I spoke recently at the OpenGov Leadership Forum in Manila and at Cloud Expo Europe. At both I started to explore a theme from my final GDS blog post: cloud-native organisations.
//speakerdeck.com/assets/embed.js
“Cloud” is a nebulous term. At the start of both talks I explained that, while there are formal definitions that I’ve found useful, for me it’s just a useful term for starting a conversation about what’s currently happening at the collision point of “the internet” and “computing tools”. Most of the time for technology, I don’t draw a distinction between the impact of cloud, agile, devops and a number of other inter-related movements.
I talked in both settings about the work that we did in GDS and across government, of which the “Cloud First” policy was a significant part. They were both short talks and I was struggling for time. I should have gone into quite a bit more detail about GDS’ work when speaking in Manila.
For Cloud Expo Europe I drew out several lessons from the work in government. The primary lesson was that changing how we approach technology changes how we think about what government does.
Learning from the government journey
I’ve talked before about how there’s a real opportunity to be much clearer about what needs to be bespoke and what we can consume. There are a set of areas where organisations think they have special requirements for their core technology. Most of the time that’s not true.
That is important, but the more transformative change is that as a set of technologies becomes more accessible it becomes part of our normal toolkit.
A much lower barrier to entry for technology lets cross-disciplinary teams incorporate technology development into policy and service design, and to iterate it in an operational setting. A set of traditional barriers are no longer necessary. Those cross-disciplinary teams–given freedom to meet their goals–are where we will get performance and innovation.
Today’s technology world
I went on to talk about some general trends in the technology world, once again citing Stephen’s Developers are the new Kingmakers. Top-down decision making isn’t enough to handle the pressure organisations are facing to change, or the scale of opportunities available through open source and cloud technologies.
What we see with the mature end of devops, and in much of the thinking behind “cloud native” architectures, is a strong focus on the team as the unit of delivery. And the team as the unit of responsibility. There are a set of architectural practices that support independent teams. They work where organisations set clear goals and principles.
Too often when we talk about “cloud” we either talk about infrastructure-as-a-service, or we talk about software-as-a-service. By and large they’re treated as very different things, but there are some common principles here that ought to be applicable to IT adoption more generally.
Software-as-a-service–particularly software that’s offered via a “freemium” model–allows a complete shift from a world where every new productivity tool has to be evaluated by a central IT department. People are expecting tools at work at least as good as those they have at home, and often will just adopt what they need whether the central department likes it or not.
As with many things we associate with cloud, that’s not new (though the scale has changed). It’s what we have been calling “Shadow IT”. With the shift to cloud we have the opportunity to change our approach.
In the talks, I cited the " Cloud Security Principles" as an example of how being more explicit about the things we care about lets us empower more people to make good decisions.
Those principles are an early illustration of how we should be thinking about cloud tools (IaaS or SaaS) in general. We should be clear about the principles that apply and help our people understand what we need to watch out for when choosing technology. Then we can help scale the use of the tools they find most useful.
There’s a lot more to do to take the thinking in those principles and really open it up to a wide audience, but they work very well as an indicator of direction.
“Software that treats people like people, not like cogs in the machine”
To illustrate what that should be letting us do, I stole a line I loved from a recent Netflix blog post. A goal of all “cloud adoption” efforts should be to give people more control of how they meet their goals.
The real challenge
As with almost everything, the real challenge for organisations is about leadership. Organisations face many challenges and it’s easy to fall back on locking things down as a way of reducing the perceived risks. That’s rarely the most productive approach.
Leaders need to focus on being open: about the real objectives, about the principles that guide how the organisation approaches issues, and about any particular risks.
Management needs a set of practices that understand what’s happening in the organisation, but that’s to keep things on course not to lock them down.
When thinking about cloud-adoption people seem to often get caught up in questions about catalogues of services and new lists of approved tools. Those things may be useful, but we would be better off spending more time explaining how we create a well-aligned, enabling environment before we get down into those sorts of details.
Update: Kush quite rightly pointed out that I didn’t go into anything specific about what to do next. In Manila, there were a set of roundtable conversations where we could begin to get into that. In London, I really just wanted to get people thinking. More blog posts to come…