NASA is a big fan of the cloud – in fact, the OpenStack open source cloud computing platform got its start there. So when NASA needed image processing infrastructure for the incredible pictures coming from Mars to Earth by way of the just-landed Curiosity rover and its mission to search for life on Mars, it’s not very surprising that the team turned to Amazon Web Services.
The initial challenge was a relatively straightfoward use-case. NASA/Jet Propulsion Laboratory (JPL) had to make sure that there was enough going on under its hood to handle “hundreds of gigabits/second of traffic for hundreds of thousands of concurrent viewers” watching the almost-live video feed of Curiosity’s landing on Mars, according to the official AWS case study. NASA sysadmins were able to achieve this with “novel use” of Amazon Route 53 and Elastic Load Balancers (ELB) shunting traffic across regions.
Over the next 8 months, Curiosity is going to be piping imagery back to NASA mission control. The mars.jpl.nasa.gov website – the Internet portal for the Curiosity mission – continues to be powered by AWS, with open source content management system (CMS) running on top of Amazon EC2 and storage from Gluster on a pool of high-I/O Amazon Elastic Block Store (EBS) volumes. It’s also using a multi-availability zone (AZ) MySQL database managed by Amazon Relational Database Service (RDS), and that same set of load balancing and routing services, in conjunction with the Amazon CloudFront content delivery network (CDN).
Again, pretty straightforward. But easily the most interesting part of this case study is also the one that gets the least detail – NASA actually uses Amazon Simple Workflow Service (Amazon SWF) to copy images from the Mars rover itself back to Earth, into an Amazon S3 storage bucket. Metadata is stored in Amazon RDS; Amazon SWF triggers the provisioning of EC2 instances for the processing of each image.
Amazon provided a handy diagram of how it works:
It’s a shame “Mars” doesn’t appear on there anywhere. Regardless, the case study carries the following conclusion:
Now that Curiosity has landed safely on Mars, the mission will continue to use Amazon Web Services to automate the analysis of images from Mars, maximizing the time that scientists have to identify potential hazards or areas of particular scientific interest. As a result, scientists are able to send a longer sequence of commands to Curiosity that increases the amount of exploration that the Mars Science Laboratory can perform on any given sol (Martian day).
So there you have it. The cloud (in this case, Amazon Web Services, but it really could have been rolled with any number of other solutions) can be used to boost efficiency, scalability and cut costs – even across interplanetary distances. And it’s worth noting that NASA was able to do it in weeks, not months.