GCP RESOURCE MANAGER- As an engineer when you are designing an architecture within google cloud, based on the google cloud global structure and the resource you want to create, that would determine the placement either a region or zone etc. This will determine the level of availability when it comes to the resources.
Google cloud has provided a global structure which is divided into regions and zonal ( AZ). A region is like New York, North Virginia, California, Texas while availability zone are smaller part within the region, the availability zones are powered by those data centres and within the availability zone you can have AZ1 AZ2 AZ3 and more, depending on google cloud services and user to either increase the AZ.
The data centres are the actual location where the actual systems resides, all the virtual machine, network resources, database systems, storage facilities, all runs within the particular data centres that are within available zone that are within regions. This is the hierarchy. Within regions you have availability zones which are deployment areas for compute related resources Google cloud has And if one AZ goes down, it will not affect their whole structure.
One of the pillars in google cloud provided is Reliability. There were two main point, FAULT TOLERANT AND HIGH AVAILABILITY. Which means creating your resources across multiple areas but it will include other practises which will make your environment to be resilient.
You might be asked why one data centre goes down, and it will not affect the other. The reason is because there are certain distance given to a new data centre and the existing data centre within google cloud. And its designed that each centres are completely dependent. The network connection will be isolated from data centre A - B. Natural disaster, power failure can cause AZ to go down.
For example, Based on the above structure, you created a VM, placed in ZONE A, and colleague mistakenly terminate ZONE A . Will you be able to login through ZONE B or will the resource be lost completely? In this scenario, your virtual machine is completely gone and that is because you specified that AZ unless you design a disaster recovery architecture to take BACK-UP, SNAP SHOTS, IMAGES etc. By default google cloud has that automatic replication across the board based on the resources.
CLOUD SQL DATABASE : Are regional not zonal, like RDS. When we talk about high availability, a service like load balancing which is an interface to distribute request to the backend. Each of the load balancing has IP to entry access and exit point of your systems. We will elaborate more when we get to networking.
For example, Take for instance, a massive storm occurred and a region goes down and nothing is assessable within the region. In this case, you are designing a DUAL REGION SOLUTIONs. With GCS you can use it to provide resources that span across two different regions. Your data will be replicated and stored across two different regions.
MULTI--REGION RESOURCE PLACEMENT: This helps you to create more region in which you have multiple AZ. And the services you need to achieve this is GOOGLE CLOUD STORAGE.
GLOBAL STORAGE: When you create a resource in any of the domain within google cloud by default that network is global. it spans across multiple regions. Even if the two structure above ( region) goes down. It will NOT affect the resource because it is global. A CLOUD VPC network is global not regional not AZ specific. Other cloud providers does not have this when it comes to networks.
Next week, we will buttress on the resource hierarchy. Direct link below. Happy Learning!!😊
References :
Questia
IAM & Resource hierarchy :
https://www.blogger.com/blog/post/edit/5428112557550405099/2839242659608842442
google cloud:
No comments:
Post a Comment