Jay Ferro, CIO of the American Cancer Society, stressed that IT has to get involved with the business side of the business. Fewer CIOs are coming from computer science backgrounds, because that training isn’t as important to the job, he claimed.
“I can teach you the boxes and wires, but if you can understand the business and how it brings in money holistically, you have a huge advantage over other IT people,” Ferro said.
Michael Cote, an analyst with 451 Research, put it more colorfully: “If you’re in charge of where tanks go, it’s bad to send them into a swamp,” he said.
Given the bring-your-own-device (BYOD) trend and its bring-your-own-app relative, it might be tempting to let employees just do whatever they want. Different groups could find the platforms that work best for them. But then, as panelists pointed out, a holistic, enterprise-wide set of tools (sorry, services) never arises. You get the predictable mish-mash of applications that don’t necessarily play nicely together.
The love for open-source was almost unanimous and almost mushy. That was true even of Libenson, the longtime IT vet working for an old-school, brick-and-mortar company. He was the one who spoke up first when the panel got asked about the four big greenfield bets they’d make.
“I never thought I would say this in my career, but: open-source all the way,” he said. “The ecosystem has evolved incredibly. … All the intellectual brainpower is in that space.” (Separately, Libenson also noted that if he could do it all from scratch, he’d rent space in colocation centers instead of building data centers, another case of taking advantage of a resource that wasn’t so certain a bet before.)
Criticisms of open-source applications did emerge. Luke Kanies, founder of Puppet Labs, pointed out that open-source projects are “rarely good at creating really new things.” (Puppet itself is an exception.) That’s not necessarily bad; NGINX could be considered a pared-down version of Apache, but that’s deliberate, because “it turns out you need a lot fewer features than in 2009,” Kanies said.
Open-source also doesn’t tend to apply to massive projects, said Alex Salazar, CEO of Stormpath, which develops APIs for authentication and management. (The OpenDaylight Project, a massive project specific to software-defined networking (SDN), was a bit beyond the purview of the discussion.) Salazar sees a lot of organizations progressing past open-source software usage into a different direction: using an API to tap a third-party service “and never think about it again.”