With California food stamp contract, Code for America shifts funding strategy
July 24, 2017
The national civic tech group is diversifying its funding sources as it deploys food assistance services in California.
Rod Davenport shares his thoughts on the cloud, the future of consolidation and the importance of seeking the middle ground.
Colin Wood is the managing editor of StateScoop. Before that, he was a staff writer for Government Technology magazine. Before that, he taught Engl...
For nearly five years, Rod Davenport has been leading the the State of Michigan's IT licensing, contracting, infrastructure and operations, and managing its enterprise architecture. It's a government known for having pursued enterprise IT early.
The state's executive branch has been calling for centralization of the state's siloed IT agencies since 1994. But in any state, data center consolidation and modernization efforts are never finished, and can even be rolled back at the whim of a fickle legislature. Each of Michigan's recent governors have continued pushing for an efficient enterprise technology presence, including current Gov. Rick Snyder, and it's showing — Michigan is frequently referenced among state CIOs as a leader in the realm of IT consolidation.
Before joining state government, Davenport spent a decade with DTE Energy, a Fortune 500 energy company near Detroit, where he worked as an enterprise technology strategist. Davenport shared with StateScoop how the Michigan's consolidation efforts are progressing and which technologies the state's Department of Technology, Management & Budget will pursue as the never-ending quest for an ideal state-sized technology rig continues.
How have the state's modernization and consolidation efforts evolved in the last five years?
Since I've been here, we have successfully contracted for a third party data center with Switch. The timing of that was fortunate because it coincided with an RFP we had already issued looking for overflow data center space. So I've got two-ish data centers right now. Part of the problem is that they're fairly close together.
One of the data centers is reaching its limits in terms of the amount of power we can bring in and as we have been trying to drive our technical infrastructure to more of a virtualized, converged approach, power densities have gone up and we're having some struggles in one of the data centers to meet that. We are one of [Switch's] early customers — we've got several rows of equipment there, we have connected them together with the redundant fiber connection and we're looking to continue to build out in that space.
It's not really consolidation, but it's more an evolution of the consolidation. I started this next generation digital infrastructure project a few years ago, which is kind of a data center within a data center idea that we would try to bring cloud-like virtualization and provisioning to our on-premise [data center] and the idea was that we could use that as the basis for a hybrid cloud strategy moving forward.
We've started some work with Microsoft Azure and we're kind of working to finish the stages of procurement to allow us access to some other public clouds. The things that we feel from a performance or security perspective are best left closer within our direct control, we'll be able to do that while still having some of the quickness of provisioning and capacity planning you get from cloud providers. And the things that are more economical we can put in the public cloud and the things that are in between, to the extent we can leverage a fed cloud certified offering from some of the other cloud providers, we can take advantage of that as well.
Longer term, we're looking to right-size our compute depending on the performance needs, the cost and the security requirements.
It seems like no matter how much progress a state makes in this area, the work is never done. What is the goal going forward?
I think IT is like that, as well. The work is never done — it's a fairly fast-evolving industry. Although, some would say IT is kind of moving out of its teenage years and becoming more mature, but we we've seen several evolutions from the mainframe to more of the distributed computing and when people started going to the web, the whole e-commerce thing and now we're looking at machine learning and the Internet of Things and I think there's always these punctuated equilibrium events in IT.
From an IT strategy perspective, because it's difficult to predict the future and because these disruptive technologies sometimes seemingly come out of nowhere and then they get traction and people see that there's a great use case for them, I think maintaining the ability to have some agility, to pick up on these trends to take advantage of them where there's performance, cost or business advantages is where we’d like to be.
So from an infrastructure perspective, I think having that sort of nimbleness built into the organization, leveraging some of these tools that we can use to both manage some of our external cloud as well as our internal cloud from not a single pane of glass, but maybe a consolidated set of tools is a good way to go.
How should the cloud fit into these plans?
There have been some predictions that no one will ever have any data centers X number of years from now because it will all go to cloud. It's true that to a certain extent the care and feeding of the hardware can be challenging, and it's true that largest cloud providers do have some economies of scale, but … I think these things generally settle out somewhere in the middle.
So, I'm trying to deliberately start in the middle with our cloud and infrastructure because I think that's probably going to position us better for the long-term.
There's certainly a place for cloud, I think it's a growing piece of our technical portfolio, but there are always going to be instances where you want to have things that are on-prem.
In the context of consolidation and modernization, which states do you look up to?
There are not many states that are consolidated. I think Utah is another state that has a consolidated approach. We speak fairly regularly with Wisconsin and Indiana and Ohio — there's a little bit a of kinship of the great lake states.
There's some sharing of best practices. We're hosting some of the Medicaid processing for Illinois, for example, but I think that although the states all sort of do the same things, I think that they have different contexts. Some of the states have more money than others. Others are more highly populated or densely populated, so there's nuances to how these things are implemented, so I don't really know if there's another state we put on the pedestal as a blueprint of things to be. Each state tries to do the best they can with the resources they have.
States that have consolidated IT, just from an organizational model, there's some advantages to that, but there's also disadvantages, right, so as you're simplifying the technical environment in the hopes of driving down costs, there tends to be a sacrifice of customization. That's another balancing act that I think centralized approaches need to take.
Has being with the state changed any of your attitudes about government? Any clues as to why these systems are often so far behind?
This is my first government role, and being on the outside of government, like everyone, I kind of looked at it and said there's a lot of things that can be improved. When I came in, I had what turned out to be a pretty aggressive plan in my mind for the pace of change. And I had originally attributed some of this to the nature of bureaucracy, that it's slow to change.
I think in its best form, as a society, we don't really want government to be able to change, because part of the benefit of it is that it provides stability for people to live their lives, so there needs to be some enduring quality to government. And I think some of that ethos around the nature of stability sort of permeates through the agencies and into the nature of government itself. So in the areas I think would benefit from faster change, they're still hampered by this cautious culture.
Originally, that was frustrating to me, but I've kind of grown to have an appreciation for it. There are areas where I think we should move more quickly, but there are sometimes good reasons for not doing things, and I think having spoken with some of the people that have been here longer and getting a better sense of the perspective from their years in government, it's not always just what the outside would see as just a fear of change or an unwillingness to embrace change. I think there's a willingness, but I think that there's also times where the roll out of a technology wasn't done as well as it could have been or maybe they mistimed the technology with the maturity and readiness of an agency to accept the change that technology would bring and so then it's led to less than stellar outcomes. It's more of an informed experience-based cautiousness, rather than just an aversion to change.
How do states balance these needs and overcome barriers to change?
It's important to stay within a certain range of the state of the practice in terms of IT. Everyone knows that if you get too far behind the curve, then you're supportability goes down, your costs go up and you have security implications from that.
And then certain agencies have a certain appetite to move slower or faster depending on how often their business requirements change.
So I think it's important to in some cases apply a little bit of pressure to make sure that your application and infrastructure are keeping pace. And even though sometimes that might be perceived as technology for technology's sake or just upgrading for the sake of upgrading, I think there's good reasons to do that and sometimes IT needs to be a little bit of a forcing function during your budget planning process around those pieces.
Probably if you talk to the other consolidated states, they would likely say the same thing. I don't know that there's a silver bullet to doing this. It takes a while, there's a lot of different perspectives that have to be listened to and harmonized to the extent that you can and it's a lot of work. There's a lot of systems that need to be converted and there's a lot of wrangling that needs to happen on the backend of that, so it's probably a worthwhile journey, but it's not a quick or easy one.