Skip navigation
Google's Transfer Appliance offers up to 480TB in 4U or 100TB in 2U of raw data capacity in a single rackmount device Google
Google's Transfer Appliance offers up to 480TB in 4U or 100TB in 2U of raw data capacity in a single rackmount device

Google Designs Data Center Appliance to Ship Client Data

For a large enterprise, one of the costliest and most time-consuming steps in moving to the cloud is transferring the enormous amount of data stored in its on-premises data centers to its cloud provider’s data centers. Network bandwidth is a precious resource, and even when you have tons of it, moving petabytes of data over a WAN can take way longer than is practical.

Amazon Web Services solved this problem two years ago by introducing a service called Snowball. If you have lots of data you want to upload to AWS cloud storage, the company ships you a rugged storage appliance, which you connect to your internal network, upload your data to it, and ship it back to Amazon.

Today, Alphabet subsidiary Google announced beta launch of a similar service, taking another step in its effort to catch up to AWS and Microsoft Azure in the enterprise cloud market. The service, creatively named Transfer Appliance, is slightly cheaper per TB than AWS Snowball, although the exact price difference will depend on your specific shipping costs.

 

Google

Google’s time estimates for transferring data over networks with varying bandwidth.

Google’s time estimates for transferring data over networks with varying bandwidth. Click chart to enlarge (Image: Google)

Another difference is in the design of the appliance itself. Besides storing more data, Google’s Transfer Appliance is designed to be mountable in a standard 19-inch data center rack, while the Snowball looks more like a PC tower built for an active battlefield.

Each cloud provider offers two models of its data migration device. The two Transfer Appliance options are 100TB in a 2U box and 480TB in a 4U box. Snowball has a 50TB and an 80TB option.

The Google service costs $3 per TB or $3.75 per TB, depending on which of the two versions of the appliance you select. Curiously, the higher-volume version of the appliance commands the higher per-TB price. You’re also responsible for shipping the Transfer Appliance (the service uses FedEx), which will run you about $500 for the 100TB model and $900 for the 480TB one.

Amazon’s Snowball service costs $4 per TB for the 50TB model or $3.12 per TB for the 80TB one. You pay for shipping to an Amazon facility too, and unlike Google, Amazon doesn’t provide set pricing for shipping, saying it will depend on your location and the shipping option you choose (e.g. 2-day or overnight).

Amazon also made the shipping process itself easier by mounting its Kindle readers on the appliances. When a customer is done uploading their data to the device, the screen on the Kindle automatically displays the correct shipping label, and a shipping company is notified that the device is ready to be picked up.

In Google’s case, the customer has to email support to request a shipping label and wait until it arrives in the mail before they can ship the appliance.

Google may be outsourcing some portion of the data migration service to the enterprise data management giant Iron Mountain, although we weren’t able to confirm this. “Google doesn’t publicly disclose this information,” a company spokesperson said in an email.

Documentation for the Transfer Appliance includes instructions for users to grant permissions on their staging bucket in Google cloud storage to [email protected]. That Iron Mountain is an official Google Cloud Platform partner is public information, but public information available about the partnership is limited only to a service for transferring customer data stored on tape in Iron Mountain facilities to Google’s cloud data centers.

Hide comments

Comments

  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
Publish