Open space for NASA data

"For the benefit of all” is NASA’s motto. Technologies developed by the agency in its 50-year history range from streamlined swimwear to robotic joints for injured horses (the pen that could write upside down, a popular allegory for over-engineering, is a myth).

Now, NASA’s duty to share its technical advancements with society is set to benefit enterprise IT departments. The agency has contributed technology it developed to support its own private cloud to an open source cloud software project.

NASA’s private cloud, named Nebula, was initiated in 2008 by chief technology officer for IT, Chris Kemp. It was designed as a repository for the vast quantities of data that the agency accumulates in its research, making that data accessible not only to its own scientists but also to researchers around the world and to the public whose tax dollars fund its work.

“With a lot of the data from our missions, we have a mandate to show it to the public,” Kemp explains. “Our strategy with Nebula is to put it on the edge of NASA’s network and to make it a great place to put large data sets and dynamic data-driven websites."

NASA employees can use the computing resources contained within Nebula – such as storage and processing – on a utility basis, making it a rare example of an internal ‘infrastructure as a service’ (IaaS) offering. “You can do whatever you want with it,” says Kemp.

Nebula far outperforms the IaaS services that are currently publically available, Kemp claims. “It’s about ten times faster and ten times bigger than anything you can do in any of those current commercial offerings,” he says. The system runs on a 10Gb Ethernet network, and individual virtual servers can use up to 72Gb of storage. “That kind of thing in the commercial environment would [involve] either sending out hard drives to providers or it would be saturating our Internet connection to get data in and out, considering the volumes that we deal with,” Kemp argues.

In its current stage, Nebula is only being used to host data that falls under the lowest of NASA’s three security risk grades. However, Kemp says that NASA is already planning to add further layers of technical and physical security so that more sensitive data can be transferred and provisioned internally by Nebula.

At present though, security is not high on the list of concerns for Nebula, Kemp says, due to the public nature of the information currently stored. “Even in the scenario that we had a breach and data would be exposed to the outside world, it’s just another way of sharing and being transparent with the public,” he believes.

Enlightened self-interest

The Nebula project, Kemp says, was based from the start on open source software. For example, some of the core functions of the system are supported by open source private cloud infrastructure software known as Eucalyptus.

As a publicly funded organisation, Kemp says NASA is keen to make sure that the taxpayers’ money is spent sparingly, and in a way that delivers the maximum return for the public. Using open software, he says, “means that the money you spend goes on research and development.”

Now, NASA is sharing the fruits of that R&D with the enterprise IT community by contributing some of its internally developed innovations to OpenStack, an open source cloud infrastructure stack backed by hosting provider Rackspace.

OpenStack is based on software that Rackspace built to support its own cloud computing offerings. Rackspace hopes that by making this software available under an open source licence, organisations will adopt OpenStack internally, thereby predisposing them to use its hosting services in future. “The longer-term benefit for us is that they could easily migrate to our infrastructure at a future date or use our cloud for extra capacity and bursting,” says Rackspace’s marketing director Fabio Torlini.

NASA’s contribution to OpenStack is its Nova fabric controller, which allows many thousands of devices to be tied together into a single utility computing environment. The agency built Nova itself as it found existing offerings were not capable of the scale it required.

Kemp believes that by providing the basis for an open cloud infrastructure standard, OpenStack will allow more organisations to build so-called hybrid clouds, combining both on-premises and hosted infrastructure. “As standards begin to emerge, I think that hybrid clouds will also begin to emerge,” says Kemp.

Indeed, part of NASA’s motivation for contributing to OpenStack was to make sure that if it were to use public cloud services in future, it wouldn’t be locked into a single supplier

“We owe it to taxpayers to ensure we’re getting the best possible value,” Kemp explains. “If you’ve only got one provider and you get locked into that provider, basic economics says that you’re probably not going to be in a good position.”

Peter Done

Peter Done is managing director of Peninsula Business Services, the personnel and employment law consultancy he set up having already built a successful betting shop business.

Related Topics

Open Source