18F shines some light on cloud.gov platform, but concerns remain

The General Services Administration’s much ballyhooed 18F organization rolled out a new service, cloud.gov, earlier this month.

While there was some initial confusion of what cloud.gov exactly is, an 18F spokesman confirmed that cloud.gov is an platform-as-a-service based on the open source technology called Cloud Foundry.

Cloud Foundry is a not-for-profit supported by a who’s who of private sector technology companies, such as IBM, VMWare, Intel, EMC and many more. The organization’s website says, “Cloud Foundry is an open standard for cloud applications. It is open source, multi-cloud, and multi-vendor. It is hardened production infrastructure for global enterprises. It is designed to make dev/ops the normal state of computing. It is built for fast-cycle innovation of cloud applications. For business, this means faster speed to market and higher rates of user adoption. For development, it means scalable microservices and continuous deployment. For operations, it means faster cycle time and higher reliability. For everyone, open source means no vendor lock-in.”

This sounds good on paper. 18F is taking advantage of a no-cost or low-cost offering to help agencies move more quickly to the cloud.

Advertisement

“Cloud.gov enables 18F deploy its cloud-based applications with baseline security and scalability concerns addressed consistently up front, without dramatically scaling the number of cloud operations experts in our organization,” an 18F spokesman said by email.

But during a conversation on Twitter with several other reporters, feds and contractors, the question came up as to whether 18F should even be offering these services. If cloud services are widely considered a commodity, then why is the government competing with the private sector to offer these services?

If you take a look at the website for the Federal Risk Authorization and Management Program (FedRAMP), six vendors have earned Joint Authorization Board (JAB) approvals for PaaS and nine vendors have received agency approvals.

So if 15 vendors and agencies are offering PaaS already, why did 18F presumably spend time and money to develop another competing platform?

One federal CIO, who requested anonymity because they didn’t want to be seen as criticizing fellow federal employees, said there are a lot of unanswered questions about cloud.gov.

The CIO asked whether this was the best use of 18F resources given the private sector is probably cheaper for this commodity cloud service. The CIO also wondered whether the resources would have been better used to help agencies tackle “harder” problems, or even working with PaaS vendors and agencies to set up pre-negotiated contracts so agencies can buy a la carte.

The 18F spokesman said cloud.gov isn’t competing with the private sector.

“By using Cloud Foundry and making cloud.gov available, we’re raising the ‘lowest common denominator’ capabilities of agencies should expect from any vendor to the Cloud Foundry capabilities, and enabling greater competition,” he said. “Commercial Cloud Foundry vendors can provide managed services and support/consulting options that are beyond 18F’s capabilities due to our funding or staffing. The Cloud Foundry application programming language provides easy mobility from cloud.gov to commercial vendors’ services for government-developed applications as warranted. To some extent, cloud.gov is seeding the government market for value-add offerings from these vendors.”

The spokesman added PaaS vendors not part of Cloud Foundry also will benefit because cloud.gov will increase the migration of government applications to a “12-factor model, and therefore the applicability of PaaS models to government.”

“12-factor applications have very little dependency on the exact platform that is supporting them, so these applications will be more easily ported to non-Cloud Foundry commercial PaaS solutions as well,” he said. “Vendors of such systems have open-book access to how cloud.gov works, and can therefore provide easy on-ramp solutions to migrate government applications to their platform.”

Another federal CIO took a different tack. The source said it looks like 18F is trying to be a PaaS cloud broker.

“It looks like they are giving you templates and ways to work with particular CSP,” said the CIO. “The more I looked at cloud.gov, the more I would like to know the details. It didn’t seem like they were trying to duplicate what industry was doing, but help us get to the cloud faster.”

The CIO said the one concern they have is around the low-security level cloud.gov seems to be offering. The source said most of the agencies need at least a moderate level, so having a low-level approved platform-as-a-service isn’t all that helpful.

A former federal CIO offered a similar take. This former CIO, who requested anonymity because they didn’t get permission to talk to the press from their current private sector company, said the value is unclear as anything built on top of the Cloud Foundry would need a separate authority to operate—meaning the software or application would have to go through the FedRAMP process.

“18F tends to find a niche where they can build better or more cheaply or one that meets needs than those that are commercially available,” the former CIO said. “18F also built a website platform for Web publishing for use by federal agencies which have an unsophisticated website and just want to maintain static content in consistent and affordable fashion. Is cloud.gov a long-term sustainable offering? It’s hard to say, but it may meet some near term needs that they believe are out there.”

The 18F spokesman said cloud.gov is part of the organization’s overall fee-for-service—or cost recovery model—approach.

“The current model is provisional for the pilot phase, and is subject to change in the future,” the spokesman said. “In [the] agile tradition, we are validating whether the model we’re proposing fits the needs of federal agencies, which often face ‘color of money’ issues around how they spend appropriated or budgeted funds. The current model is also a test to validate that cloud.gov is cost-recoverable in terms of the 18F/GSA expenditure to develop, provide and support it for other agencies, rather than just 18F’s internal use.”

The former federal CIO said in general cloud.gov seems like a good thing for agencies to use and build cloud applications on.

A third federal CIO tells me they are “kicking the tires” on cloud.gov to see if it would be useful for their agency.

So why are vendors and some in government so concerned? It is a matter of better communication from 18F? Or does the entire concept of 18F make feds and contractors uncomfortable?

Cloud.gov will be worth watching to see if 18F can attract customers beyond the work they are doing for agencies. It also will be interesting to see when Congress decides to take a look at the 18F concept and whether the government is unfairly competing with the private sector.