You might have wondered, where is "the cloud" located anyway? In this topic, we explore where the AWS cloud platform is located. We will see that the AWS global infrastructure is the basis for its high availability, reliability, and low latency. We will introduce the key components of this global infrastructure which includes regions, availability zones, edge locations, and more. Finally, we discuss its benefits and drawbacks.
Regions and availability zones
First, let's talk about regions. Regions are physical locations around the world where AWS has set up its data centers. Every region is geographically separated from others by great distances. This ensures that if one region fails, you can still access services from other regions.
Each region has clusters of data centers known as availability zones (AZs). Every AZ within a region operates independently in terms of power, cooling, networking, and other physical aspects. This independence means that if one AZ fails, the others will not be affected. Generally, an AWS region contains at least three AZs.
On AWS, you begin by creating virtual networks within availability zones in a specific region. You then define subnets within these virtual networks and launch AWS services, as shown in the diagram above.
A region's name reflects its geographical location. For example, the first AWS region in South Africa is named af-south-1; and the second region in the U.S. West area is named us-west-2. For AZs, their names correspond to the region they are placed in plus a letter of the alphabet (a, b, c, d...). For example, the first AZ in the South African region is named af-south-1a. Later, we will launch AWS services in various regions and AZs.
When choosing regions and availability zones, keep these points in mind:
- Not every service is available in all regions;
- Pricing for AWS services varies by region;
- The closer the region is to your users, the lower the latency;
- Make sure you meet any compliance requirements for running workloads in a selected region;
- While most services are region-specific, some, like AWS IAM, are global;
As of October 2023, AWS has 32 regions and 102 availability zones:
Regions can effectively communicate with each other through fast, redundant, and low-latency network links. You can design applications to automatically replicate across regions or AZs, ensuring their resiliency and continued availability during failures or scheduled/unscheduled maintenance.
Local and wavelength zones
The closer you are to where an application is hosted, the faster the access time. AWS local zones are extensions of AWS regions designed to bring applications as close to users as possible thus reducing latency. The image below shows an AWS Local Zone in Nigeria with South Africa as its parent region. Users in Nigeria can enjoy fast response times compared to accessing the services from the AZ in South Africa.
Now let's go over wavelength zones. These zones provide access to AWS services over 5G networks. Imagine not having to wait even a few milliseconds for your favorite show on Netflix to load. Sounds impressive, right? AWS Wavelength Zones offer less than 10ms latency for data access on 5G-enabled devices like mobile phones, IoT devices, or smart vehicles. This is particularly useful for video streaming, gaming, and virtual reality applications. This greatly improves user experience.
In the diagram above, the mobile devices are connected to AWS media services over a 5G network, and the users can stream content with ultra-low latency for a seamless viewing experience.
AWS Outposts
What if you love AWS but need to run its services in your organization's data centers? This could be for compliance reasons, low latency access, or perhaps you want to make your cloud migration smoother. With AWS Outposts, you can use the same services, APIs, and tools used in the AWS cloud on your on-site servers. However, this requires AWS-approved hardware.
Consider the architecture diagram below:
In this case, we have an AWS database service running on an AWS Outposts rack connected to the AWS cloud. This database might be used to store sensitive data that needs to remain in-house for compliance reasons. With AWS Outposts, we can set up a hybrid cloud deployment model.
The edge
The AWS global infrastructure also includes regional edge caches and edge locations. These hold a copy or cache of content close to users for quick retrieval. The caches are used by Amazon CloudFront, a content delivery network (CDN) service that accelerates content delivery. With this CDN, websites can store assets like photos and videos in edge locations for fast access when required.
Let's look at a simple schematic representation of this:
In this setup, the static website housed in the Amazon S3 storage bucket is cached in edge locations closer to users. When a user sends a request, the Amazon Route 53 DNS service directs the request to the nearest edge location for that user. Therefore, users can access content faster compared to retrieving it from a more distant availability zone.
Pros and cons of AWS global infrastructure
The AWS global infrastructure comes with several benefits. Here they are:
- High availability of resources.
- Lower latency when accessing AWS services.
- Flexibility in where to run applications.
- The AWS infrastructure is designed to be very secure.
- Companies can scale as needed due to AWS's large number of data centers.
- High performance due to AWS's high bandwidth, fully redundant global network.
However, there are also some drawbacks:
- Deploying applications to multiple regions, AZs, and zones can be expensive.
- Deploying and managing applications in multiple regions and zones can be complex.
Conclusion
We have just explored the AWS global infrastructure. Now you know where the cloud is! We learned that AWS regions are geographically distributed locations across the globe that contain clusters of data centers. These data centers host cloud services. However, data centers can be too far from our end users, resulting in high latency. To solve this issue, local and wavelength zones, as well as edge locations and caches, can be employed to bring content closer to users. We can even run cloud services on-site with AWS Outposts!
With data centers located worldwide, AWS can guarantee high availability and a global footprint for its users. Developers can harness the AWS cloud to create resilient, highly responsive, and globally distributed applications. This enhances user experience and minimizes downtime. Unfortunately, achieving the needed level of redundancy and speed can be costly, but for sensitive workloads, it might be worth it.