Government ICT: The problem with Cloud

As I have commented  before, the potential savings by moving to a Cloud based service, for New Zealand Government, is a conservative $150,000,000 per year. I suspect that its higher when we start to look at the sheer scale of infrastructure investment across seventy-six core agencies let alone the next layer of outlying government including Councils and the like. As an example, we know that as of twelve months ago those seventy-six agencies had more than eight hundred Exchange servers between them. As a comparison, Microsoft has two global clusters to manage their email.

I worked on the Government Cloud Programme that was developed by the Department of Conservation and saw New Zealand adopt a “Cloud First” policy at the time. I’m not claiming credit for that, I joined the programme as it was moved to the office of the Government CIO within the Department of Internal Affairs. More about that later.

Around the world, Government Cloud is taking off and is well-advanced, New Zealand is now lagging in terms of adoption. In the U.S. the Government CIO pushed for, and built, the FedRamp initiative. Effectively a method by which companies could certify themselves to be able to sell Cloud to government. And it works. The U.S. government is moving large tranches of it’s compute power to the Cloud including the CIA looking to purchase $615m of Cloud services from Amazon.

In the UK. The G-Cloud initiative is now in its third iteration. Once again, they provide a FedRamp style certification for companies that want to sell Cloud to government agencies. Those that are certified are added to an online store whereby agencies can search for Cloud services and engage directly with those companies.

Canada, India, Sri-Lanka, China, Japan, the European Union, the Philippines, Vietnam, African Nations, South Africa, and nearly every other populous country on the planet is advancing Government Cloud Computing in a safe, secure, and conservative way. So why so slow in New Zealand?

There are a bunch of reasons for this, and a lot of it is to do with the confused messages that come from central government.

Anyone in Government ICT knows about the All of Government offering systems of which the “IaaS” has been a keystone, at least in promotion, for the last three years. And while the ideology of the offering is sound, the execution is less so.

The way that the “IaaS” works is relatively simple. Government went out and selected, via a lengthy RFP process, three separate vendors that could deliver Infrastructure as a Service. This means that if you are an agency that wants to take up that service then you sign a contract with the Department of Internal Affairs that allows you access into the fold. As a result, the agency doesn’t need to go to market to buy that service, they can approach either Revera, Datacom, or IBM directly to get pricing and a service.

Now. This is no different to the All of Government One.Govt service that started as the Government Shared Network (GSN), many years ago. What one.govt drove was the price for telecommunications, down, a lot. The more that signed up to the service, the less the cost. However, as of eighteen months ago, the market reacted around the one.govt and lowered their prices to kill it. So, the ideological concept of government agencies buying an all of government service, died.

IaaS as offered in an all of government service, is not Cloud, in it’s true sense. It is more like a capacity on demand model of the early 2000’s with some capability, depending on provider, to host virtual machines. It isn’t Cloud. Cloud is an on-demand, scalable, flexible, elastic service that can be switched on and off, by the minute, for various services.

There are advantages to the service, on average, it costs half of what the public offering from those three organisastions does. However, that cost is still forty times more expensive than competing Cloud services, particularly those that are delivered out of Australia. That means that in year four, or five, government IaaS is more expensive and has less functionality than the market can deliver. That’s a problem.

Why though? Well, it’s reasonably simple. AoG IaaS is effectively a product that the government has created and then injected into the market in competition with companies that do this as their day-to-day bread and butter. They’ve created a product that is in competition with super-massive Cloud heavyweights like Azure, Google, Amazon, and Rackspace. And before you start on the Data Soverienty issue and the fact that these competitors are offshore, in Australia, consider the fact that Amazon is recruiting New Zealand based staff and already has a country manager.

This means that government IaaS is pretty much dead and cold compared to competing products in the market, which are not all offshore, there are a significant number appearing onshore. More expensive, less functionality, more complex contracts, and no roadmap for the future.

Now, here is the next issue. When I talk to the government CIO’s, they all tell me that there is an expectation placed on them to pick up the government IaaS service, political pressure, and trying to explain to the powers that be that the service is not up there, could be damaging. So they stall. They want to take up Cloud, but are pressured to take up the lesser kind of Cloud that the government IaaS is offering.

Here is what the Cloud Programme was originally tasked with developing. Firstly, a feasibility study on building a Government Community Cloud, actual Cloud, not a Cloud washed alternative. Second, a business model reminiscent of the G-Cloud. Something that had a risk and assurance model that allowed for Cloud Service Providers, regardless of who they were, to be certified for government use and an “application store” of some description that allowed agencies to purchase those services. Lastly, the first cab of the rank, an Office Productivity Offering (read Microsoft 365) for agencies to utilise.

Here’s what happened.

The feasibility study found that unless central government mandated the build of a Cloud service for all agencies, then there was no point in building it. I.e. You can’t build it and expect agencies to consume it when it is less flexible and more expensive voluntarily. You have to mandate them using it. At its heart, the government cannot build a competing Cloud product in a market that is utterly cut throat. If they do, it must be mandatory, regardless of cost and functionality.

The business model, G-Cloud lookalike, collapsed. What could have been a service that would drive Cloud adoption forward, failed. Why? Because at it’s heart, it competed with the existing government IaaS offering. The IaaS offering would have just become one of many products that was certified for government. It would have killed it.

The Office Productivity Project is still running, as far as we know. However, certain interests in government mandated that the data be stored onshore. That meant that Microsoft, Google, or anyone else interested, would have had to build a node here. Short answer, yes, but at a significant cost that likely blew the business case to pieces.

So what should central government do?

Step back, stop trying to compete in a market that is incredibly competitive and instead issue guidelines to agencies on how to safely purchase and consume Cloud. Only then can we see the benefits that Cloud can offer be unlocked. Until this happens, we’ll stall any progress.

Anything else is a waste of money and doomed to failure.


One comment

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: