At the Amazon Web Services (AWS) Summit in New York City today the company announced it can now do EC2 data processing on its Snowball edge device.
AWS announced Snowball in 2015 as a way to store and transfer massive amounts of data at the edge. Snowball is a physical box with up to 100 terabytes of local storage that AWS offers for massive data uploads, the kind that would take weeks to do remotely. Customers load their data onto the device and then physically ship it to AWS for uploading to the cloud.
Today, AWS announced that Snowball Edge can now run Amazon’s Elastic Compute Cloud (EC2) instances to collect and process data in edge environments. It can do this with limited or non-existent internet connectivity before shipping the processed data back to the AWS cloud for storage, aggregation, and detailed analysis.
Each Snowball Edge device is powered by an Intel Xeon D processor running at 1.8 GHz, and supports any combination of instances that consume up to 24 virtual CPUs and 32 gigabytes of memory. Customers can build and test Amazon Machine Images (AMIs) in the cloud and then preload them onto the device as part of the ordering process. They can use the EC2-compatible endpoint exposed by each device to programmatically start, stop, resume, and terminate instances. This allows them to use the existing CLI commands and to build tools and scripts to manage fleets of devices.
“By using compute EC2 you will be able to load your Snowball Edge with AMIs, enabling your device to function as a mobile data center,” according to a blog posted by AWS Chief Evangelist Jeff Barr.
At the event, Werner Vogels, VP and CTO at Amazon, said “AWS Snowball Edge has AWS Greengrass capability integrated into it. It can collect data from remote locations, locally, do some pre-processing using Lambda functions, and send that data back to AWS.”
Greengrass is the company’s software that helps edge devices process data and communicate with the AWS cloud. And Lambda is the company’s serverless compute service that allows developers to run code without managing servers.