Google Cloud Pricing Guide: Models, Calculations, and Tips

Google Cloud is a collection of services that includes storage and compute resources, Infrastructure as a Service (IaaS), and Platform as a Service (PaaS), all provided by Google. Google offers several pricing options that cater to different demographics and use cases. These options include a free tier, pay-as-you-go instances, and discounted long-term reservations. Because so many pricing options are available, it can be difficult for organizations to predict how much their cloud usage will actually cost.

The Google Cloud Pricing Calculator can help organizations estimate costs and better manage their usage. In this guide, we’ll examine Google’s pricing models and share cost optimization strategies. We’ll also explore common pitfalls related to Google Cloud cost management.

If you’re using Google Cloud for any of your organization’s applications, take a look at Veeam Backup for Google Cloud. This solution offers end-to-end secure backups and fast, reliable recovery tools for Google Cloud instances.

Understanding Google Cloud Pricing

Rather than using a flat fee structure, Google charges for its cloud services based on usage. These are some common use cases, along with information on how fees are calculated for each one.

Compute: Compute Engine is a service used to run virtual machines (VMs) on Google’s infrastructure. Users are charged based on the amount of time the VMs run each month. Short-lived, preemptive instances carry a significant discount (up to 80%) and are a good choice for fault-tolerant workloads. Sustained use discounts are available in other instances, providing a smaller discount on VMs that run for more than 25% of the month.
Storage: The Google Cloud Storage program offers storage for large, unstructured data. Pricing is calculated based on the volume of data retained, network usage required to read or move the data, and the number/type of operations performed on the buckets. You’ll also incur retrieval and early deletion fees for data held in Archive, Coldline, or Nearline storage and for replicating data across regions. VMs can have persistent disks attached to them as well. This is a higher-performance storage solution with fees that are based on the size of the disk.
Networking: If you wish to use the Cloud CDN to speed up your website, you’ll pay based on the number of requests and the volume of data that’s sent to the cache (cache fill) or retrieved from it (cache egress). For transferring data within Google’s Virtual Private Cloud, you’ll page for ingress and egress traffic, with rates depending on the type of IP used, whether the traffic crosses regions, and whether the route the traffic takes involves leaving the Google network.

Google Cloud Pricing Models

Google Cloud offers a tiered pricing structure, including:

Free Tier Option

The free tier includes 20+ cloud services and products, each with a monthly usage limit. In addition to these offerings, new customers receive $300 of free Google Cloud credit, which can be applied to premium cloud services or usage that exceeds the free tier.

Pay-as-You-Go Option

This on-demand pricing model lets users pay for services based on their usage. For example, compute engine products are charged based on the length of time they run, while storage fees are calculated based on metrics such as the volume of storage used, network usage, and operations performed on that storage. The pay-as-you-go option offers the most flexibility, but it costs more than long-term reservations.

This model can be confusing for some users because the pricing structure differs across products. This usage-based pricing measures several factors, including computing time, data storage, data egress and ingress, operations performed on storage, instance size, and even how the data is transferred.

Long-Term Pricing

Organizations that plan to use the cloud to run their operations continuously may benefit from a long-term pricing plan. This plan requires the organization to commit to a cloud deployment of between 1 and 3 years. In return, they receive a committed use discount, which can be as high as 70%.

If you use multiple services on Google Cloud, it’s a good idea to read the pricing terms for each option and take advantage of the pricing calculator to estimate fees based on your projected usage. We’ll break down some of the factors that influence Google Cloud costs below.

Google Cloud Compute Pricing

Compute Engine allows users to create VMs that are tailored to their needs by customizing the machine’s GPU, RAM, storage, CPU, and networking. This service also offers several general-purpose machine types, as well as options for AI/machine learning, rendering, and other specialist workloads.

Google Cloud Computer machine configurations are divided into several different categories:

General purpose: Machines with pre-defined vCPUs and memory for general-purpose computing.
Memory-optimized: High-memory machines for data-intensive workloads.
Storage-optimized: Designed for workloads high in storage density but low in core usage.
Accelerator-optimized: Intended for GPU-accelerated workloads such as AI or machine learning.
Shared core: Intended for small and non-resource intensive applications. These machines timeshare a physical core, making them relatively inexpensive.

Prices are listed per hour/month, and some machine types and configurations offer a discount for long-term commitments, although this isn’t always available.

Google Cloud Storage Pricing

Google Cloud Storage pricing is based on the type of storage and its usage. Pricing elements include:

Data storage: This is based on the volume of data held in the bucket, with rates depending on the storage class chosen.
Network usage: Your network usage is based on the volume of data transferred between buckets over Google’s network. Fees are based on server location. For example, there’s usually no fee for transferring data within the same region using an internal IP address.
Operations: This metric measures the number of times certain activities, such as listing objects in a bucket, are performed. Operations are divided into two classes, with operations being more expensive to run on Class A than B.
Inter–region replication: Inter-region replication fees are based on the cost of replicating data across multiple locations (i.e., both the volume of data and respective locations).
Retrieval/early deletion fees: Archive, Coldline, and Nearline storage are designed for long-term file hosting of less-frequently-accessed data. This storage is less expensive than hot storage, but there’s a surcharge to access or delete data in these storage classes.
Google persistent disks: This storage product is designed to be attached to Google Cloud VMs and used for frequently accessed data. Persistent disks can be platter hard drives or SSDs. Pricing is based on GB/month of available storage, with optional add-ons, such as multi-region redundancy and additional IOPS/month.

The Google Cloud Free Tier offers up to 1GB of network egress and 5GB of storage per month, with a limited number of operations per month.

Google also offers a database service for MySQL, PostgreSQL, and SQL Server. Dedicated instances are priced based on the number of vCPUs and the amount of memory they have, and shared-core instances are billed based on how long they run. In addition to these fees, there are network and storage costs depending on the location of the instance. If you pull data from a Cloud SQL server using a client application that’s running on a machine and isn’t hosted in Google’s Network, you may be billed for internet egress.

Google Cloud Network Pricing

Fees for Google’s Virtual Private Cloud and CDN are both based on the type of traffic:

Data egress and ingress: Outgoing network traffic (i.e., egress) is chargeable in most cases. It’s important to be aware that if you transfer data between two VMs, the traffic is classified as egress traffic when it leaves the first VM and classified as ingress traffic when received by the second VM. Ingress traffic isn’t charged. However, there may be a fee for resource use when processing that traffic. It’s important to plan data transfers carefully to account for the charging rules. For example, if you transfer data between two VMs in the same zone and use the internal IPs, you won’t be charged. However, if you use an external (internet-facing) IP, you’ll be charged because the data will be routed over the internet.
Data egress when transferring to users: Google uses a tiered pricing system for transferring data between users and Google Cloud instances. The Premium tier, which is active by default, uses Google’s points of presence to provide the fastest service, but it costs more. The Standard tier uses public internet infrastructure, making it less costly, but Google has less control over the speed of the transfer.
Cross-regional transfer fees: Data transfers to Google Cloud instances in a different region are chargeable by volume. Rates vary depending on the source and destination regions.
Content Delivery Network: CDN services are priced based on the amount of data sent to/from the cache and the number of requests. Both cache fills (i.e., updating the CDN cache) and cache egress (i.e., filling user requests) are chargeable.

Google offers two tiers for data transfer between users and the platform’s cloud instances. The Premium tier relies on Google’s own infrastructure, routing data through the company’s points of presence to try to maximize performance. The Standard tier uses public ISPs. Transferring data in this way is less expensive, but Google can’t guarantee the speed and performance of the public network. Note that the Premium tier is activated by default for egress traffic.

Introducing the Google Cloud Pricing Calculator

The complexity of Google Cloud’s pricing and the difficulty of estimating your usage on a month-to-month basis means it’s important to carefully monitor your usage and take steps to optimize your cloud deployments. Simple changes, such as setting usage limits and shutting down idle resources, can save your organization a lot of money in the long term.

The Google Cloud Pricing Calculator is a free tool provided by Google. It’s kept up-to-date with the latest pricing changes, making it a reliable source of price information. This calculator is an invaluable resource for anyone who hosts data or applications on Google Cloud.

You can access the pricing calculator online, and you don’t need to be an active Google Cloud user to take advantage of it. You can use the tool to estimate the cost of running your chosen services on Google Cloud versus other cloud platforms or even on-premises.

The pricing calculator is easy to use and offers several handy features, including:

Instant estimates for pricing based on instance type, size, and configuration.
Coverage for all services, including storage, compute, databases, machine learning, etc.
A way to simulate different usage patterns or add and remove VMs and licenses to see how that affects your bills.
Cost estimates in your local currency.
Regular updates to reflect new pricing structures.
The ability to compare costs across regions.
API support.
Export estimates in CSV format.

Using these tools, you can get an at-a-glance overview of the likely costs of porting your workloads to Google Cloud. You can also forecast how changes to your workloads or Google’s pricing will impact you.

Navigating the Google Cloud Pricing Calculator

The Google Cloud Pricing Calculator is split into several tabs that represent different types of services, such as compute, cloud workstations, and storage. Each tab asks for details about the service and how you expect to use it. The information you provide is used to produce cost estimates.

Step-by-Step Guide to Using the Pricing Calculator

Open the calculator and select the service you’re interested in from the list of icons at the top.
You’ll be shown a list of questions relating to the service. In this example, we’ve selected Cloud TPU:
Choose the number of chips, generation, and class.
Select the location for the instance.
Enter the number of hours per day you expect the instance to be running.
Click “Add to Estimate.”
If you wish to add another service, such as storage, repeat this process.
Each time you click “Add to Estimate,” the estimated costs for the service will be shown on the right-hand side of the screen.
If you’d like to change the details of any estimate, click the “Edit” button under the product’s listing in the “Estimate” selection.
You can change the currency for the estimate by selecting the currency dropdown at the top of the estimate table.
Some products offer sustained use discounts or are included in the free tier. Where this is applicable, a checkbox will ask whether you wish to include these discounts in your cost estimates.
If there are forthcoming pricing changes, you’ll be alerted and shown an estimate for the price under the revised pricing structure.

The more information you provide, the more accurate the estimates will be. Try to have information about your usage habits available before using the calculator.

How to Create and Save Different Scenarios for Comparison

Once you’ve picked the products you’re considering and configured them, you can save or share the estimate for future use. At the bottom of the pricing estimate list on the right-hand side of the screen, you’ll find links to print or email the estimate, copy the URL, or download the estimate as a CSV to share with colleagues or import into external applications.

If you’re comparing prices for different products or configurations, saving the URL of each estimate makes it easy to flip between them in different browser tabs.

Exploring Advanced Settings for Accurate Estimates

As you work through the options in the pricing calculator, you’ll see you’re required to provide answers to many of the questions, such as the number of instances and the machine types for the Compute Engine product. However, many details can be left blank or on the default option.

To get the most accurate estimates, provide as much information as possible. If you’re planning a migration to the cloud, you may not know exactly how many requests you’re likely to send or how much data you’ll transfer, but use your best guess. If you already have workloads, you can take information from your admin panel to get a clearer picture of your usage.

Tips for Cost Optimization

Optimizing your Google Cloud workloads requires careful planning:

Choose the correct instance for the workload you’re running. Scale up when — and only when — you need more computing power/storage.
Use pre-emptible instances for noncritical workloads.
If you know you’ll be running an instance long term, take advantage of the sustained use discounts, which can be up to 70%.
Use compression to minimize storage/transfers where possible.
Optimize any code you run in the cloud to keep chargeable operations at a minimum and ensure processes shut down gracefully once they’ve completed.
Instead of storing snapshots for long periods, create a backup from a snapshot, and store the backup on a cheaper object storage tier.
Use Google Cloud’s Cost Management tools to automate resource management.
Set alerts and alarms for usage patterns outside the norm.
Review your instances regularly and downsize or shut down anything that’s not actively being used.
Review your cloud workloads when prices change to see if you could make any changes to save money.
Where appropriate, try routing traffic within Google’s network rather than the public internet to reduce data transfer costs.

Keeping Up with Pricing Changes

The prices for Google Cloud services are reviewed regularly. Google offers notice of pricing changes to give users time to adapt their usage if required. To avoid being surprised by any pricing changes, it’s a good idea to regularly review current prices and cost estimates. Even if Google hasn’t changed its prices recently, your usage may have changed, especially for user-facing applications.

Subscribing to the official Google Cloud Newsletter is a good way to get the latest news about any changes to the platform. In addition to the official newsletter, the independent GCP Weekly newsletter is a useful source of news and information about the platform.

Case Studies: Real-World Cost Savings with the Pricing Calculator

Even major tech-focused brands find it difficult to manage the costs of their cloud deployments. Zero-code app development company Apxor cut its costs by 30% when it switched from its previous cloud partner to Google Cloud while also benefiting from some significant performance improvements. The company used the switch as an opportunity to optimize its workloads, which include Compute Engine, Dataflow, and TensorFlow tools.

Arabesque AI, a financial asset management company, saw even greater savings. The company slashed its cloud bills by 75% with preemptible instances and dynamic scaling. This allowed it to use pay-as-you-go instances for maximum flexibility, maintaining high performance without paying more than it needed to for the workloads it was running.

Conclusion

Google Cloud is a useful platform for organizations looking for a scalable hosting service with an international presence. The variety of IaaS and PaaS products available on the platform makes it versatile and powerful. However, the pricing structure can be confusing. Learning how to estimate pay-as-you-go pricing with the Google Cloud Pricing Tool can help you manage your costs and decide whether a move to long-term pricing makes sense for your organization.

With careful planning, it’s possible to optimize your resource usage and your bills. If you choose to go this route, stay up to date with any announcements from Google regarding pricing changes. Try to follow best practices for security to reduce the risk of a breached account running up a significant bill. In addition, try Veeam Backup for Google Cloud, which offers free backups for up to 10 instances and powerful tools for hybrid and multi-cloud environments, helping you protect your valuable data.