< Back to Blog

How To Reduce the Cost and Improve the Performance of Your GCP Infrastructure

Are you currently using 50TB or more in Google Cloud Platform (GCP)?  Are those workloads comprised primarily of analytics and transactional databases that power high performance critical applications, with some dev/test environments?

These environments typically require high performance SSDs combined with powerful compute engines. The resulting footprint can cost $80,000 a month or more, depending on how many replicas you use.

Reducing Cloud Spend with Silk on GCP

Silk’s Cloud Data Platform can reduce the cost of your GCP infrastructure while also increasing data performance. We do this by separating compute from capacity, removing the underlying interdependencies, and by implementing rich Tier 1 data services.

This can dramatically reduce your footprint and your GCP data costs by 30% or more.

Let’s take a look at a typical Silk customer: They have a high-performance data footprint of around 250TB, costing about $70,000 a month.  Even with a small number of replicas, their GCP bill can balloon to over $80,000 a month.  But by deploying Silk on GCP, our customer is able to turn off expensive cloud resources and reduce their monthly spend to about $20,000 a month! A huge savings with a significant increase in performance.

How does our platform do this?  Our secret sauce is leveraging existing cloud infrastructure in an extremely efficient manner to increase performance and reduce the footprint at the same time.

We deliver rich Tier 1 data services like in-line variable deduplication, in-line compression, pattern removal, zero detect, and thin provisioning.  Together, these services deliver major reductions in resource utilization (2x-4x) while increasing performance and reducing latency.

And there are more savings…

Our platform also eliminates the need to overprovision capacity to support high compute requirements. In GCP, performance thresholds are interdependent on the amount of compute and capacity provisioned. To support high performance, but low capacity, data sets, organizations must radically overprovision capacity — leading to additional costs and waste.

Silk solves this problem by disaggregating compute and capacity performance dependencies. You can now scale either compute or capacity independently, allowing for more granular control of provisioned infrastructure. With the Silk Cloud Data Platform, the smallest (or largest) data sets will get the same high performance.

Rich Tier 1 Data Services without a Hit to your Budget

In GCP, to create copies of your data, you make replicas. The more replicas you create, the more costs you incur—and it’s not a small cost increase. Using replicas can also add to your performance requirements, further increasing costs.  Silk gives you unlimited zero-footprint replicas at no cost. A win-win.

Finally, GCP uses a shared nothing architecture which restricts data resources and applications to one-to-one mappings natively.  You must add both compute and capacity resources for each new application. Silk provides a shared resource capability supporting one-to-many mapping, allowing the same resources to support hundreds of applications simultaneously.  This reduction in resource overprovisioning delivers even more saving.

The bottom line: we can save 30% or more off your public cloud data charges.

In today’s current climate, where we are all looking to reduce costs, these savings will flow right to your bottom line. That’s the power of the Silk Cloud Data Platform.