Former VMware CTO Steve Herrod is now mobile API 'appy
Tell marketing to shove it and build your apps right, says virtualiser-turned VC
Your marketing department wants to build mobile apps to expose your business processes to world+dog, but if you do it their way you'll fail and set back future mobile development efforts.
So says Steve Herrod, former VMware chief technology officer and now venture capitalist with General Catalyst.
Speaking to The Reg today, Herrod said he's seen plenty of companies – among them Facebook, LinkedIn and Intuit - outsource the build of a web app for mobile platforms, watch it suffer thanks to a small feature list, then try a hybrid mobile app and again fail before finally dedicating a team to a native mobile app.
That third effort, he said, will only succeed if organisations understand that a native mobile app needs to communicate with core apps using tailored APIs. Those can be RESTful – there's no need to build anything exotic – but the point is not to rely on existing means of exposing data to apps because those older approaches will either lack the security needed for mobility or will harm the user experience by assuming a LAN as carriage rather than the weirdness that is public data networks.
“On the desktop you can make an app chatty,” he said. “In the mobile world you need to think about that: APIs need to make one call that returns a lot of data,” so that mobile devices don't need to go back and forth to servers.”
Herrod thinks building apps in this way will change data centres because it will mean a lot more East-West – or server-to-server - traffic within data centres. Which isn't a surprise given VMware acquired network virtualisation pioneer Nicira on his watch. Nicira is often used to dynamically build networks within a rack, so that once traffic reaches a top-of-rack switch it's ready to do something useful instead of bouncing around some more within a data centre's walls.
He therefore sees building mobile apps and the APIs to power them as a chance to re-engineer data centres and software, the better to cope with the cloud/mobile future.
He's also convinced such thinking isn't fad-following, insisting that his role as a VC doesn't differ much from his time as a CTO. Both involve chatting to CIOs about business problems, forming a view about the future and then making a bet about innovations. At VMware than meant directing product development. At General Catalyst he flings cash at startups he thinks can eventually turn a quid, and says he has a couple of API 'appy ventures ready to roll.
Your reaction to those ventures will determine if he's right. And maybe your comments will shed some light on the likelihood of that outcome. ®