Google said today that the persistent high IOPS solid state disk storage it showed off in June is now available in open beta. Compute Engine users can now try out SSD storage to help speed up data-processing applications as well as their webscale apps that are reliant on big databases.
Among the benefits of using the local SSD feature is the higher IOPS (input/output operations per second) than Google’s persistent disk and the lack of redundancy that can come with using spinning discs.
From the blog:
[blockquote person=”Google” attribution=”Google”]Performance scales linearly from 1 to 4 partitions. The full 4 partitions can execute up to 680,000 random 4K read IOPS and 360,000 random 4K write IOPS. This is 8x more write IOPS/GB and 15x more read IOPS/GB than SSD Persistent Disk.[/blockquote]
Google said that its local SSD feature is “very competitively priced” at $0.218 per GB/month, which translates to $0.0003 per GB/hour.
If [company]Google[/company] decides to do some maintenance in its data center, it promises that your local SSD data will not disappear as the company will use “live migration technology to move your VMs along with their local SSD to a new machine in advance of any planned maintenance, so your applications are not disrupted and your data is not lost.”
As Barb Darrow reported in the summer, Google’s SSD feature will duke it out with [company]Amazon[/company] Web Services’ provisioned IOPS EBS service. Darrow corresponded with Server Density CEO David Mytton who played around with the SSDs and wrote via email that “Performance is based on volume size and grows linearly, which makes it easy to predict cost and what throughput you need.”
Fast SSD storage is critical particularly for Cassandra, MongoDB and Hadoop workloads.
Mytton also wrote that Google’s price model was less complex than Amazon’s, which is something the search giant is trying to capitalize on to attract users with the cloud wars in full effect.