The Ultimacy of Cutting Down Cloud Costs

One of the significant advantages of utilizing cloud technology is the ability to easily adjust the scale of resources based on demand, resulting in reduced operational costs. This is particularly valuable when facing unforeseen fluctuations in customer demand.

In Incentro Africa, we have a dedicated team of Solutions Architects who specialize in assisting companies with optimizing their cloud operating expenses.

Throughout our years of collaboration with companies across 34+ countries in Africa, we have established that optimizing your Google Cloud necessitates a continuous and iterative approach, involving round-the-clock monitoring of cloud resources, aimed at minimizing cloud computing waste without compromising on operational requirements and ultimately achieving the desired business value.

Optimizing your Google Cloud offers several immediate advantages, including:

  • Enhanced visibility into the latest cloud cost trends and forecasts.

  • Increased accountability throughout the organization.

  • Strengthened cost governance policies and improved control over the cloud infrastructure.

  • Efficient usage and cost optimization through real-time intelligence and automated recommendations.

GCP stands out from its competitors with its highly competitive pricing, constantly adapting to market dynamics through annual changes to its pricing model. Initially, GCP users were billed per hour and minute. However, when AWS introduced the concept of per-second billing, GCP swiftly followed suit to provide even greater business value competitively. GCP went a step further by offering second-wise billing, surpassing AWS, as this pricing model isn't limited to specific instance types. In contrast, AWS's per-second billing option is only available for Linux-based compute instances.

In partnership with Incentro, Google Cloud has the following pricing models:

1. Free Tier

The Free Tier is designed for individuals and organizations exploring the world of cloud computing and various cloud vendors without an immediate need for extensive platform usage. This tier offers access to a broad selection of predefined resources, some of which are always available for free. Within the Free Tier, users can utilize 24 different Google Cloud services and solutions and receive a $300 free credit to explore paid services. Each service and product in this tier comes with a predetermined monthly usage limit.

The GCP Free Tier is particularly suitable for users with low usage requirements who are open to occasional service disruptions as they familiarize themselves with a range of Google products, including IoT, AI, compute, database, storage, and more.

2. Usage-Based Billing

Google Cloud's pay-as-you-go is an on-demand pricing model, well-suited for organizations with infrequent platform usage. This option provides the flexibility to add or remove services as needed, but it comes at a higher cost, making it the most expensive pricing option.

3. Long-term Commit Billing

Incentivizing and acknowledging users for making long-term commitments, Incentro & Google Cloud provides more significant discounts compared to the pay-as-you-go option. Google's long-term pricing, known as Committed Use, applies to commitments ranging from 1 to 3 years. By opting for long-term pricing, users can potentially save up to 70% of their computing costs compared to on-demand pricing.

The key elements of Google Cloud pricing are:

  • Storage

  • Networking

  • SQL and

  • Compute

The foundation of Google's optimization principles revolves around three core objectives: enhancing visibility, optimizing resource usage, and improving price efficiency. All of their cost optimization tools are aligned with these three key objectives to deliver favorable outcomes for their customers. Now, let's delve into how these principles work in practice:

1. Enhancing Visibility

This makes it possible to:

  • Foster shared accountability across teams.

  • Conduct regular cost reviews to assess expenditure.

  • Analyze historical and current trends in costs.

  • Visualize the financial and operational implications of cost-related actions.

  • Align total costs with the operational structure.

  • Establish a chargeback model for cost allocation.

  • Employ budget alerts, quota limits, and other tools for effective cost management.

2. Optimizing Resources

This makes it possible to:

  • Employ precise benchmarks to strike a balance between cost and performance.

  • Assess services for application deployment to ensure efficiency.

  • Select the appropriate custom VMs for optimal resource utilization.

  • Utilize Google Cloud Recommendations to identify idle resources or VM instances that are under or over-provisioned.

3. Enhanced Price Efficiency

This makes it possible to:

  • Employ applicable discounts, including:

  • Sustained and committed use discounts

  • Flat-rate pricing

  • Per-second billing

  • Collaborate with technology teams such as CCoE or in-house FinOps teams to achieve significant savings while optimizing usage across business units.

1. Google Cloud's Cost Management Suite

This free application is accessible to any GCP user through the Google Cloud Console, though it may interact with other GCP services that could incur charges. For instance, the tool can utilize Pub/Sub to configure and send budget alerts. Additionally, Google has recently introduced a new tab in the Console to optimize Google Kubernetes Engine (GKE), providing benefits to all Kubernetes users. The console boasts several essential features, including:

  • Recommendations for usage optimization

  • Insights and predictions regarding cloud cost trends

  • Analysis of spending patterns

  • Alerts for actual and forecasted thresholds

  • Metrics for policies and authorization to facilitate financial policy implementation.

2. Google Cloud's Pricing Calculator

The Google Cloud Pricing Calculator is a crucial cost management tool that empowers organizations to create precise estimates for specific GCP solutions. With a user-friendly interface, the GCP Pricing Calculator enables users to group cloud resources, specify configurations with cost impacts, select desired timeframes, and generate accurate cost estimates.

3. Cloud Billing Reports
  • Google Cloud Billing Reports is a crucial built-in tool accessible from the Google Cloud Console. This analysis tool offers users visibility into cost history, cost drivers, cost trends, and cost predictions through a variety of user-friendly reports. Some of the reporting mechanisms of the billing reports and what you can achieve with them include:

  • Cost Table Report: Offers detailed access and analysis of invoices and statements down to the finest details.

  • Cost Breakdown Report: Presents a waterfall view of monthly charges, credits, and savings, providing a clear overview of expenses.

  • Pricing Table Report: Provides in-depth pricing information on SKU prices associated with selected billing accounts.

  • CUD (Committed Use Discounts) Analysis Reports: Enables understanding of CUD impact, identification of additional cost-saving opportunities, and visualization of commitment effectiveness.

  • Custom Billing Reports: Empowers users to visualize custom cost information using Google Data Studio, export detailed reports with BigQuery, and access billing data for further analysis.

With the help of these detailed reports, users can pinpoint the products or locations that significantly contribute to their Google Cloud costs. This information allows them to effectively manage their spending by organizing it according to projects, folders, and labels.

While right-sizing VM instances and terminating zombie assets are well-known and commonly practiced by organizations, let's delve into some lesser-known but effective strategies. Here are some lesser-known but effective cost-saving practices that Incentro suggests:

  • Remove unattached block-level storage discs:

Identify and terminate unattached block-level storage discs that may continue to incur expenses even after VMs are discontinued.

  • Get rid of unused IP addresses:

Regularly check for disassociated static external IP addresses and eliminate them to avoid unnecessary charges.

  • Schedule non-production virtual machines:

Optimize costs by scheduling non-production virtual machines to be active only during office hours, saving up to 65% to 80% on expenses.

  • Maximize Committed Use Discounts (CUDs):

For predictable workloads, commit to 1 to 3 years and take advantage of CUDs, which offer significant discounts of up to 57% without requiring upfront payment.

  • Utilize low-cost storage for infrequently used data:

Store less critical data in lower-cost storage tiers to cut down on expenses while still benefiting from various storage features and benefits.

The three key objectives we previously discussed for GCP cloud spend management—enhanced visibility, resource optimization, and price efficiency—apply to evaluating and setting standards for every crucial aspect of the infrastructure. This includes computing, network, and storage, which form the backbone of a GCP environment, as well as advanced analytics.

1. Compute Costs Optimization

Adopting Google Compute Engine is often the initial and primary step in Google Cloud migration, streamlining the process of procuring and setting up large VMs. Currently available in 23 regions and 70 zones, it offers a wide range of VM types. Additionally, GCP provides Custom Machine Types, allowing users to follow various processes, such as ideal VM recommendations, rightsizing VMs, object versioning, cloud scheduler, preemptible VMs, committed and sustained use discounts, and specific machine types. Organizations that have successfully achieved cost optimization understand how to leverage sustained use discounts while scaling their computing needs. Here are some recommendations from Incentro:

  • Get acquainted with the various VM instance pricing types, the Compute Engine billing model, and resource-based pricing.

  • Leverage Billing reports to enhance visibility into your Compute Engine expenses. Customize your views based on projects, labels, etc. for detailed insights into different machine types, CUDs, cloud usage, and more. Export this data to BigQuery for further analysis opportunities.

These are ways in which you can optimize your compute engine:

  1. Achieve cost efficiency by paying only for the necessary compute resources, identifying idle VMs, and automating their scheduling for auto start and stop.

  2. Consider purchasing commitments for frequently used VMs and allocate additional VMs for seasonal bursts, following a 70:30 ratio to optimize spending.

  3. Automate spend optimization using tools like Cloud Functions to clean up overprovisioned, underutilized, or unattached resources, and unused IP addresses, optimizing every cloud function.

  4. For fault-tolerant workloads such as big data, CI/CD pipelines, machine learning, data analytics, and media transcoding, utilize short-lived preemptible VMs to save costs.

  5. Choose autoscaling to handle sudden surges in traffic efficiently by dynamically adjusting resources based on monitoring metrics, CPU utilization, or load balancing capacity.

2. Achieve Storage Cost Optimization

Regardless of a company's size, data storage is a fundamental need. For GCP users, Google Cloud Storage provides a scalable and extensive storage solution. However, not all data is accessed with the same frequency; some are used daily, while others are accessed infrequently. Nevertheless, both types of data contribute to your Google Cloud costs. Therefore, periodically cleaning up unnecessary data is essential.

  • During GCP migration, our first suggestion is to clean up data and, while migrating, select appropriate storage classes based on budget, resiliency, and durability. For example, using a multi-region standard class will store data in one of the expensive multi-region locations.

  • Consider retention, access patterns, and performance when choosing cloud storage options.

  • Optimize storage costs by utilizing object lifecycle management rules for storage buckets to effectively manage data and save on storage expenses.

3. Network Costs Optimization

The network serves as the third pillar of the backbone, analogous to the four wheels of a four-wheeler. Google offers its users a global, scalable, flexible, and secure network for migrating and managing next-gen workloads and various services. However, organizations must invest in designing a robust, reliable, and cost-effective cloud network architecture and function.

Here are some approaches to building a cost-effective network:

  1. Understand network traffic flows: Utilize VPC Flow Logs and export data to BigQuery via Cloud Logging to gain insights into traffic trends and make informed decisions.

  2. Analyze global deployments: Use Network Topology to assess network charges in different regions and identify inefficient deployments or 'top talkers,' enabling cost optimization of network egress.

  3. Evaluate the need for VPN: Decide whether a VPN is necessary based on the requirement to upload large data volumes daily, as it can impact costs.

  4. Utilize Network Service Tiers: Optimize your network by leveraging Network Service Tiers to strike a balance between performance and cost-effectiveness.

  5. Optimize network usage: Make use of tools like Private Google Access and external IP addresses to optimize network utilization and minimize costs.

Are you ready to kickstart your Cloud Optimization journey? Get in touch through [email protected]