One of the best reasons to use public clouds is the ability to run large-scale workloads.
It’s all great and all, but what if you have a large workload running in your datacenter that you want to move to the Amazon Web Services cloud? Transferring petabytes worth of data over the Internet is not practical.
You don’t have to transfer large amounts of data across the Internet if you need to move them to AWS. AWS has a service called Snowball that is designed to assist AWS subscribers with large-scale data transfers. It was first revealed at the 2015 AWS reInvent conference. AWSInsider reported that Snowball was launched in 2015.
[Snowball] provides organizations with a briefcase that allows them to physically transfer large amounts to AWS. This is a process that can take several days or even months if done over the internet. Organizations can order Snowball appliances through their AWS Management Console. The appliances start at 50TB and feature “tamperproof” closures, durable materials, and digital shipping labels to reduce the chance of shipping errors.
Snowball is not suitable for all situations. Snowball is not the best choice if you need to transfer large amounts of data. Data transfer jobs that are too large to be done in another way should be done with Snowball.
Logging into AWS Management Console is required to use Snowball. Select the Snowball option on the home screen and click Get Started. AWS will launch a wizard to guide you through the planning of a Snowball job. Figure 1 shows that a Snowball job can be used for data transfer to Amazon Simple Storage Service (S3), or, as a feature update in March, export data from S3.
[Click on the image to see a larger version] Figure 1: Snowball can be used to move data to and from Amazon S3. You will need to create an import task in the case of an import operation. This involves providing your shipping information as well as details about the data to be transferred. AWS will send you a physical appliance to connect to your network. Most organizations will choose to transfer data to the Snowball appliance using the Snowball client. However, it is possible to transfer data programmatically. After the transfer is completed, you can ship your appliance back to AWS. The contents of the appliance will then be transferred to S3.
After I have covered the basics of Snowball, there are some important points you should know. Snowball has a limit on its capacity, just like any other appliance. The USA has two sizes of Snowball appliances: 50TB and 80TB. Snowball appliances can store 80TB of data in other regions.
Despite the Snowball appliance’s limitations, it is possible transfer very large data sets. The transfer operation must be broken down into several jobs, each job representing one Snowball appliance. Transferring large datasets requires multiple appliances. AWS’ default service limit allows only one Snowball appliance per organization. If you need to use multiple appliance simulators, you will need contact AWS to make arrangements.
The most important obstacle to Snowball’s use is the requirement that data be kept static during transfer operations. If data is altered during copy operations, the data will be invalidated and can’t be copied to S3. This is a serious issue as any data set can be modified during the copy process.

Oswald