Introduction to Cloud Computing and the AWS Global Infrastructure

As discussed previously, computing refers to on-demand access and delivery of IT services, which customers can consume over the standard public internet or some form of wide-area network. These services will include compute, network, storage, databases, and Software as a Service (SaaS) products. Cloud computing has enabled businesses to design and deploy applications without requiring expensive hardware upfront. Instead, they lease/rent required IT infrastructure from such third-party providers.

Of the various providers of cloud computing services, AWS is the largest provider, offering a variety of cloud IT services. These services fall into various categories: Infrastructure as a Service (IaaS), Platform as a Service (PaaS), and Software as a Service (SaaS) solutions. Businesses can consume those services rather than create their own dedicated environments to host applications. With established infrastructure across multiple countries and massive economies of scale, AWS can offer infrastructure and software service at a fraction of the cost, with redundancy, scalability, high availability, and security.

In the following section, you will learn about the AWS Global Infrastructure, which enables customers across the globe to access AWS services. Furthermore, customers in one location can consume cloud services worldwide, enabling access to a global customer base, and fulfilling any compliance or regulatory needs.

AWS Global Infrastructure

The AWS Global Infrastructure is a vast collection of data center facilities across multiple countries globally. The geographical locations where AWS hosts its data center services comprising compute, storage, and network, along with its vast array of cloud services, are known as AWS Regions. You will find the map at https://aws.amazon.com/about-aws/global-infrastructure/.

Within each AWS region are small groups of data centers that are logically and physically separated by a distance that falls within 100 kilometers (km) (60 miles) of each other. These logically and physically separated groups of data centers form what we call Availability Zones (AZs). Furthermore, AWS designs its regions with multiple AZs per region. Most AWS regions have at least three AZs, and some have even more. For example, the North Virginia region (us-east-1) has six AZs.

In the next section, you’ll look at the core components of the AWS Global Infrastructure in more detail.

Regions

As explained earlier, AWS regions are physical locations across the globe where AWS hosts its infrastructure facilities. These comprise data centers designed to enable customers to access a vast collection of infrastructure services with which they can deploy cloud resources, such as compute, network, storage, and database services. Customers can connect to a given region anywhere across the Global Infrastructure.

Choosing a region to provision cloud resources ultimately depends on the use case of the business. Often, this will be based on multiple factors, including the following:

• The requirement to host infrastructure resources closer to your end users, where you can host your applications with reduced network latency.
• The requirement to host infrastructure within political and national borders to adhere to strict data sovereignty and compliance regulations.
• The requirement to isolate groups of resources from each other to facilitate disaster recovery and business continuity use cases.

Note

In the case of our fictitious client, TodoPlus, the initial offering for the application will be based in the US to fulfill compliance and regulatory requirements for storing data within the US borders. Should the product be successful, TodoPlus would be looking to expand into Europe and Asia Pacific once all necessary laws and regulatory requirements have been analyzed, and measures are taken to adhere to them.

AZs

Within each AWS region, you will find multiple AZs, which are metropolitan areas housing one or more data center facilities in each region. Each AZ will host hardware components such as servers, storage, and network equipment, all fitted with redundant power, connectivity, cooling, and security controls.

The primary purpose of having multiple AZs in each region is to enable customers to host their applications and workloads in a manner that offers high availability, fault tolerance, and scalability. With multiple AZs, you can host copies or replica application resources across these AZs, which ultimately means that you can continue to serve your customers even if there is an outage of one AZ in the given region.

This is all possible because, although each AZ operates independently, they are still connected over high-speed, high-bandwidth, low network latency, and fully redundant, dedicated metro fiber connectivity.

Note

Concerning our company, TodoPlus, their initial choice of region to host their application will be set to the US-East-1 (North Virginia) region. This region is selected because the company will start promoting its new productivity application in local markets. However, they plan to make the application available to customers across the US and later globally.

Edge Locations and Regional Edge Caches

The AWS Global Infrastructure also comprises edge locations and Regional Edge Caches. Edge locations or points of presence (POPs) offer massive amounts of storage, high-bandwidth networking equipment, and edge computing services that enable data to be accessed, processed, and analyzed closer to the end customers’ physical location.

These edge locations are connected to AWS regions through the AWS backbone network. This comprises fully redundant, multiple 100-Gigabit Ethernet (GbE) parallel fiber connections that substantially improve throughput and offer low-latency connectivity. You can review the current list and types of edge locations (POPs) at https://aws.amazon.com/cloudfront/features/.

Edge locations are different from standard regions and AZs. You cannot connect directly to a given edge location to set up resources. Instead, you consume certain AWS services that use these edge locations’ storage, caching, and high network connectivity. One service that uses these edge locations is Amazon CloudFront.

Regional Edge Caches are like edge locations. However, they are strategically placed and have a larger storage capacity to hold cache data longer than individual edge locations. Individual edge locations have a shorter time-to-live than Regional Edge Caches, ensuring that stale data isn’t hosted too long. If the same cache data is later accessed, an attempt is made to see whether it is still available at regional edge caches before sending a request to the origin.

Amazon CloudFront is a content delivery network (CDN) service that enables you to efficiently distribute content to end users in a manner that reduces overall latency. With CloudFront, regularly accessed content is cached in the edge location and in regional edge caches, which offer the lowest latency to end users who attempt to access your content. This means those users do not have to fetch frequently accessed content from the origin if it resides in the cache.

Note

In the case of our fictitious company, TodoPlus Limited, the application will be hosted in the N.Virginia (us-east-1) region. CloudFront can be used to cache static content such as images, videos, and user guides to help reduce the latency for end users as they access the application from various parts of the US and, ultimately, from across continents when our client expands its offering globally.

Edge locations can allow customers to upload data to AWS storage services such as Amazon S3 over the AWS backbone network, offering low latency and high-bandwidth throughput using a service known as S3 Transfer Acceleration (S3TA).

The AWS Global Infrastructure also comprises other infrastructure services, including the following:

• Local zones are special zones designed to bring compute, storage, database, and other select AWS services closer to end customers’ physical locations. This is particularly useful if you require very low latency access to cloud services. Regarding TodoPlus Limited, this will not be necessary as its customer base is geographically dispersed across the US and potentially across the globe.
• Wavelength zones are zones where AWS has deployed infrastructure services such as compute and storage services within 5G network providers to help optimize mobile edge computing applications.
• Direct Connect locations are designed to establish high bandwidth network connections between clients’ data center facilities and the AWS cloud. TodoPlus Limited may wish to set up a Direct Connection to the AWS cloud. This will improve data transfer speeds between on-premises applications and the cloud due to the higher bandwidth capability.
• Outposts enable true hybrid cloud computing design by extending AWS infrastructure services, APIs, and tools to customers’ on-premises locations. If TodoPlus Limited plans to continue with a hybrid cloud model, an Outpost configuration will enable them to access AWS services locally, allowing low-latency access to certain applications that cannot be hosted in the cloud. If they have applications that need to follow strict compliance or regulatory requirements, then, again, an Outpost setup will help achieve this as all the data will be held locally on-premises.

This section examined the AWS Global Infrastructure and identified some of its core components. Understanding how the Global Infrastructure is architected will enable you to design applications for high availability, scalability, security, and cost-effectiveness.

In the next section, we will look at how you can access the vast array of AWS services via an AWS account as a customer.