What’s the Big Deal about Google Nearline?

Google Nearline enters the cloud storage market

Google Nearline announced beta availability in March 2015 and general availability in July 2015. The timing was no accident. Google’s GA announcement dovetailed with Amazon’s success story earnings report. It sounds like an odd timing but Google garnered a lot of positive attention by riding on AWS’ announcement coattails. Google clearly means to compete with AWS and specifically positions Nearline against Glacier. They do not limit themselves to Glacier or to AWS for that matter; Google also offers a Switch and Save program that enables easy migration from S3 and MS Azure to Nearline.

Clearly Google did not want to play a commodity me-too game. Google Nearline fills a gap between low-priced cold storage in the cloud that has long restore times, and premium cloud services with high performance and a high price tag.

Cost and Performance

Nearline is part of Google’s Cloud Platform. Unlike Glacier, which runs on its own mystery disk environment (AWS isn’t talking), Nearline shares the same Google infrastructure as the rest of the Cloud Platform services. Google saves money by using the same storage environment as the rest of Google Cloud Platform. This saves money on economy of scale, by using the same APIs, encryption and other storage services, and not having to migrate data between standard storage and cold storage.

Cloud Storage Nearline and Glacier both charge about 3 cents per gigabyte per month to store data. Retrieval pricing gets a lot more complicated and both providers can make a case for being cheaper. However, Google promises recovery times in seconds while Amazon’s service takes hours to restore the same amount of data. This fast recovery is Nearline’s claim to fame in the cold storage market.

Base Nearline payment is calculated in units of 4 MBs per TB. Customers can choose to pay for an extra I/O feature that scales speed linearly with the amount of data the customer stores. These accelerated I/O speeds will speed up throughput for customer restores. For example, 3TB of stored data returns 12 MBs read times while 100 TB of data achieves 400/MBs.
However, nothing is perfect. Nearline claims 99% uptime and Glacier 99.99% uptime. Although we are not talking active production data here, if you store highly regulated documents with long retention periods you might consider Glacier instead.

Use cases

  • Low cost storage tiering. Nearline is inexpensive enough to use as a long-term repository for less active data, and is fast enough to recover data in a reasonable timeframe.
  • DR. Customers do not want to wait days to restore data and many of them balk at receiving hard drives full of data off the data truck. Nearline scales performance to the amount of storage, which protects against unacceptable recovery times with larger data sets. Extra I/O resources are granted from nearline buckets.
  • Retention management. Nearline can help customers manage retention lifecycles. Customers can manually submit data to Nearline and can also schedule automatic deletion and archiving.

What does this mean to you? That if you have data you are primarily keeping for regulatory reasons, and if that data cannot be lost, and if you already have a lot of data stored with AWS, then there is no compelling reason to switch services. However, if you have data that you want to store in the cloud and you need fast recovery, then Nearline’s superior performance is a much better deal.

Also consider integrations with existing data protection and sharing providers. A number of them signed up for a Nearline integration program during its beta period including Unitrends, Actifio, EMC, CommVault and Egnyte. If you already have a partner product installed on site, or are planning to do so, then this is another reason to consider adopting Nearline services.