Google this week announced substantial price cuts on local storage for customers running virtual machine instances on the company’s Compute Engine cloud platform.
Starting Aug. 8, enterprises will pay 63 percent less for local SSDs (solid-state disks) that are attached to what Google calls its on-demand Compute Engine VMs. That brings the price down to between $0.08 and about $0.11 per GB of data per month for most of Google’s cloud regions in the U.S.
Google has also slashed prices for storage associated with its preemptible VM instances. The preemptible VM is a very low-cost cloud option that Google has designed for customers looking to run certain types of workloads, such as batch processing and fault-tolerant applications in the cloud. Preemptible VMs cannot run for more than 24-hours and Google has the right to shutdown instances if it needs extra capacity for other workloads.
Prices for local SSDs associated with this class of VMs in most US regions is now between $0.064 and $0.086, or about 71 percent cheaper to Google’s previously quoted prices.
“At Google we’re always looking to reduce total cost of ownership for our customers,” said Google product managers Chris Kleban and Michael Basilyan in announcing the price reductions on Google’s Cloud Platform Blog.
The cuts are the result of technology advances that Google has achieved in field of cloud storage. They reflect the company’s commitment to adjust pricing so customers can benefit from technology innovations they said.
Google describes its local SSDs as a class of high-performance physically attached block storage for cloud VM customers. The storage is attached to the server running a customer’s VM instance and is designed to persist for as long as the VM instance exists on the system.
Local SSDs, according to the company have lower latency and better throughput compared to SSD persistent disks and standard persistent disks. But the performance gains come with certain trade offs in availability and durability, Google has noted.
Local SSD storage for instance cannot be automatically replicated so all data on it can be lost under certain circumstances. Google itself has recommended that companies using its Local SSD services to store important data also backup the data on a more reliable physical or cloud option in case of failure.
Google markets the Local SSD service to customers seeking to run workloads with extremely high bandwidth and input/output requirements. In preemptible VM environments Local SSDs are typically being used for applications like financial market analysis, genomic data analysis, movie rendering and media transcoding or digital conversion.
Google’s current Local SSD configurations start at 375 GB. Enterprises can add up to eight of the SSDs per VM instance, or up to 3 TB of local storage.