Amazon—aren't they an online bookstore? This is what someone in the late 1990s might have asked. Nowadays, with nearly 40% of the e-commerce market in the US and over 30% in India, everyone knows who they are. But not so well-known outside of technology circles is Amazon Web Services (AWS). Amazon describes this part of the business as follows:
AWS is a secure cloud services platform (https://aws.amazon.com/what-is-cloud-computing/), offering compute power, database storage, content delivery, and other functionality to help businesses scale and grow. Millions of customers (https://aws.amazon.com/solutions/case-studies/) are currently leveraging AWS cloud products (https://aws.amazon.com/products/) and solutions (https://aws.amazon.com/solutions/) to build sophisticated applications with increased flexibility, scalability, and reliability (https://aws.amazon.com/what-is-aws/).
Amazon is currently the leader in the cloud infrastructure business with 33% market share—about the same as Google Cloud, Azure, and IBM combined. They're using their huge lead, and the profits from it, to invest in new services to further differentiate themselves. Since the publishing of the first edition of this book, they've added hundreds of new offerings. Millions of customers are taking advantage of these services. The initial target market of startups and small businesses has spawned web-scale enterprises such as Netflix, Pinterest, and Airbnb. The Fortune 500 and governments around the world are also on board.
Note
On an annual basis, all of Amazon's operating income derived from AWS in 2017 (https://www.zdnet.com/article/all-of-amazons-2017-operating-income-comes-from-aws/).
How we got here provides valuable lessons for building your product and scaling it globally. The retail business is peaky. Amazon needed capacity to meet its highest demand period during Thanksgiving and Black Friday. Come January, it had a whole lot of hardware sitting around, underutilized. Having worked for years providing e-commerce services to partners, they were able to take their internal offerings and make them available to the public. Utility computing was born.
This pattern has been repeated many times. Amazon is constantly adding services to tackle new and unique problems that they have encountered within their own business. Those internal services then get previewed and customer feedback is collected. When they meet a certain service level, they're released for use by anyone with a credit card. You need to use this same strategy for your products.
Product-driven development will continue to evolve within the Amazon ecosystem. The benefits that are identified internally will trickle down to all of the AWS cloud users. Newer services will continue to simplify the consumption of complex offerings. Economics of scale will push prices lower while moving resources closer to your customers.
Whether you're developing new features for Amazon or your local football club, or hoping to create the next Netflix, you need to be focused on your product. The patterns in this book have helped people develop and scale their products. Throughout this book, we'll build Minimum Viable Products (MVP) (http://www.startuplessonslearned.com/2009/08/minimum-viable-product-guide.html) on AWS.
Amazon will continue to use this pattern to develop new services, and you should too.
The third part of this book looks at persistence. The item we are persisting is data. AWS offers a number of services to help us manage the volume, velocity, and variety of data. Other products will help identify the veracity and value in the flood. In Chapter 11, Data Processing - Handling Your Data Transformation, we'll use another set of tools to predict how our customers are going to react to new features. Continuous experimentation and measurement allow our product to become more successful.
AWS also provides a number of open source projects as services (https://www.slideshare.net/AmazonWebServices/aws-reinvent-2016-open-source-at-awscontributions-support-and-engagementarc213). This allows them to leverage large active communities for product development and bug hunting. In the case of the AWS Software Development Kit for Python, Boto (https://github.com/boto/boto3), Amazon has taken over the project. We're not going to go into all of the benefits of open source here, but remember that many of the services included in our patterns are available as source code to everyone.
Coding is going to be an important part of your product, the systems that run it, and the people that use it. We'll be writing code to allow the reuse and modification of our cloud patterns. To get into the spirit of things, we'll be using a cloud-based development tool that was originally written by the open source Mozilla foundation. A group of the developers decided to create a freemium model to provide support for people using their Integrated Development Environment (IDE), Cloud9. Amazon bought the company—another great pattern for applying to your product.
The cloud is an extension of the practice of co-locating hardware at a site away from your company. Historically, you would put your stuff closer to a fast connection in order to get your pages to load more quickly. Amazon stretched the market by allowing you to run your product on their hardware. They further innovated by creating a utility model that lets you pay for only what you need.
Obviously, IDE as a service was not the first service offering of AWS, but it fit well with their philosophy of Anything as a Service (often abbreviated as XaaS). By going beyond their initial offering of co-located virtual machines, Amazon has redefined existing infrastructure services. Providing configurable services and software implementations of hardware found in data centers has allowed them to develop complementary products such as load balancing, content delivery networks, and security plus, failover and replication solutions. Extending their internal tools or purchasing synergistic products such as Cloud9 has provided the full stack developer an opportunity to use a well thought out, low cost, pay for use, integrated suite.
AWS provides three primary categories of service offerings:
- Infrastructure as a Service (IaaS)
- Platform as a Service (PaaS)
- Software as a Service (SaaS)
IaaS can be described as a service that provides virtual abstractions for hardware, servers, and networking components. The service provider owns all of the equipment and is responsible for its housing, running, and maintenance. In this case, AWS provides Application Programming Interfaces (APIs), SDKs, and a web console for creating and modifying virtual machines; their network components; routers; gateways; subnets; load balancers; and much more. Where a user with a physical data center would incur charges for the hardware, power, and connectivity, this is removed by IaaS with a payment model that's per-hour (or per-use). Specific IaaS components will be the building blocks for our product. A discussion of the key pieces will take place in Chapter 2, Core Services - Building Blocks for Your Product.
While AWS doesn't provide a full-blown PaaS offering such as Cloud Foundry (https://www.cloudfoundry.org/) at the moment, it has a number of products that can simplify the product development experience. Its broad service catalog can also be used to compose a DIY platform. PaaS is described as the delivery of a computing platform, typically an operating system, programming language execution environment, database, or web server. With AWS, a user can easily turn a code base into a running environment without having to worry about any of the pieces underneath, such as setting up and maintaining the database, web server, or code runtime versions. It also allows your product to be scaled without having to do anything other than set policies through the configuration. We'll survey the services for on-demand expansion in Chapter 3, Availability Patterns - Understanding Your Needs.
The best interpretation for the SaaS model is on-demand software. The user need only configure the software to use and interact with it. The draw to SaaS is that there's no need to learn how to deploy the software to get it working in a larger stack, and generally the charges are per-usage-hour. In the case of AWS, much of the maintenance burden is also removed.
The AWS suite is both impressive and unique in that it doesn't fall under any one of the cloud service models we described previously. Until AWS made its name, the need to virtualize an entire environment or stack was usually not an easy task and required integration of different providers offerings, each solving a specific part of the deployment puzzle. The cost of using many different providers to create a virtual stack might not be cheaper than the initial hardware cost for moving equipment into a data center. Besides the cost of the providers themselves, having multiple providers also created the problem of scaling in one area and notifying another of the changes. While making applications more resilient and scalable, this Frankenstein method usually did not simplify the problem as a whole.
When you build your first MVP for venture capitalists or an internal review board, we recommend you do it fast and focus on the user experience. In most cases, the monolith you create isn't going to be able to achieve web scale. The Cloud-Native Maturity Model is based on the twelve factor design principals. Its focus is on how to run your applications on the cloud: https://pivotal.io/replatforming
AWS supplies all of the services that are needed to launch cloud-native applications. We'll start with a cloud-friendly product, then in Chapter 8, Application Virtualization - Using Cloud Native Patterns for Your Workloads, we'll go into great depth. First, let's look at some of the trends driving cloud adoption.
In most cases, the monolith you create is not going to be able to achieve web scale. As you break off pieces to refactor, you'll develop smaller services. This allows for the decoupling of components upgrade schedules, which speeds up product development. By adopting the continuous delivery pattern espoused by Jez Humble, you can deploy new features and measure their success quickly. We'll take a peek at these architectures in Chapter 5, Continuous Deployment - Introducing New Features with Minimal Risk. Isolation of your microservices will be the subject of Chapter 4, Security - Ensuring the Integrity of Your Systems.
By providing easy to use, low cost SaaS options, AWS lets you focus on building software. High value feature development replaces low level toil. In Chapter 6, Ephemeral Environments - Sandboxes for Experiments, we'll present patterns that help foster collaboration between developers and operators. AWS composable services empower all team members to participate in the product development they know best.
Amazon has introduced thousands of services since the introduction of the Elastic Cloud Compute (EC2), their first publicly available multi-tenant virtualization offering.
Note
Hardware virtualization or platform virtualization refers to the creation of a virtual machine (https://en.wikipedia.org/wiki/Virtual_machine) that acts like a real computer with an operating system.
Most of them are enhancements to their MVP. Many of the services are tucked away in the management console. All of them are constantly monitored to ensure that customers are happy and to identify opportunities for new features. The container and serverless patterns in Chapter 8, Application Virtualization - Using Cloud Native Patterns for Your Workloads, promote self-contained service development for your product. Some, such as the Internet of Things (IoT) services, are out of scope for this book.
If you don't have an AWS account, you can sign up for a free one at https://aws.amazon.com:
You'll need a valid email, phone number, and credit card (in case you go over your allotted capacity):
The exercises in this book will use many of the free AWS services. We'll note when there are pay-per-use services that may be a better solution for certain classes of problems.
AWS learned from the Amazon e-commerce team the value of partnerships. Just like you can shop the Crocs store on the parent side, you can rent an in-memory data grid powered by Apache Geode for a few hours through the AWS marketplace. The AWS marketplace provides an avenue where users can purchase or trial third-party images that provide configurable services such as databases and web applications. As a product developer, you can easily add your offerings to the exchange:
In the spirit of competition, a lot of the software in the marketplace offers competition to the AWS native services. Operations engineers may find it easier to move workloads to the cloud by reusing components that they're familiar with, such as F5 load balancers, instead of the AWS Elastic Load Balancers (Chapter 2, Core Services - Building Blocks for Your Product). Other products, such as the trend Micro Deep Security Agent, fill in gaps where AWS doesn't currently have a solution. Partnerships with Microsoft, VMware, and Tableau let you easily migrate internal workloads or extend your data center to the cloud.
The adoption of a DevOps culture for your product team provides you with similar cooperation benefits. Getting operations engineers involved in product development helps eliminate hand-offs. Developers being on call for production issues acquaints them with troubleshooting practices and boosts empathy for users. By working together, the product team can establish a clear contact for use. Well-defined APIs are an excellent way to ensure that the contract can be met. AWS publishes extensive documentation for its interfaces, as well as defining service-level objectives for its products.
For a new product, you should start development in the cloud to take advantage of Amazon's well-thought-out offerings. For existing companies, the AWS suite can handle almost everything a modern data center can while significantly reducing your operating costs. The shift suits small teams but, for mid-sized teams, the effort that's saved begins to outweigh the cost. In larger organizations, AWS can also help you get rid of some of the process bottlenecks that come with a data center. Requesting, provisioning, repairs, and scheduling downtime are all streamlined. There are many different answers to why moving to the cloud might be beneficial. We'll start at mid-sized because this is the size that usually includes the two teams that benefit the most. Let's take a look at our personas:
- The software engineer
- The operations engineer
- The product owner
For a software developer, the ability to share code and delivery tools across teams is critical. AWS gives you the ability to enforce consistency across environments. Different versions of core languages or libraries can be avoided by having uniform templates. Chapter 5, Continuous Deployment - Introducing New Features with Minimal Risk, emphasizes how these good continuous delivery practices in the cloud can lower business risk. Chapter 3, Availability Patterns - Understanding Your Needs, goes deeper on making sure your product is available, while Chapter 4, Security - Ensuring the Integrity of Your Systems, covers how a secure groundwork can be inherited across products and services. We'll lay out how developers can easily build cost-effective products, which can grow as demand intensifies. Data management patterns in Chapter 10, Databases - Identifying which Type Fits Your Needs, and Chapter 11, Data Processing - Handling Your Data Transformation, will help software engineers understand the trade-offs in architecting their MVPs.
The ability to spin up entire environments for testing and simulation in AWS is amazing. Providing developers with their own sandboxes improves their ability to innovate and deliver new features at a high velocity. In the cloud-native world, a single environment can be encapsulated on a single EC2 instance. We'll explore how application virtualization can further improve developer productivity in Chapter 8, Application Virtualization - Using Cloud Native Patterns for Your Workloads.
Throwaway environments are also be beneficial to the operations teams. In Chapter 6, Ephemeral Environments - Sandboxes for Experiments, we'll show you how to take centralized configuration, which describes AWS resources, map them into a dependency graph, and create an entire stack. This flexibility would be nearly impossible in traditional hardware settings and provides an on-demand mentality—not just for the base application, but also for the entire infrastructure, leading to a more agile core.
The operations team's responsibility differs greatly from company to company, but it's safe to assume that the team is heavily involved with monitoring the applications and systems for issues and possible optimizations. AWS provides enough infrastructure for monitoring and acting on metrics, and an entire book could be dedicated to the topic. For an operations team, the benefit of moving to Amazon might justify itself only to alleviate all of the work involved in duplicating this functionality elsewhere, allowing the team to focus on creating deeper and more meaningful system health checks. There will be more coverage of operations in Chapter 7, Operation and Maintenance - Keeping Things Running at Peak Performance, while observability will be addressed in Chapter 12, Observability - Understanding How Your Products Are Behaving.
As a product owner, the low cost of entry into very complex technology is a huge win. Chapter 3, Availability Patterns - Understanding Your Needs, will analyze the cost benefits of using AWS to provide global availability. The ability to move from hourly pricing to longer term subscriptions will give you a good baseline for product operating expenses. Avoiding inferior and counterproductive solutions will be the topic of Chapter 9, Antipatterns - Avoiding Counterproductive Solutions, and Chapter 13, Anti-Patterns - Bypassing Inferior Options. The shared responsibility model AWS provides will be explored in Chapter 4, Security - Ensuring the Integrity of Your Systems. By inheriting industry standard security controls, you can concentrate on user-centric product aspects.
Letting AWS manage the services at scale lets you avoid investing in dedicated support staff. Amazon has also adopted the freemium pattern for their own support. Start small and expand if you need more help:
