Reducing prices for local SSDs in “crowded out” and on-demand instances

A week ago, on August 8, in our blog , we announced a reduction in prices for local SSDs. Now our customers pay 63% less for local SSDs , which are used in virtual machines at the request of the Google Compute Engine . The price in most regions will be $ 0.80 per 1 GB per month. Local SSDs used with preempted virtual instances will become even cheaper : the price will decrease by 71% and in most regions will amount to $ 0.064 per 1 GB per month.

We are constantly looking for ways to reduce customer operating expenses and talk about what we managed to do in this direction thanks to the development of technologies that allows us to reduce them. We adjust prices so that our customers can use the latest developments and create something new without extra effort.

Local SSDs are highly efficient, physically connected block data storage devices that continue to work as long as the instance exists. These drives support NVMe and ISCSI interfaces , they have high IOPS (I / O operations per second) and high bandwidth - what is needed for the most demanding programs. Often, local SSDs can be the optimal solution for temporary disks when caching layers or scaling databases, such as NoSQL.

A key feature of local SSDs is that you can assign any number of local SSDs to a machine of any configuration. You do not have a limit on the capacity of drives for a certain number and memory of virtual processors in virtual machines. In addition, local SSDs are available for the same instances for which GPUs are available . This provides additional opportunities for creating systems with high performance.

We have reduced prices not only for local SSD-drives for ordinary instances on demand, but also for drives for “crowded out” instances. Preempted instances are normal Compute Engine virtual machines, but they cannot run for longer than 24 hours. In addition, we can supplant (disable) a VM earlier if its resources are needed for other purposes. Thus, the capacities of our data center are used more efficiently - and you can use it too! You can request a specially allocated memory capacity of a local SSD-drive in order to take advantage of a “crowded-out” instance, even though you already have a allocated memory capacity of a local SSD-drive (for more detailed information click here ).

Google Cloud Platform customers turn to preemptive virtual machines to lower the cost of computing. In our practice, there are many interesting cases of use. Our clients use extruded VMs with local SSDs to analyze financial markets, process data, create special effects for movies, study the genome, transcode media and solve a variety of business or design problems using thousands of extruded virtual machines in one project.

We hope that lowering the price of local SSDs for virtual machines on demand and crowded out VMs will open up new opportunities for you to solve interesting problems in the field of business, design and science.

You can find more detailed information in articles about local SSD-drives and extruded virtual machines . You can find prices for SSDs on the corresponding Compute Engine page . Alternatively, you can use our price calculator . If you have any questions or would like to leave a review, go to the help page .

We are glad that you use our products for your development. If you want to share your impressions and show what you created using the Compute Engine, share it on Twitter , Facebook or G + .