Amazon Will Ship Your Cloud Data to You … on a Truck

It takes a long time takes up a lot of expensive bandwidth to push 100 terabytes of data across a Wide Area Network.

Amazon’s answer to moving those kinds of data volumes from customer data centers to its cloud data centers has been to ship its customers high-capacity storage servers. The customer uploads their data to the server, which then gets shipped back to Amazon for upload to the cloud.

Amazon announced the service last year. Today, the company started offering the same service, but in reverse. If a customer has accumulated a lot of data in their AWS environment and wants to move it elsewhere, Amazon will put it on its Snowball data shipping servers and ship them to the customer.

It saves a lot of time. According to Amazon’s estimates, it can take more than 100 days to transfer 100TB of data over a WAN, while it takes less than one day to upload that much data to two Snowball appliances – each can currently hold 50TB. Add however long it takes to ship it from point A to point B, and the total time to move that much data is much shorter than 100 days.

A Snowball server is in a rugged container and has all the necessary cabling. Data on it is encrypted using encryption keys provided by the customer. They are not stored on the Snowball itself. A Kindle displaying the shipping label is attached to the container, and the label switches automatically every time the container is ready to be picked up to go to its next destination.

The new service determines automatically how many appliances will be needed to ship the customer’s cloud data. The data has to be stored on Amazon’s cloud storage service called S3.