I would like to learn how to design an infra-structure based on Google Cloud?

[Mikolaj Ruczynski]

I will have users around the world. The database has about 50 GB. Pageviews about 20 million per month. I have invented such services:

• GOOGLE APP ENGINE - Build scalable web and mobile backends in any language on Google’s infrastructure
• CLOUD SQL (database)
• CLOUD CDN (Low-latency, low-cost content delivery using Google's global network

I have the following questions:

1. How much does it need application servers?

2. In what locations do they have to be application servers?

3. Is there a concept of replication for application servers?

4. How much does it need SQL Servers and in what locations?

5. Is there a concept of rep for SQL Servers?

Thank you for your help

[Shivam(Google Cloud Support)]

Welcome to Google Cloud Platform. I am addressing the questions in the order asked:

1. With App Engine, you don’t have to worry about servers. Google handles most of the management of the resources for you. For example, if your application requires more computing resources because traffic to your website increases, Google automatically scales the system to provide those resources. If the system software needs a security update, that's handled for you too. 

2. You can currently run your app from the Central US or Western Europe regions.

3. App Engine will scale/spin up more instances based on the incoming traffic patterns to make sure that all the requests are served. App Engine is regional which means the infrastructure that runs your apps is located in a specific region and is managed by Google to be redundantly available across all the zones within that region.

4. You can choose storage size based on the requirements to start with and can later increase the size as the database will grow. You can also enable ‘Automatic storage increase’, which will check the available storage every 30 seconds and if available storage falls below a threshold size, additional storage capacity is automatically added to your instance.

 

5. All data is replicated in multiple locations for great availability and durability. Data is stored in datacenters in the EU or the US. You have the ability to replicate a master instance to one or more read replicas. A read replica is a copy of the master that reflects changes to the master instance in almost real time.

I hope this helps.