Why do We Select GCP?
The new Google Cloud Platform (GCP) region has been opened in Jakarta, Indonesia, since 2020. Currently, some PaaS/SaaS services, such as Amazon Web Service(AWS) and Microsoft Azure, are available there. But this time, I mention why we select GCP from many other services.
1. High Flexibility and Cost Reduction
Using GCP, we are flexible to scale our server to handle suddenly increasing users. Up until now, this took time and cost. With GCP, we save much time and cost as we don’t have to do the messy settings on-premises.
Of course, other cloud providers can also handle the scale management. However, GCP is a low-cost yet flexible server expansion. We can expect a 20–30% cost reduction since GCP sets are cheaper than other cloud providers. They provide discounts for continuous use too. One of the biggest attractions is the low price.
2. Environment Preparation & Deployment Speed
By using cloud computing, we can build, compile, and scale on our private network faster. The development environments on the cloud such as virtual machines, PaaS, storage, and data migration services, are fully prepared. Programmers can start development immediately. Database services such as Cloud SQL and Cloud Spanner, networking services of Virtual Private Cloud are also available. It is possible to immediately set up a development environment on each device.
3. Data Science
As data science has become a hot topic, more and more companies are dealing with data, and the data management cost is increasing. Those companies are looking for a way to do it in-house, and so do our company. We are trying to use Cloud Dataprep, which can explore and clean up data for analysis.
“BigQuery”- BigQuery is a so-called analysis platform of big data.
It can execute a batch to import CSV/ text data and compile terabytes of data inserted by JSON with SQL *1.
We can store everything such as logs and database data in BigQuery, visualize it on the data portal, and share it at meetings. Even non-engineers can start it immediately with client mode.
In addition, what is surprising is its speed and cost. A few GB of data is executed in a few seconds. If you collect 200–300MB of logs and daily reports every day, the storage fee is under 1 USD and the query fee can be covered free of charge*2.
*1 There are two types of BigQuery original and standard.
*2 BigQuery can be used from other clouds and on-premises. When it comes to the ease of authentication and network problems, there are many advantages to using GCP.
4. Advanced Technology
As I mentioned in the data science section, many cutting-edge technologies are usable other than statistical processing. For example, Google made MapReduce open-source for distributed processing framework and also made group open-source for grouping process management. Kubernetes, which is probably the most used container management platform, is also necessary for us.
In the field of machine learning, Google has been building its theme on “Democratizing AI”.
Now there are four services of AI.
・TensorFlow : an open-source machine learning software library
・Cloud ML Engine : allows you to learn TensorFlow using Google’s infrastructure
・ML APIs : a set of API services that can use trained models
・AutoML : that can be customized from the learning part and create its learning model
We already adopted Cloud Vision API, one of ML APIs, into our business to extract individual data from certifications.
In summary, GCP has a wide range of services, and the price is based on the usage fee. It is second to none when compared to other similar services.
It’s free to start with a trial. It’s worth a try!
