Mathias Rosenthal - Fotolia
Google continues to fill the gaps in its enterprise strategy with a new cloud data transfer service that follows a path forged by its biggest competitor.
Transfer Appliance is a rackable server for customers to transfer as much as 1 terabyte of compressed data from their own data centers to Google Cloud Platform (GCP). The high-capacity servers are shipped to users, filled with their data and mailed back to Google where the data is uploaded to Google Cloud Storage. It's akin to the popular Amazon Web Services (AWS) Snowball, though with differences in design and capabilities.
The service targets enterprises that want to adopt public cloud more readily, but wrestle with massive amounts of on-premises data.
"Migration is really hard and one of the challenges around any type of workload portability is not moving the applications, it's moving the data," said Terri McClure, an analyst at Enterprise Strategy Group in Milford, Mass. "There's a data gravity equation that needs to be considered."
An enterprise with 10 petabytes of data to transfer over a typical bandwidth of between 100 Mbps to 1 Gbps would need somewhere between three and 34 years to complete that migration, according to Google. With Transfer Appliance, the time to get that data into the cloud is reduced to weeks. Data is encrypted before it moves to Transfer Appliance and the customer decrypts the data at its final destination. Google doesn't have access to the keys for that data.
The ability to more easily transfer huge amounts of data to GCP will help accelerate large enterprises' migration to the public cloud, said Dave Bartoletti, an analyst at Forrester Research, a market research firm in Cambridge, Mass.
"This bulk loading method should be attractive for customers who have shied away from cloud for large data analytics or machine learning jobs because they have limited bandwidth and/or security concerns about transferring data out of their data centers," he said.
Even with highly secure fiber channels, moving large amounts of data to the cloud is a challenge for many enterprises, said Dave Tucker, senior vice president of product development at Workiva, a financial services software provider in Ames, Iowa, that uses GCP. The company doesn't necessarily need the Transfer Appliance, but Tucker sees the value in the new service and says it definitely fills a gap.
"Any time you can simplify and streamline processes like this you are going to open new opportunities and markets," he said.
Sizing up cloud data transfer services
Google's Transfer Appliance and AWS Snowball address the same need for enterprises that want to shift large amounts of data to the cloud, but there are some differences between the two.
The Transfer Appliance works within the confines of users' existing setup, which is important because data center space is at a premium for most enterprises, Google said. Snowball, by contrast, comes in a large hardened case intended to handle any bumps between data centers, although the newer 100TB Snowball Edge is also rack-mountable.*
Transfer Appliance comes in two sizes of raw storage capacity: 100 TB for $300 or 480 TB for $1,800. With compression, they can hold between 200 TB and 1 PB. Google estimates shipping will cost $500 and $900, respectively.
Snowball, by comparison, comes in 50-TB and 80-TB versions, with a charge of $200 and $250, respectively.
Snowball, which has been available since 2015, has several features not found in Google's Transfer Appliance, such as compute capabilities and the ability to transfer data from the cloud back to on premises.
Cloud data transfer widens Google's reach
For Google, this is more than just a storage play. It has positioned its cloud as an advanced analytics platform for modern applications in areas such as IoT or machine learning, and it sees this device as a means to squeeze intelligence out of archived data wasting away inside customers' data centers.
Transfer Appliance joins a growing list of services aimed to attract customers to put their data on Google Cloud Storage. Others include Google BigQuery Data Transfer Service to automatically load data into the query service, and Cloud Storage Transfer Service to move data over the internet from cloud to cloud.
In the past, Google has touted its network capabilities as a differentiator, saying its global fiber was sufficient to meet customer demands. This move is also an acknowledgement of the limits to move data from a private data center.
"It's just a reflection of the customers that Google is now addressing," said Dave Nettleton, lead product manager for Google Cloud Storage. "As Google is building and getting much more traction with traditional enterprises, we see net new use cases we need to address."
* Information updated after publication
Trevor Jones is a news writer with SearchCloudComputing and SearchAWS. Contact him at [email protected].
Explore Google's lineup of cloud storage services
Look into offline data migration services from top cloud providers
Set up a virtual private cloud network with Google