Mastering Cloud Economics with Google Cloud Tools

Cloud computing holds the promise of revolutionizing the way businesses compute, store, and process data by offering scalable, flexible, and cost-effective solutions that enhance efficiency, collaboration, and innovation. Furthermore, the Cloud has transformed IT costs from a capital expense (CAPEX) to an operating expense (OPEX).  However, many businesses are challenged with optimizing cost in the cloud and managing their spend.

This post will explore the arsenal of native Google Cloud tools that are available to master cost management in the cloud—especially if you leverage services such as FinOps.

Choosing the right compute engine instance

With Google Cloud, you have options for compute engine instances, including on-demand and spot/preemptible instances. Basically, an instance is a virtual machine (VM) hosted on Google infrastructure, and you can create an instance using the Google Cloud console, the Google Cloud CLI, or the Compute Engine API.

Using on-demand instances allows you to pay for the capacity you need by the second, at a fixed rate—without having to make any long-term commitments. This option could be a good fit for short-term or unpredictable workloads. You’ll need to choose a cloud region that best suits your requirements, balancing network latency, data residency, compliance, and service availability with the cost for each region. The Google Cloud Pricing Calculator and Network Intelligence Center can be used to estimate costs and analyze network performance across different regions.

Committed use discounts (CUDs) can also be used to unlock cost savings. CUDs provide deep discounts in return for committing to a minimum spend or resource usage for a term of either one or three years. CUDs can be flexible or fixed, which offers a tradeoff between flexibility and savings. With a flexible CUD, you’ll receive a predictable flat-rate discount based on a spent commitment (28% off for one year, and 46% off for three years) that apply across multiple VM families and regions.

Sustained use discounts (SUDs) are another option; they can be used on Compute Engine resources that make up more than one-quarter of a billing month (so long as you aren’t using any other discounts). The discount increases incrementally with usage, up to a 30% net discount for VM instances during that month.

Spot pricing, or preemptible instances, use excess capacity, so their availability varies with usage. There’s no guarantee an instance will be available when you need it, since preemptible instances operate on a bidding system. But they can provide significant discounts—to the tune 60% to 91%, according to Google. They’re best suited for fault-tolerant applications or batch-processing jobs.

You can also optimize costs by purchasing reservations for specific services like BigQuery, Cloud SQL, and storage classes. BigQuery offers three editions that support different workloads, as well as an on-demand model. Each edition offers different capabilities at a different price point; you can use editions and the on-demand model at the same time on a per-project basis, depending on what works best for you.

It’s also important to provide more visibility and raise awareness of cloud cost in the organization. This can be done in several ways, such as exporting billing data to BigQuery, creating FinOps dashboards using BigQuery data through Looker Studio (which is free) to raise awareness and identify anomalies, and using Google Cloud quotas and budget alerts to monitor runaway costs.

Saving through architectural efficiencies

The Google Cloud Architecture Framework provides design recommendations to help optimize the cost of workloads in Google Cloud. For example, it offers strategies when provisioning, monitoring, and managing resources.

Whether you’re doing this in-house, or with a third-party partner, architectural efficiencies can create long-term savings.

For example, Unilog—a global technology company specializing in e-commerce and product data management in the B2B marketplace—needed rapid deployments, unlimited scale, and reduced support costs for its customers.

While Unilog was struggling with administering and scaling client deployments to support its rapid growth, it also needed assistance in controlling the cost of its traditional infrastructure. So Unilog approached Pythian to modify its design, both to support and to optimize Google Cloud.

As a result, the company has seen up to 50% in cost reductions for Infrastructure as a Service (IaaS), as well as improved support costs and operating margins with multi-tenancy.

Optimizing workloads in Google Cloud with FinOps

FinOps, a blend of finance and DevOps, is also known as cloud financial management. For those without the internal resources to deploy, operate, monitor, and manage cloud workloads, Pythian offers both a self-managed and fully managed FinOps service.

Our self-managed FinOps service—free with a Google Cloud billing commitment to Pythian—provides visibility, analysis, and forecasting of Google Cloud costs and usage, as well as recommendations for cost optimization, reservations, and commitments.

Our fully managed FinOps service uses a powerful combination of advanced data analytics, AI/ML tools and human technical and business expertise to offer immediate, and often substantial, cost savings. With predictable billing, combined with cost-saving opportunities and architectural efficiencies, Pythian’s managed FinOps services can generate cloud cost savings as high as 20 to 35%.

And, in addition to managing your Google Cloud environment, Pythian can also manage your AWS, Azure, and/or OCI environments. With a unified view across all your clouds, you’ll understand what you’re spending, how you’re spending it, and how to optimize it.

Contact us to learn more about our FinOps Cost Management & Optimization.