Dr Steven Johnston from Southampton University explain how his team have used cloud computing to predict satellite collisions in space.
10,000 space objects are currently being tracked. These objects are satellites, typically space debris. Johnston’s team are using Azure to project the orbits to predict collisions.
Cloud computing is elastic – it can be scaled up or down, which can be useful for processor-intensive work such as Southampton’s Space Situational Awareness System Tech.
There is a scale of cloud computing, from Amazon Web Services – where you are responsible for managing a virtual machine – to Google Docs, in which you have no control at all over the machines you are working with. In the middle lies Windows Azure. You can’t configure Azure yourself, but it does have a more managed infrastructure.
While Southampton’s space-tracking system relies on Azure, Steven freely admits that both cloud computing in general and Azure in particular have limitations: there’s no standard API for moving a VM, which essentially locks in a user to a particular system; there are bandwidth problems with moving data; and the inherent access limitations make it more difficult to create clusters.
Image: Known orbit planes of Fengyun-1C debris one month after its disintegration by a Chinese interceptor.