One of the major advantages of using AWS for the data science team has been flexibility. Sternberg notes, “We don’t have to go through a long ramp-up time to bring hardware in-house, or go through a purchasing cycle, so we can get going really quickly with new ideas.”

Scalability is an essential component as well. “The scalable nature of the cloud allows us to fire up clusters of computers to process data using Hadoop,” says Sternberg. “And those clusters can be very configurable in terms of the number of nodes, the amount of storage, the amount of processing power. So it can be a very scalable solution.”

Finally, security is vital. Sternberg explains, “With AWS, we’re able to use SSL to transfer data securely. We use Amazon’s Server Side Encryption Support (SSE) to encrypt data stored at rest in Amazon S3 as well. We appreciate the general ability to lock down and control access to our compute nodes. This is obviously very important to us, as we can’t risk any unauthorized access.”

Sternberg anticipates continued use of AWS: “We have a long queue of experiments and analyses that we’re undertaking and we envision using Amazon Elastic MapReduce, Amazon EC2, Amazon S3, and some other services from AWS to help us get there. We can envision using those products consistently in the future.”