Tim O’Reilly brought two folks to the Gov 2.0 Summit stage this morning: Randi Levin, CTO for the city of Los Angeles, CA and Dave Girouard of Google. L.A. is halfway through a migration for their city-wide email system, moving from Novell Groupwise to Google Apps and Randi estimates an eventual savings of $5M over the next 3-5 years. Tim asked her what she would be able to do with such savings, and her answer was poignant if blunt: Survive.
The budget cuts facing L.A. and other cities/towns/states are substantial and potentially transformational, for good or ill. This brief panel conversation showed how Ms. Levin’s decision to embrace cloud computing could lead to a transformation that is positive in the long run, in ways not yet obvious. Cloud computing is simply a means to an end, in the words of Girouard – and L.A. chose this means out of budget necessity, but may reap benefits that spark the imagination. In particular, Tim’s imagination seemed ignited as he asked if Google would consider combining their auto-translation services with speech recognition to provide the “polyglot” that is L.A. with Star Trek ‘tricorder’ capability in L.A. employee smart phones.
Sure, and that’s the word Dave used to describe some of Google’s technology investments. However, Google’s strategy is to use their sheer bulk to scale these fanciful capabilities (machine learning, computational linguistics) into cloud services. What is fanciful in component enabling technology may be combined into consumer services – think for a minute about the layers of once-fanciful technology that provide you with a GPS system that not only knows the best route, but knows the quickest route at the moment, based on current traffic patterns and/or accident reports.
When Randi mentioned survival, she also touched on the need to get data out of costly data servers located in earthquake zones, and into a more survivable yet cheaper hosting solution. Tim asked about vendor lock-in – what is the competitive landscape when easy and portable cloud solutions become default destinations, as happens with network-driven applications (would you invest in a Facebook replacement?).
The answers were promising: Randi compared migrating data from an SAP financial system to a competing suite – vendor lock-in is nothing new. Dave revealed Google’s strategy revolves around “data liberation.” Simply put: companies own their data, and should they want to move off a Google platform, they should receive that data quickly and in an open-standard format to facilitate migration. (This sounds like a core Principle to be included, in case anyone out there is working up The Cloud Manifesto!)
The conversation was both remarkable and familiar. Asked to provide the core advice for anyone attempting a similar move, Randi Levin reminded us to “know your requirements.” Not realizing her requirements included a political aspect led to an elongation of the schedule. What appeared to be simple decision to move to a solution that was cheaper, allowed for continuity of operations in the even of disruption to city data centers, and allowed for a device-agnostic approach to city email – was anything but simple to execute in the political sphere.
The remarkable aspects were hinted at by Tim O’Reilly in his questions. Having transformed to email as a cloud service, the related and future cloud services that could be linked in for increased value are near limitless. L.A.’s custom Tricorder app may be a few years off, but the slight increase in clouds over L.A. is a hopeful sign in an accidentally transforming era.