GCP (Google Cloud Platform) is a cloud service operated by Google. In this article, we will explain the features, how to use each service, what you can do, and the advantages of introduction compared to AWS. If you are interested, please refer to it.
1. What is GCP (Google Cloud Platform)?
GCP is an abbreviation for “Google Cloud Platform,” a cloud service operated by Google. Google operates a search engine “Google” and a video sharing platform “YouTube”, and uses GCP as its infrastructure. In other words, even general companies and individual developers can realize the same infrastructure environment as Google by using GCP.
By the way, GCP is often compared with “AWS” operated by Amazon. Both AWS and GCP are world-famous public clouds, and each company often discusses which one to use when introducing cloud services.
In conclusion, GCP is cheaper for simple services. On the other hand, in terms of robustness and generous Japanese language support, AWS is often favored.
1.1 Examples of typical services of GCP (Google Cloud Platform)
We would like to introduce two typical services of GCP.
■ Google App Engine
Fully managed application platform. The standard application languages available are PHP, Python, Ruby, Node.js, Java, C # and Go.
It also comes with a Docker container that allows you to install your own language and library framework.
■ Google BigQuery
A multi-cloud data warehouse that supports large-scale data analysis. It has a built-in machine learning function and can be flexibly scaled up according to the amount of analysis data.
In addition, according to a survey conducted by The Enterprise Strategy Group, Inc, the total cost of holding large-scale analytics over a three-year period can be reduced by 26% to 34% compared to other cloud data warehouses.
1.2 Is GCP (Google Cloud Platform) charged? free?
GCP has a free tier for each service. Therefore, if you use it within the free tier, you can use it for free.
You will be charged if you use the service beyond the free tier. The fee structure is a pay-as-you-go system, and you will be charged for the excess.
1.2.1 Free trial and free tier
GCP offers two free trials and free slots. A free trial is a free trial service worth $ 300 for 90 days after you start using GCP.
The free tier provides a separate usage amount that can be used free of charge for each GCP service.
As an example, I will introduce a typical App Engine and BigQuery free tier.
■ App Engine
| resource | Free allocation |
|---|---|
| “F” instance | 28 hours free per day |
| “B” instance | Free for 9 hours per day |
| Data transfer (downlink, outward) | 1GB of daily traffic is free |
See also: App Engine resources
■ BigQuery
| resource | Free allocation |
|---|---|
| storage | 1TB free per month |
| “B” instance | Free for 10GB per month |
See also: BigQuery resource
1.2.2 Estimated cost
GCP costs are calculated on a pay-as-you-go basis for each product. Please check the official website for details on each product .
An English version of the pricing tool is also available. With this tool, you can simulate costs based on what you use.
▸ Price calculation tool
2. [By service] What you can do with GCP (Google Cloud Platform)
![What is GCP? Explaining how to use, what you can do, and merits of introduction by features and services [Freelance engineer project information | Professional engineer]](https://proengineer.internous.co.jp/topics/wp-content/uploads/2020/09/column_image17019_2.jpg)
There are many GCP products and services, and many people don’t know what they can actually do. Here, we will introduce what you can do with GCP for each service that has many GCP users.
Unless otherwise specified, the price list introduced in each item is based on the assumption that the Tokyo region is used.
2.1 Computing
GCP has several computing services that provide scalable virtual machines. Here, we will introduce three services.
2.1.1 Google Compute Engine
Source: Compute Engine ｜ Google Cloud
Google Compute Engine is a service that provides compute resources in the form of IaaS (Infrastructure-as-a-Service). Virtual machines can be combined with OS, CPU, memory, etc. according to the purpose.
In addition, virtual machines are built on Google’s infrastructure, so you can take advantage of large workloads, performance, and stability.
In addition, the cost is a pay-as-you-go system that is charged according to the specifications of the environment and the usage time.
The hourly rate is as follows.
| Machine type | Number of virtual CPUs | memory | Fee (US $) | Preemptive fee (US $) |
|---|---|---|---|---|
| n1-standard-1 | 1 | 3.75GB | $ 0.06 | $ 0.01 |
| n1-standard-2 | 2 | 7.5GB | $ 0.12 | $ 0.03 |
| n1-standard-4 | Four | 15GB | $ 0.24 | $ 0.05 |
| n1-standard-8 | 8 | 30GB | $ 0.49 | $ 0.11 |
See also: General purpose machine type
2.1.2 Google App Engine
Source: App Engine ｜ Google Cloud
Google App Engine is a platform for PaaS (Platform-as-a-Service) format application development and hosting. It provides a mechanism to easily start the application.
In addition, the layers below applications such as middleware are completely managed by Google App Engine. Therefore, the feature is that the burden on the developer such as environment update and patch application is reduced.
The hourly rate is as follows.
| Machine type | Number of virtual CPUs | memory | Fee (US $) | Preemptive fee (US $) |
|---|---|---|---|---|
| n1-standard-1 | 1 | 3.75GB | $ 0.061 | 0.01325 |
| n1-standard-2 | 2 | 7.5GB | $ 0.122 | 0.0265 |
| n1-standard-4 | Four | 15GB | $ 0.244 | 0.053 |
| n1-standard-8 | 8 | 30GB | $ 0.488 | 0.106 |
Reference: General purpose machine type family
2.1.3 Google Kubernetes Engine
Google Kubernetes Engine is a cloud service that provides an execution environment for Docker containers. By creating a Kubernetes cluster on Google’s infrastructure and deploying a Docker container, you can deploy the container in seconds.
Two types of operation modes are provided, and the charge system changes depending on the mode.
[Autopilot mode]
This is a fully provisioned cluster configuration mode. Cluster configuration options are created automatically.
[Standard mode]
This mode allows the user to flexibly set the infrastructure.
The hourly rates are as follows. The price will be in Autopilot mode.
| item | price | 1 year guaranteed discount price (US $) | 3-year guaranteed discount price (US $) |
|---|---|---|---|
| GKE Autopilot vCPU charge (vCPU-hours) | $ 0.06 | $ 0.05 | $ 0.03 |
| GKE Autopilot Pod Memory Charge (GB-hours) | $ 0.01 | $ 0.01 | $ 0.00 |
| GKE Autopilot Ephemeral Storage Fee (GB-hours) | $ 0.00 | $ 0.00 | $ 0.00 |
See: Autopilot mode pricing
2.2 Storage
Storage is the place where you store your files and data. GCP is developing various services to store files and data on the cloud.
Here, we will introduce three famous services.
2.2.1 Google Cloud Storage
Google Kubernetes Engine is a cloud service that provides an execution environment for Docker containers. By creating a Kubernetes cluster on Google’s infrastructure and deploying a Docker container, you can deploy the container in seconds.
Two types of operation modes are provided, and the charge system changes depending on the mode.
[Autopilot mode]
This is a fully provisioned cluster configuration mode. Cluster configuration options are created automatically.
[Standard mode]
This mode allows the user to flexibly set the infrastructure.
The hourly rates are as follows. The price will be in Autopilot mode.
| item | price | 1 year guaranteed discount price (US $) | 3-year guaranteed discount price (US $) |
|---|---|---|---|
| GKE Autopilot vCPU charge (vCPU-hours) | $ 0.06 | $ 0.05 | $ 0.03 |
| GKE Autopilot Pod Memory Charge (GB-hours) | $ 0.01 | $ 0.01 | $ 0.00 |
| GKE Autopilot Ephemeral Storage Fee (GB-hours) | $ 0.00 | $ 0.00 | $ 0.00 |
See: Autopilot mode pricing
2.2 Storage
Storage is the place where you store your files and data. GCP is developing various services to store files and data on the cloud.
Here, we will introduce three famous services.
2.2.1 Google Cloud Storage
Cloud SQL is a cloud service that provides relational databases such as MySQL, PostgreSQL, and SQL Server. Because it is fully managed, maintenance costs such as patching can be reduced.
The advantage of using Cloud SQL is not only high availability and performance, but also that it can be easily linked with other GCP services such as Compute Engine.
Cloud SQL pricing consists of three pricing items.
<Hourly charge for CPU and memory>
| Fee (US $) | 1 year guaranteed usage discount | 3-year guaranteed usage discount | |
|---|---|---|---|
| vCPU | $ 0.0537 per vCPU | $ 0.04 | $ 0.03 |
| memory | $ 0.0091 / GB | $ 0.01 | $ 0.00 |
| HA vCPU | $ 0.1074 per vCPU | $ 0.08 | $ 0.05 |
| HA memory | $ 0.0182 / GB | $ 0.01 | $ 0.01 |
<Hourly storage and network charges>
| price | ||
|---|---|---|
| storage | SSD storage capacity: $ 0.221 / month HDD storage capacity: $ 0.117 / month Backup capacity: $ 0.104 / month per GB (usage) |
|
| HA storage | SSD storage capacity: $ 0.442 / month HDD storage capacity: $ 0.234 / month per GB Backup capacity: $ 0.104 / month per 1GB (usage) |
|
| Serverless export | Free until February 1, 2021, after which $ 0.01 per GB of instance storage capacity |
|
| network | Up to Cloud SQL (inward): Free * IPv4 address: Idling $ 0.013 / hour |
|
* Downstream (outward) charges may apply to the source. For example, outbound traffic from the Compute Engine is charged at the rate of the external IP address.
Downbound from Cloud SQL (outward): See network outbound (outward) charges
<Hourly charge for instance>
| Destination | price |
|---|---|
| Compute Engine instance and Cloud SQL cross-region replica | Within the same region: Free Between regions within North America : $ 0.12 / GB Between regions outside North America: $ 0.12 / GB |
| Google products (excluding traffic to Compute Engine and Cloud SQL cross-region replicas) | Intracontinental: Free Intercontinental : $ 0.12 / GB |
| Internet downlink (outward, when using Cloud Interconnect) | $ 0.05 / GB |
| Internet downlink (outward, when not using Cloud Interconnect) | $ 0.19 / GB |
See: Cloud SQL pricing
2.2.3 Google Cloud Datastore
Google Cloud Translation API is a service that provides API for Google Translate. We support more than 100 languages and provide 3 types of APIs to suit your needs.
<Translation API>
Translation API is an API that translates sentences into more than 100 languages using Google’s neural machine translation technology. There are two types of APIs, Basic, which allows you to use standard functions, and Advanced, which provides customization functions in addition to standard functions.
<AutoML Translation>
AutoML Translation is an API that allows you to set up your own translation model. It is effective when specialized knowledge is required or when you want to deal with slang.
<Media Translation API>
The Media Translation API is an API that enables real-time speech translation.
Google Cloud Translation API pricing is based on the API used and the amount specified. The charge system differs depending on each API, so if you want to know the details, please check the charge of each API.
▸ Translation API pricing page
▸ AutoML Translation pricing page
▸ Media Translation API pricing page
3. Benefits of introducing GCP (Google Cloud Platform)
![What is GCP? Explaining how to use, what you can do, and merits of introduction by features and services [Freelance engineer project information | Professional engineer]](https://proengineer.internous.co.jp/topics/wp-content/uploads/2021/09/column_image17019_3.jpg)
The benefits of introducing GCP include:
• Infrastructure equivalent to Google search, YouTube, and Gmail can be used
• No initial investment in infrastructure required• Lower overall cost than AWS
• Abundant machine learning related services
I will explain each in detail.
3.1 Google search / YouTube / Gmail equivalent infrastructure can be used
The first advantage of GCP is that you can use the same infrastructure as Google services such as Google search and YouTube / Gmai at low cost.
Google’s services perform large-scale processing speedily, and can be said to be a highly reliable infrastructure.
• Google Search handles billions of searches instantly
• YouTube plays 6 billion hours of video a month
• Gmail provides storage for 425 million users
It takes a huge amount of time and cost to build and operate a highly reliable infrastructure from scratch, but with GCP, you can use it by paying only for what you use when you want to use it, so you can save time and cost.
3.2 No initial investment in infrastructure required
GCP already has a reliable infrastructure, so you don’t have to spend time and money building it in advance. In an on-premises environment, securing a monitoring system and personnel is a bottleneck, but such fixed costs can also be reduced.
3.3 Easy to scale up and down
All GCP services can be scaled up or down according to your needs. Also, depending on the service you are using, if you set it to adjust automatically, it will automatically scale up and down according to fluctuations in the number of accesses and the number of processes.
Charges are also calculated based on machine resource status and usage. Therefore, you will be charged only for what you use, so you do not have to worry about overpayment.
3.4 Overall lower cost than AWS
The cost performance of GCE instances is particularly good, and long-term discounts are automatically applied. In addition, since you can get free credit for 30,000 yen, you can operate it virtually free for the first few months to a year, especially for small sites and apps.
In addition, the server is scalable, so it can handle a rapid increase in access. Since GCP is a service provided by Google, it can also provide other services, and there is also the advantage that it is not complicated because google is the contact point in terms of support.
GCP can reduce operational costs by reducing complex requirements by keeping in mind a simpler design and operation than AWS.
3.5 Abundant services related to machine learning
You can easily get the infrastructure you need to use TensorFlow. TensorFlow is a library released by Google as open source for use in machine learning. All the infrastructure is left to GCP, so you can focus on implementing machine learning algorithms.
4. Disadvantages of introducing GCP (Google Cloud Platform)
![What is GCP? Explaining how to use, what you can do, and merits of introduction by features and services [Freelance engineer project information | Professional engineer]](https://proengineer.internous.co.jp/topics/wp-content/uploads/2020/05/column_image15475_8.jpg)
There are two disadvantages to introducing GCP:
• There is little information in Japanese
• There are few regions
The first disadvantage is that there is little information in Japanese.
There are few Japanese translations of GCP official documents, and Japanese translations are a little difficult to read because they are translated by Google Translate. Also, since there are few Japanese engineers who use GCP and there are few technical articles in Japanese, it will be a little difficult for engineers who are not good at English.
The second disadvantage is “fewer regions”. Since there are few regions, it will be necessary to take measures when expanding globally or responding to failures. Also, be aware that the types of services available vary depending on the location of the region.
5. Comparison of GCP (Google Cloud Platform) and AWS (Amazon Web Services)
![What is GCP? Explaining how to use, what you can do, and merits of introduction by features and services [Freelance engineer project information | Professional engineer]](https://proengineer.internous.co.jp/topics/wp-content/uploads/2020/09/column_image17019_4.jpg)
You may be wondering which is better, GCP or AWS. The bottom line is case by case. AWS is good for “reliability and robustness”, and GCP is good for “simple and low cost” systems. I will explain in detail.
5.1 AWS if reliability and robustness are required
AWS is feature-rich and flexible, and if you understand the features and build them thoroughly, it will be easier to operate and manage. You can flexibly customize the configuration and security, so if you need reliability and robustness, AWS is a good choice.
However, be careful not to set it too complicated.
5.2 GCP for a simple and low cost system
GCP permission management is simpler and easier to understand than AWS. Therefore, if you want to operate a simple and low-cost system, GCP is a good choice.
However, complicated authority management is not possible, so it is necessary for the administrator to design the policy definition to check the security requirements in advance.
6. GCP and AWS service correspondence table
| GCP | AWS | |
|---|---|---|
| IaaS | Compute Engine (GCE) | Amazon EC2 (EC2) |
| PaaS | App Engine (GAE) | AWS Elastic Beanstalk |
| Object Storage | Cloud Storage (GCS) | Amazon S3 (S3) |
| Load Balancer | Cloud Load Balancing | Elastic Load Balancing |
| Virtual network | Virtual Private Cloud (VPC) | Amazon VPC |
| DNS | Cloud DNS | Amazon Route 53 |
| CDN | Cloud CDN | Amazon CloudFront |
| WAF | Cloud Armor | AWS WAF |
| Single sign-on | Google Workspace | AWS Single Sign-On (SSO) |
| Monitoring / Logging | Cloud Monitoring | AWS CloudWatch |
| cron | Cloud Scheduler | AWS Batch |
| Infrastructure as Code (IaC) | Cloud Deployment Manager | AWS CloudFormation |
| CI / CD | Cloud Build | AWS CodePipeline |
| API development / management | Cloud Endpoints | AWS API Gateway |
7. Trends in job offers related to GCP (Google Cloud Platform)
![What is GCP? Explaining how to use, what you can do, and merits of introduction by features and services [Freelance engineer project information | Professional engineer]](https://proengineer.internous.co.jp/topics/wp-content/uploads/2020/08/column_image16488_3.jpg)
The professional engineers operated by our company have posted a large number of GCP related projects.
There is a need for GCP skills in a wide range of industries, including BI tool development projects for the pharmacy industry and development projects for well-known economic information platforms. With the expected need to operate the cloud at low cost, there will be strong needs for GCP deals.
▸ Search for GCP related deals with professional engineers
8. Summary
This time I explained about GCP. GCP is superior to AWS in terms of cost. However, AWS will be higher in terms of reliability and abundance of information in Japanese.
There is a strong need to operate the cloud at low cost, so strong needs are expected. We hope that you will refer to this article and consider which one is more suitable for you.