Multi-Cloud Notes - AWS, Azure, GCP Global Infrastructure

This is the continuation of a series of posts where I post notes on Multi-Cloud computing. I am trying to cover what I learn on AWS, Azure, GCP from various learning portals such as PluralSight, Udemy, Qwiklabs etc... I hope you find it useful.

AWS Global Infrastructure:

• AWS locations are called regions. Each region is a geographic area in the globe.
• AWS has 77 availability zones within 24 Geographic Regions around the world. They are planning to add 12 more availability zones and 4 more regions in places such as Bahrain, Cape Town, Jakarta, and Milan
• Northern Virginia, Oregon, Ohio, Northern California, Canada are some of the North American regions. GovCloud regions in North America are only open to government customers.

AWS Global Infrastructure Elements:
AWS Region

• They are independent off of each other.
• They reduce network latency for end user because the platform and assets are in same location where the end user / customer is based.
• AWS Regions allow infrastructure to be near the geographic boundary of the country where the customer resides. This is to meet local government compliance and regulatory requirements.

Availability Zone (AZ)

• Each region consists of two or more AZ.
• This gives AWS redundancy, availability and fault tolerance. In the event, there was a disaster, AWS customer would still be able to carry on their business.
• All AZ have one or more data centers behind its compute power

Data Center

• This where physical servers, storage, switches, loadbalancers, machine learning, iot equipments are fielded
• Any resources provisioned on the cloud gets created here.

Edge Location

• They are used for caching content near users so that they can experience faster execution of programs.
• AWS has more than 205 edge locations and 11 regional edge caches.
• They reduce network latency to access static images, files.

Azure Global Infrastructure:

• Azure can scale globally to reach more users.
• Azure provides reduced time for user to access data. They have very little network latency.
• Azure Cosmos DB is claimed to scale at planet scale.
• Azure provides data compliance services.
• Azure also has redundancy and availability.

Azure Global Infrastructure Elements:
Data Center

• Just like AWS, these are physical buildings with servers, storages, switches, loadbalancers, and various compute equipments
• When we provision cloud resources like Virtual Machines, Azure SQL DB etc, they all get created in these secure buildings.

Availability Zone (AZ)

• Just like AWS AZ's contain 1 or more data centers
• All AZ's are resourced with independent power, cooling, networking equipments
• App and data are distributed across AZ for redundancy.
• When one availability zone goes down, another availability zones takes over to keep operation active.

Region

• Just like AWS, the Region in Azure have two or more AZ.
• AZs within a region are connected with low-latency regional network. This ensures low latency.
• Presently, there are 54+ regions in 140 countries. That's higher than AWS or any cloud provider.

Geographies

• This is a group of regions in a geographic area.
• Many countries require data to be resident for compliance reasons. That's what this service covers.
• Data remains within the country or the geographic border.

Sovereign Region

• Azure offers this service to Governments only. This is equivalent to AWS GovCloud
• Isolated Region that's not open to public usage.
• They built and certified it as per government requirements.

Edge Zones

• Azure's closes equivalent to Edge Locations is Edge Zones
• Azure has introduced Edge Zones in preview in 2020
• There are three types of Azure Edge Zones: 1) Azure Edge Zones, 2) Azure Edge Zones with Carrier, 3) Azure Private Edge Zones

GCP Global Infrastructure:

GCP Global Infrastructure Elements:

Region

• GCP has 22 regions
• GCP regions are independent geographic area where application and data resides.
• They consist of 2 or more zones connected with each other using high speed network. Locations within the same region have network round trip latency under 1 milisecond.

Zones

• 61 Zones in total
• Unlike AWS, and Azure... GCP zones are individual data center.
• Each zone is supplied with independent power, cooling, networking infrastructure.
• Application should deployed across multiple zones in a region to make it fault tolerant.
• Data must be copied across multiple zones as well for redundancy and high availability.

Edge Network Locations

• GCP has over 142 edge locations around the world.
• Just like AWS, they cache data near the user to enable lowest of latency.

GCP Type of Resources:

• Zonal, Regional Reource, & Multi-region Resources

Zonal Resource

• It runs within a single zone.
• When a zone fails, all resources within it fails.
• Example of this resource includes: Google compute engine.

Regional Resource

• Redundantly deployed across multiple zones in a single region.
• Example of this resource is App engine

Multi-region Resources

• GCP manages it
• Deployed against multiple regions for high availability, redundancy, low latency.
• Example of this resource is Google Cloud Storage, Google Cloud KMS, Big Table