Build modern

applications on AWS
Manage less. Build faster. Innovate more.
Modern applications are
changing how you deliver
customer value
Organizations worldwide are focusing on innovation, customer
satisfaction, and operational efficiency as today’s top business
priorities. To achieve these goals, decision makers understand the
need for their applications to play a key role in leading the way.
In fact, 64 percent of US enterprise decision makers cite modernization
as their top IT/operations priority for the next 12 months.¹ However,
many companies continue to struggle with building and running
their applications—with 78 percent of organizations citing a lack of
technology management expertise as a top challenge.²

For many businesses, a move to the cloud has helped accelerate

the pace of digital innovation by providing on-demand capabilities,
shifting the burden of server management to the cloud provider,
and reducing the total cost of ownership (TCO). But moving to the
cloud is just the first step toward modernization. As business and
technical leaders consider an increasingly competitive market, many
are building modern applications on Amazon Web Services (AWS) to
leverage the full value of the cloud.

In this eBook, we will guide you through three pathways that will
help lay the foundation for modern application development in your
own organization. We will also explore how modern application
development with AWS can help your organization innovate, reduce
costs, accelerate time to market, and improve reliability.

¹ Sustar L., et al., “Aiming For Innovation: Cloud In The US Today,” June 2022
2
² “2023 State of the Cloud Report,” Flexera, 2023
Modern applications empower “Taking advantage of new MACH-ready
architectures — microservices, API, cloud,
digital innovators and headless — allows organizations
to quickly plug in new capabilities that
Innovation means listening to your customers fit their technology stack and address
In a recent IDC Market Perspective, the importance of retaining a opportunities of growth as a digital-first
customer-centric viewpoint was referenced as a critical element in customer-centric enterprise.”3
driving successful digital innovation. A digital innovator does not present
a solution in search of a problem. Rather, innovation is driven by viewing Marci Maddox, Research VP, Digital Experience Strategies, IDC
the customer journey through the customer’s lens and building with the
goal of removing points of friction in their experience.

Focusing on your customer means making business decisions by

working backwards from the customer’s point of view. It means
Listen
constantly evolving products and services to better deliver the
outcomes that delight customers. It also means listening to what your
customers truly care about so you can anticipate the experience that
will keep them engaged as you continue inventing and iterating on
their behalf. This approach is called the “innovation flywheel.”

The basic idea of the innovation flywheel is that meaningful innovation Innovation
is created out of customer demand, improves with customer feedback, Flywheel
and constantly repeats (and profitably) until the demand changes and
the cycle begins again. Modern application development technologies
and techniques help unlock your ability to iterate quickly and scale Iterate Experiment
continuously. The deeper your teams adopt modern application
development, the faster you can get your innovation flywheel
spinning—and the sooner you can leverage innovation to differentiate
your business.

³ Maddox, M., et al., “Reframing the Digital-First Customer Journey: Key Takeaways
3
from IDC’s Global CIO Advisory Board – April 2022 Edition,” IDC, May 2022
40%
Building modern applications on AWS can help you get to market faster.
By accelerating the build cycle and offloading operational overhead,
your developers can quickly build new features. You will increase
of firms will be adopting innovation with a modular architecture that frees teams to experiment
a cloud-native-first independently with individual application components without risking
strategy in 2023⁴ the entire application. By automating test procedures and monitoring
at every stage of the development lifecycle, you will improve reliability.
And you will improve TCO with a pay-for-value pricing model that

45%
reduces the cost of over-provisioning or paying for idle resources.

While digital transformation and the shift to modern application

development can feel complex at an organizational level, the process
of YoY growth in cloud does not need to be overly complicated. Many organizations, inspired by
and service adoption⁵ the promise of the cloud, move all their new application development
to the cloud in a single step. Plenty of others take a hybrid approach,
which involves building applications to run on premises and in the

82%
cloud. These two paths don’t have to be mutually exclusive. In fact,
the process typically involves parallel paths—the first, a team-by-team
experience; the second, a workload-by-workload journey—both moving
of organizations are opportunistically, a step at a time.
struggling to manage
their cloud spend⁶

⁴ Sustar, L., et al., “Predictions 2023: Cloud Computing,” Forrester, October 2022
⁵ “Closing the cloud strategy, technology, and innovation gap: Deloitte US Future of Cloud
Survey Report,” Deloitte, 2022
⁶ “2023 State of the Cloud Report,” Flexera, 2023 4
3 modernization pathways that
generate real business value
Through our experience building applications for Amazon.com—and for millions of
AWS customers—we’ve observed three pathways customers often take to translate
their vision of application modernization into a reality, generating value for their
business in the process:

1. Run managed Kubernetes on AWS and extend existing Kubernetes to AWS for
hybrid distribution: Organizations running Kubernetes in the cloud or on premises
can run these workloads on AWS to increase reliability, security, and scalability.
This approach provides server-level control, common operational tooling across
environments, and integrated access to the vast AWS services portfolio.

2. Build new apps with a serverless operational model: As organizations build

new applications or features, a serverless operational model provides the easiest
approach for scaling and managing cloud applications. The serverless model
shifts as much management of the underlying compute resources to AWS as
possible, so development teams can focus on writing code that solves business
problems.

3. Transition to a modern DevOps model: To empower the cultural shift that

produces modern applications at scale, organizations can leverage DevOps
services and tools. Ensure security and governance while rapidly building and
delivering new products.

We will explore each pathway in more detail, demonstrating how each can help
lead to increased agility, lower costs, and building better apps that support business
success. While you can modernize applications from any starting point, the outcome
needs to be the same: applications that are secure, reliable, scalable, and quickly
available for your customers and partners from the onset of your combined efforts.

5
Modern applications in practice
The 3 pathways of modern application development
Modern application development is a powerful approach to designing, building,
and managing software in the cloud. In practice, it increases the agility of your
development teams and the reliability and security of your applications, allowing your
teams to build better products faster.

We’ve worked with a wide range of organizations, from large, experienced operations
teams looking to maintain maximum control with portable orchestration tools to
resource-constrained teams preferring to offload operational tasks as they shift their
focus to building and iterating on applications.

From these experiences, we’ve identified three solution pillars of modern application
development to help you on your journey toward modernization:

1 Run managed Kubernetes and extend existing on-premises investments

2 Build with a serverless operational model

3 The Modern DevOps model

6
1 Run managed Kubernetes and extend existing on-premises investments

Today, more and more customers are taking a path of reinvention, building and security-conscious government agencies and financial institutions trust
new applications that enable their organizations to take full advantage of the it to run their most important production workloads. According to the Cloud
cloud. In fact, there’s no single way to modernize; applications can coexist in Native Computing Foundation (CNCF), more containers run on Amazon EKS
different states (some highlighted in this eBook). Organizations worldwide than any other cloud. Amazon EKS runs native Kubernetes and manages cluster
are continuing to adopt and grow their use of containers as a lightweight and lifecycle and other key tasks. And it includes integrations with AWS services
portable way to run and deploy many new applications. For many, containers that enhance management and security. For example, Amazon GuardDuty
have become the preferred route to support complex application architectures threat detection continuously monitors and profiles container runtime activity
that require flexibility, rapid deployment, and continuous innovation in an to identify suspicious behaviors. By choosing Amazon EKS, organizations can
environment that can run almost anywhere. eliminate undifferentiated heavy lifting and focus on delivering value to their
end users. Additionally, Amazon EKS is customizable to support any needs. For
When moving to containerized applications, organizations need to consider example, it helps customers take full advantage of the broadest selection of
where and how to invest their operational resources. Many development teams compute choices and instance types, such as AWS Graviton-based instances,
have strong opinions about how their applications run. Seeking to standardize which enable up to 40 percent better price performance.
the deployment, scale, security, and management of their containerized
applications, these dev teams have turned to Kubernetes. Due to its vibrant For on-premises or hybrid deployments, Amazon EKS Anywhere brings the
open-source ecosystem, community, consistent APIs, and broad flexibility, Amazon EKS experience to your data center, allowing you to retain latency-
Kubernetes has gained significant adoption. sensitive or regulated applications in a private cloud on premises all while
leveraging a common toolset across environments.
Organizations planning to or already running Kubernetes—or thinking about
building new applications on Kubernetes—can run them on Amazon Elastic
Kubernetes Service (EKS). Amazon EKS takes full advantage of AWS’s
operational excellence to provide organizations with a highly scalable,
reliable, and secure Kubernetes platform in the cloud. It is the best way to run
Kubernetes, which is why the fastest-growing startups, largest enterprises,

7
Why use Kubernetes today
Many successful modern application developers have already made the transition to Kubernetes. These developers like Kubernetes because
the technology makes it easier to automate and standardize deployment, scaling, security, management, and observability of containerized
applications. This helps them reduce development time, increase operational efficiency, and improve their overall reliability and security.
Kubernetes enables customers to leverage the same APIs and skill sets for running infrastructure across environments across their
organizations or in the cloud and on premises.

Run applications Seamlessly move Run your apps Add new

at scale applications anywhere functionality
Kubernetes lets you define Containerized applications Run highly available and Kubernetes is an open-source
complex containerized can be seamlessly moved scalable Kubernetes clusters project, so adding new
applications and run them from local development on AWS while maintaining functionality is easy. A large
at scale across a cluster machines to production full compatibility with your community of developers
of servers. deployments in the cloud Kubernetes deployments and companies build
using the same operational running on premises. extensions, integrations, and
tooling. plugins that help Kubernetes
users do more.

8
Organizations possessing in-house Kubernetes and operational
expertise can run Kubernetes on AWS using Amazon EKS to leverage
scalable, secure, and highly available infrastructure. Amazon
EKS automatically manages the availability and scalability of the
Kubernetes control plane responsible for scheduling containers,
managing application availability, storing cluster data, and
automating other key tasks. For organizations that have chosen
to use Kubernetes, allowing AWS to operate and maintain the
Kubernetes cluster has made it easier to achieve their goals.

For organizations that want to use Amazon EKS without the

complexity associated with running and scaling the underlying
infrastructure, AWS Fargate is a serverless compute environment for
containers that removes the requirement for you to own, run, and
manage the lifecycle of a compute infrastructure yourself. With
Fargate, you can:

• Eliminate operational overhead of scaling, patching, securing, and

managing servers
• Improve security through workload isolation by design (Amazon
EKS pods run in their own dedicated runtime environment)
• Pay only for what you use and automatically scale to meet
business demands by launching thousands of containers within
minutes across worldwide AWS Availability Zones

9
C U ST O M E R S N A P S H O T

New Relic
New Relic is a leader in the observability space, providing engineers “Consuming AWS lets us focus on our core
with a software-as-a-service (SaaS) platform that analyzes, competencies, enabling us to release better
troubleshoots, and optimizes their entire software stack. In 2020,
products faster and more frequently.”
New Relic began a journey to transform its business, moving from a
host-based pricing model to a consumption-based pricing model. In Andrew Harnett, VP of Engineering, New Relic
addition, the company introduced a free service tier to all customers
and prepared for a significant increase in demand on its platform.
New Relic made the decision to migrate its entire platform to AWS to
accommodate the projected trajectory of growth. In a period of just
eight months, the business leveraged its own tools and processes and
AWS capabilities and offerings to migrate over 20,000 servers and to
refactor its entire services platform. The refactoring of the platform
utilized Amazon EKS. In this migration and refactoring program,
New Relic saw material improvements in engineering efficiency and
platform resiliency, along with a path for long-term scalability to
support its aggressive growth targets.

Read more ›

10
C U ST O M E R S N A P S H O T

Volkswagen Group
Global automotive manufacturer Volkswagen Group is building “It’s simple and effective for us to use AWS
augmented reality and virtual reality (AR/VR) applications to save because there are so many native services that
time, reduce costs, and improve performance for use cases across the
support our goals and architecture drafts.”
product lifecycle—from design reviews and training simulations to
remote guidance in the factory. To scale AR/VR, Volkswagen needed Jan-Paul Brückmann, Product Owner, Volkswagen Digital
a faster, simpler, and more efficient way to prepare and deliver 3D Realities Hub
content. The company developed a cloud architecture on which it
could accomplish two things:

1. Automate Volkswagen’s 3D data preparation pipelines for higher

optimization of 3D models

2. Test the ability to remotely render and stream 3D graphics to

AR/VR headsets, replacing headsets tether-constrained to
expensive workstations

To achieve its goals, Volkswagen is using AWS to migrate specific VR

rendering and 3D data optimization workloads to the cloud. It’s also
using Innoactive Portal, a custom application developed alongside VR
software company Innoactive to deploy and manage VR applications.
Leveraging several AWS services, including Amazon Elastic Container
Service (Amazon ECS)—a fully managed container orchestration
service that automates the deployment, management, and scaling
of containerized applications—Volkswagen is making its 3D data
preparation pipeline more efficient and increasing its rendering
power and speed in the cloud.

See the full story ›

11
2 Build with a serverless Key considerations for building
scalable modern apps
operational model
Architectural patterns: microservices
When running applications in the cloud, some teams and organizations
Monolithic apps can be easy to manage at first, but as your app
may choose to manage the cloud infrastructure themselves. However,
begins to change and evolve, this growth can create challenges
this means they need to invest resources in cloud administration
for scalability and agility, including how to distribute ownership
tasks to ensure that their infrastructure is provisioned, maintained,
of the app across your teams. You can build a strong culture
upgraded, patched, load balanced, secure, available, and scaled
of ownership but still struggle to scale up if your application
based on application needs and customer demands during the build
architecture includes hard dependencies that prevent teams
and run phases of their application. For teams looking for faster
from taking ownership of the final product. This is why we
ways to turn ideas into modern production applications and reduce
often recommend building microservices architectures for
ongoing infrastructure maintenance in production, these tasks are
applications that grow and change rapidly. Microservices are
operational overhead and slow their pace of innovation. They need to
the architectural expression of a culture of ownership—they
build applications rapidly with the ability to scale quickly, have global
neatly divide complex applications into components that a
availability, manage large amounts of data, and respond in milliseconds,
single team can own and run independently.
all while lowering total costs and maintaining robust security postures.
With a monolith, you have many developers all pushing changes
A serverless operational model alleviates this challenge and provides the
through a shared release pipeline, which causes friction at many
easiest model for scaling and managing cloud applications by shifting
points of the lifecycle. Upfront, during development, engineers
as much management of the underlying compute resources to the
need to coordinate their changes to make sure they’re not
cloud provider as possible so development teams can focus on writing
breaking someone else’s code. To upgrade a shared library to
code that solves business problems. With AWS serverless solutions,
take advantage of a new feature, you need to convince everyone
you leverage AWS best practices and expertise on performance,
else to upgrade at the same time—a tough ask! And if you want
scalability, availability, and security as AWS takes on tasks such as server
to quickly push an important fix for your feature, you still need
management, resource allocation, and scaling on your behalf. With
to merge it with changes in progress.
those tasks taken care of, development teams can accelerate time to
production while lowering TCOs by paying for only what is used and After development, you also face overhead when you’re
focusing valuable human resources on innovation. pushing the changes through the delivery pipeline. Even when
making a one-line change in a tiny piece of code, engineers
need to coordinate their changes ahead of time, merge their
code, resolve conflicts within releases, rebuild the entire app,
run all of the test suites, and redeploy once again.

12
Monolith apps Do everything the app requires in a single Agility, scalability, and resilience
self-contained unit
With a microservices architecture, applications are
Single app composed of independent components that run each
application process as a service. Services are built for
Must deploy entire app
business capabilities, and each service performs a
One database focused function. Because services run independently,
each can be updated, deployed, and scaled to meet
Organized around technology layers the demand for specific functions of an application.
For example, an online shopping cart built as a
State in each runtime instance
microservice can be used by many more users during
One technology stack for entire app a sale as it independently scales to support demand.
Microservices communicate data with each other
via well-defined interfaces using lightweight APIs,
events, or streams. Our customers are increasingly
relying on event-driven architectures (EDAs)—those in
Decompose the app into smaller, loosely coupled which actions are triggered in response to changes in
Microservices
components, each responsible for a specific capability data—to improve scalability and resiliency while also
reducing costs.
Each component does one thing

Minimal function services

Deployed separately, interact together

Each has its own database

Organized around business capabilities

State is externalized

Choice of technology for each microservice

13
C U ST O M E R S N A P S H O T

Bosch Thermotechnology
North America “We use AWS to achieve our business goals
and to innovate in the technology space.
Bosch Thermotechnology North America (Bosch TTNA) built a
smart source of heating, ventilating, and air-conditioning (HVAC)
Using AWS, we accelerate the change that
systems by modernizing and migrating its business to AWS. The we’re driving.”
new solution monitors products remotely while removing the
Pablo Ferreyra, Head of Software Development for Bosch
undifferentiated heavy lifting of infrastructure management.
CI Americas, Bosch Thermotechnology North America
Bosch TTNA was new to smart device development and wanted
a cost-effective solution to expand its infrastructure capacity and
scalability while creating new smart technologies. The company
used AWS to build solutions to connect its devices to AWS Internet
of Things (AWS IoT). The solution uses AWS serverless technologies
for data processing, application integration, and the scaling
required to manage its business. Bosch TTNA can now improve
customer service by remotely monitoring its new smart energy and
building devices with minimal operational overhead.

See the full story ›

14
C U ST O M E R S N A P S H O T

Liberty Mutual
With $40 billion in annual revenue, Liberty Mutual is the world’s “Our collaboration to figure things out feels like
sixth-largest property and casualty insurance company and is an more than a customer-vendor relationship. It
industry front-runner in technological innovation. To achieve its
genuinely feels like AWS is part of our team.”
goal of becoming a global digital company, the insurance provider
focused on three main areas of transformation: customer centricity, Dave Anderson, Former Director of Technology, Liberty Mutual
agility, and cloud-native development. To that end, the company
made a strategic decision to pursue a serverless-first approach—a
move designed to give it an edge in its competitive global market.
Liberty Mutual used AWS to migrate its on-premises systems to
the cloud to modernize and drive business-wide transformation. By
using serverless architecture and letting AWS handle infrastructure
management tasks like capacity provisioning and patching, Liberty
Mutual reduced its operational burden and realized substantial cost
savings. The company also used serverless solutions so that it could
rapidly build more agile, high-quality applications. By eliminating
the operational overhead, the serverless architecture facilitates
experimentation, empowering Liberty Mutual’s teams to release
quickly, get feedback, and iterate to get to market faster.

Read more ›

15
As serverless as possible
As your architectural patterns and software delivery processes change, you will 1. No server management – There is no need to provision, operate and
probably want to adopt an operational model that enables you to offload any patch, or maintain any servers. There is no software or runtime to install,
activity that isn’t a core competency of your business. To gain agility for rapid maintain, or administer.
innovation, we recommend building microservices architecture and operating
and deploying software using automation for things like monitoring, 2. Flexible scaling – Your application can be scaled automatically or
provisioning, cost management, deployment, security, and governance of by adjusting its capacity through toggling the units of consumption
apps. Choosing a serverless operational strategy—opting for serverless (throughput, memory) rather than units of individual servers.
technologies wherever possible—enables you to maximize the operational
3. Pay for value – Instead of paying for server units, pay for what you value—
benefits of AWS. Whether you are building net-new apps or migrating legacy
consistent throughput or execution duration.
software, choosing to build with serverless primitives for compute, data,
analytics, and integration will enable you to benefit from the most agility the 4. Automated high availability – Serverless provides built-in availability and
cloud has to offer. fault tolerance. You don’t need to architect for these capabilities because
the services running the application provide them by default.
How do we define serverless at AWS?
5. Native integration – Built-in service integrations with AWS solutions, third-
When we say serverless, we mean the removal of the undifferentiated heavy party SaaS solutions, and the ability to connect with first-party custom-
lifting that is server operations. This is an important distinction because it built applications make it easier to access and use the resources necessary
allows you to focus on building the application rather than managing and to build your applications.
scaling the infrastructure to support the application. The tenets of a serverless
operational model are: 6. Secure environment – Serverless offers built-in security best practices
that natively integrate with AWS’s identity and governance tools that you
already trust with compliance to SOC, HIPAA, PCI, and ISO.

A serverless operational model is ideal for high-growth companies that

need to innovate quickly. Serverless enables your teams to move even
faster while keeping a laser focus on the activities that differentiate your
business—all so you can accelerate your innovation flywheel.

16
Build with serverless containers
Leveraging AWS Lambda and Amazon
ECS with Fargate

With the rise of containers and serverless computing, AWS continues

to provide the broadest and deepest offerings for compute to ensure
you have the right compute option to fit your needs. Choosing the
optimal compute for your modern application starts with exploring
several questions. Does self-managing infrastructure improve your
business results? Do you have the expertise to do it? And will the
extra effort ultimately drive value?

Increasingly, customers are choosing to offload server management by

adopting serverless container services like Amazon ECS on Fargate or
just running code with serverless function services like AWS Lambda.

Most customers use a combination of services. About 80 percent of

AWS container customers have also adopted Lambda.⁷ Leveraging
both options has its benefits, including fully managed services
that have deep integration with AWS infrastructure, support for a
wide range of use cases, abstraction from complexity, and a broad
ecosystem of partners.

⁷ "The State of Serverless,” DataDog, June 2022

17
So how do you frame the decision?

Customers choose Lambda when they have teams focused primarily on writing code
and no limitations on existing instances or container platforms. Lambda can scale
automatically from zero to millions of simultaneous processes in real time with a built-in
security posture. With Lambda, you can take full advantage of the most trusted, reliable,
and secure cloud provider. Lambda manages almost all server management tasks on
your behalf, including hardware, software, networking and facilities management, data
source integrations, and provisioning. That’s why, every month, Lambda has over a
million monthly active customers who generate trillions of invocations.

Customers often choose Amazon ECS on Fargate when they want to leverage containers
and container orchestrators for development but do not need to manage servers or
clusters. Amazon ECS delivers the easiest way for organizations to deploy and manage
containerized applications at any scale.

With Fargate, Amazon ECS supports serverless container orchestration, so you

can leverage more of the operational excellence of AWS when it comes to scaling,
maintaining availability, and securing your containerized workloads. Amazon ECS
provides tooling and built-in support that makes it simple to build and run containerized
applications on AWS. When more control over the characteristics of how your applications
run is desired, Amazon ECS on Amazon Elastic Compute Cloud (Amazon EC2) is
available, as well as Amazon ECS Anywhere for when you want to run container
workloads on your infrastructure. It’s no surprise that over 65 percent of all new AWS
container customers use the fully managed and versionless Amazon ECS.

18
 AWS Manages Customer Manages

Least Data source integrations Application code

AWS Lambda
Serverless functions Physical hardware, software,
networking, facilities

Provisioning

AWS Fargate Container orchestration control plane Application code Management tasks
Serverless containers and provisioning
Data source integrations
What you manage

Cluster scaling
Security config and updates, network
Physical hardware, host OS/kernel, config
networking, facilities

Amazon ECS/ Container orchestration control plane Application code Work clusters
Amazon EKS Physical hardware, networking, Data source integrations Management tasks
Container management facilities
Security config and updates, network
as a service config, firewall

Amazon EC2 Physical hardware, networking, Application code Security config and updates, network
Most
Infrastructure as a service facilities config
Data source integrations
Provisioning, managing scaling, and
Scaling
patching of servers
Management tasks

19
 Snoop Futuready
Snoop is a UK-based fintech startup that helps customers cut their bills, Futuready, an online insurance broker operating in Indonesia and Thailand,
pay off debt, grow their savings, and save where they spend, all without migrated to the AWS Cloud, seeking to improve its reliability and gain the
changing banks. The company achieves this by harnessing open banking agility to modernize its monolithic architecture. The company redesigned
data to provide hyper-personalized financial insights on its cloud-based its architecture using Amazon ECS with Fargate, which helped expand its
app in seconds. Working with lean resources, Snoop’s co-founders business regionally without additional IT investments. Futuready also reduced
looked to AWS for solutions to hasten time to market and build an app its infrastructure and operations costs by 35 percent and has accelerated its
that remains secure, highly scalable, and available 24/7. By leveraging deployment cycle for new insurance providers and products. The time taken
Amazon ECS with Fargate, Snoop scaled from zero to one billion to onboard new insurance partners was reduced from six weeks to only two
transactions in just two years, optimizing costs and reducing overhead. weeks, and the team can now release new features weekly instead of biweekly.
Additionally, Furtuready leverages Lambda to automate the process of
See the full story › completing insurance documentation, helping to reduce the amount of time
customers spend filling in their account details when purchasing a policy.

“All of our Amazon ECS instances use AWS Fargate, See the full story ›
which takes off a huge piece of overhead. As a fast-
scaling startup, that’s exactly what we need.”
“We can move faster and have gained 99.999% uptime
Jamie West, Senior DevOps Engineer, Snoop with AWS and its range of available managed services.
This leads to higher revenues by running more
projects and closing deals more than 50% faster.”

Dr. Nicholas Ruwan Dias, CTO, Futuready

20
 Taco Bell Coca-Cola
As one of America’s most iconic food brands, Taco Bell operates over When the COVID-19 pandemic hit in 2020, consumer habits changed literally
7,000 restaurants in the United States. During the COVID-19 pandemic, overnight. Coca-Cola responded rapidly with a no-touch drink dispensing
Taco Bell needed to rapidly shift to meet consumer demand for delivery. experience to go along with its innovative Freestyle drink dispensers. Coca-
The company runs nearly all of its infrastructure on AWS and uses Cola opted to build with Lambda, and as a result, its team was able to focus
serverless on AWS to focus less on managing servers and more on on the application rather than security, latency, or scalability. (With Lambda,
building business logic and data transformations to deliver real-time it’s all built in.) The new application launched in just 100 days, and now more
menu and restaurant information to its delivery partners. than 52,000 machines have the touchless capability.

Watch the video › Watch the video ›

“We have a menu that is very complex and (it) has to “Low latency is essential to the user experience, which is
be shared across multiple digital channels. Serverless why we’re committed to a serverless solution on AWS.”
fits that model really well.”
Michael Connor, Former Chief Architect, Coca-Cola Freestyle Equipment
Vadim Parizher, VP of Technology, Taco Bell Innovation Center

21
3 The Modern DevOps model Developer agility: abstraction, automation,
and standardization
Modern DevOps is the combination of cultural philosophies, Microservices architectures make teams agile and enable them to move
practices, and tools that enable an organization to quickly faster, which means you’re building more things that need to get released—
and safely develop software, release it to production, and great! However, you won’t get new features to your customers faster if your
maintain its target availability and performance. build-and-release process does not keep pace. Traditional development
processes and release pipelines are slowed mainly by manual processes
AWS has identified a set of common, broadly accepted
and custom code. Custom code is ultimately a long-term liability because
practices that, when adopted, provides a mechanism for
it introduces the possibility of errors and ongoing maintenance. Manual
building a high-performing DevOps organization. This
steps—from code changes and build requests to testing and deploying—are
approach takes a simple idea—continuous improvement—
the greatest drags on release velocity. The solution involves abstraction,
and applies it to everything in the DevOps lifecycle, from
automation, and standardization.
planning and code writing to deployment and monitoring.
We call this approach modern DevOps, and it’s centered To speed the development process, abstract as much code as possible,
around bringing developers and operations closer by sharing particularly the lines of non-business logic code required to develop
operational tasks like compliance, observability, resilience, and deliver production-ready apps. One way to do this is by employing
and infrastructure earlier in the development process and frameworks and tooling that reduce the complexity of provisioning and
enhancing it with artificial intelligence (AI) and machine configuring resources. This gives developers the ability to move quickly while
learning (ML). also enforcing best practices for security, privacy, reliability, performance,
observability, and extensibility throughout the development process.
Development frameworks give you confidence that your architecture will
support your business growth long term.

22
By defining your software delivery process with best-practice
A bit more detail
templates, you can provide a standard for modeling and provisioning all
infrastructure resources in a cloud environment. These infrastructure-
as-code (IaC) templates help teams get off to a good start because Continuous integration (CI) is a software development
the templates provision the entire technology stack for an application practice in which developers regularly merge their code
through code rather than using a manual process. changes into a central repository, then automated builds and
tests are run. CI most often refers to the build or integration
Through automation, you can create a repeatable motion that speeds stage of the software release process and entails both an
up your software delivery lifecycle. Automating the release pipeline automation component (a CI or build service, for example)
through continuous integration and continuous delivery (CI/CD) helps and a cultural component (learning to integrate frequently,
teams release high-quality code faster and more often. Teams that for example).
practice CI/CD ship more code, do it faster, and respond to issues quicker.
Continuous delivery (CD) is a software development
In fact, according to the 2020 Puppet State of DevOps Report, teams
practice in which code changes are automatically prepared
that employ CI/CD have a failure rate that is five times lower, a commit-
for a release to production. CD expands on CI by deploying
to-deploy rate that is 440 times faster, and a rate of deployment that is
all code changes to a testing environment or a production
46 times more frequent. Most notably, teams that practice CI/CD spend
environment after the build stage.
44 percent more of their time creating new features and code instead of
managing processes and tools.⁸

CI/CD pipelines have become the new factory floor for building Learn how Amazon automates safe,
modern applications. At Amazon, we started using CI/CD to increase hands-off deployments ›
release velocity, and the results were dramatic—achieving millions of
deployments per year and growing faster every year. To help companies
benefit from our experience, we built a suite of developer tools based on
the tools we use internally, so our customers can deliver code faster.

⁸ “2020 State of DevOps Report,” Puppet, 2020

23
C U ST O M E R S N A P S H O T

HyperTrack
HyperTrack is a self-serve cloud platform for live location tracking
through applications. When it launched, the company needed to build
a platform that could scale automatically to meet its anticipated
growth without reducing the time its developers spent on building
new features. HyperTrack opted to use AWS Amplify for a mobile
development framework and a serverless architecture, allowing it to
scale up and down automatically without engineering intervention. As
a result, the company has realized a 30 percent cost savings compared
to its previous architecture. A big part of that savings comes from
the elimination of operational resources for server management;
HyperTrack saves 40 hours of work every week while managing
millions of events.

See the full story ›

24
Transform to a Modern DevOps model
Creating a culture of ownership: Manage less,
innovate more with modern DevOps There are four foundational elements of CodeCatalyst that accelerate
Ultimately, innovation comes from people. So, enabling your people to deliver the transition to DevOps:
better customer outcomes is where modern application development starts.
At AWS, we use the concept of “products, not projects” to describe how this 1. Project blueprints that automatically set up everything you need to
cultural mindset impacts team structure. Simply stated, it means that the start a new project
teams that build products are responsible for running and maintaining them.
2. Flexible CI/CD automation for managing day-to-day software
Such a mindset makes product teams accountable for the development of
lifecycle tasks
the whole product, not just a piece of it.
3. On-demand cloud dev environments for a consistent build experience
After more than a decade of building and running the highly scalable web
application that is Amazon.com, we understand the importance of giving 4. Project and issue management for streamlined team collaboration
autonomy to our teams. When we gave our teams ownership of the complete
application lifecycle—from taking customer input and planning the road map
to developing and operating the application—they became owners, fully
empowered to develop and deliver customer outcomes. Autonomy creates One of the most difficult aspects of the move to DevOps culture is that teams
motivation, opens the door for creativity, and develops a risk-taking ethos must spend a great deal of time and effort constructing and maintaining
in an environment of trust. bespoke software platforms out of disparate tools, which can lead to
frustration. Software teams that use CodeCatalyst get pre-integrated code
While not inherently technical, embracing a culture of ownership remains repositories, cloud infrastructure, DevOps tooling, and collaboration tooling,
one of the most challenging aspects of modern application development. all delivered as a service they can quickly weave into the fabric of their day-to-
Empowering teams to become product owners involves changing the day work. CodeCatalyst does all of this in a flexible system that can grow with
motivation of your organization, the structure of your teams, and the work for them: It combines an opinionated experience that works out of the box and
which teams are responsible. Knowing firsthand how difficult yet impactful provides tooling modularity, so you can integrate or swap out tools as needed.
this transition can be has led AWS DevOps services team to make everything it By drastically reducing the amount of time and effort software teams spend
does about accelerating our customers through it. configuring software platforms, CodeCatalyst has the ability to turn IT from a
cost that is necessary to support business growth to a strategically competitive
That design philosophy culminated in Amazon CodeCatalyst, a complete
advantage that organizations cannot live without.
software development service that brings together everything teams need
to plan, code, build, test, and deploy applications on AWS.

25
Contino
Contino helps customers migrate their applications to the cloud,
improving their architectures by taking full advantage of cloud-native
features to enhance agility, performance, and scalability. Contino
is using CodeCatalyst to bring together everything its software
Building a culture of innovation
development teams need to plan, code, build, test, and deploy
applications on AWS. Providing a streamlined, integrated experience,

1
CodeCatalyst has helped Contino improve collaboration and increase Start with customers – Every innovation should start with
productivity, allowing its engineers to focus more time on rapidly a customer need and, ultimately, lead to delighting your
developing application code that captures customer value. customers. Relentlessly prioritize customer demand.

2
Hire application builders and let them build – Remove any
See the full story › obstacles that slow the process of building and releasing
products and features for customers. The faster you iterate,
the faster your innovation flywheel spins.

3 Support application builders with a belief system – Don’t pay

lip service to innovation. Live and breathe innovation in all
areas of the business, from leadership to sales to support.

26
Accelerate development with generative AI-powered tools
Part of enabling your developers to build better applications that deliver more creative aspects of coding. Amazon CodeWhisperer, an AI coding
exceptional customer experiences is making use of tools that can improve companion, radically improves developer productivity by generating code
productivity and free up more time for your team to innovate. It is very suggestions in real time based on developers’ comments in natural language
common for software developers to spend a significant amount of their time and prior code in their integrated development environment (IDE). Developers
writing code that is pretty straightforward and undifferentiated. They also can simply tell CodeWhisperer to do a task such as “parse a CSV string of
spend a lot of time trying to keep up with a complex and ever-changing tool songs” and ask it to return a structured list based on values such as artist,
and technology landscape. All of this leaves developers less time to develop title, and highest chart rank. CodeWhisperer provides a productivity boost
new, innovative capabilities and services. Developers try to overcome this by generating an entire function that parses the string and returns the list
by copying and modifying code snippets from the web, which can result in as specified. AWS ran a productivity challenge, and participants who used
inadvertently copying code that doesn’t work, contains security vulnerabilities, CodeWhisperer completed tasks 57 percent faster, on average, and were 27
or doesn’t track the usage of open-source software. And ultimately, searching percent more likely to complete them successfully than those who didn’t use
and copying still take time away from the good stuff. CodeWhisperer. This is a giant leap forward in developer productivity, and we
believe this is only the beginning.
Using tools that leverage capabilities like generative AI can take this heavy
lifting out of the equation by “writing” much of the undifferentiated code,
allowing developers to build faster while freeing them up to focus on the

27
Manage less, Why build modern applications on AWS?

innovate more Faster to By speeding up the build-and-release See the proof:

Modern applications help create
market cycle and offloading operational Urbanbase launches
competitive differentiation by enabling overhead, developers can quickly build services 20x faster
rapid innovation. By adopting services, new features. Automated test-and- with AWS ›
practices, and strategies that prioritize release processes reduce error rates, so
speed and agility, you can shift resources products are market-ready faster.
from business-as-usual operations to
differentiating activities with deep
Increase With a modular architecture, See the proof:
customer value. Experiment more and
innovation changes to any individual application iRobot uses Lambda and
turn ideas into releases faster. Nurture
component can be made quickly the AWS IoT platform to
an environment where builders spend
and with a lower risk to the whole manage its Roomba robotic
more time building and less time
managing. Modern applications are application, so teams can experiment vacuum cleaners ›
how organizations, including Amazon, with new ideas more often.
innovate with speed and agility.

Improve By automating test procedures and See the proof:

reliability monitoring every stage of the dev Siemens decreases
lifecycle, modern apps provide greater customer control system
reliability at deployment. Any issues alerts by 90% and reduces
are evaluated and addressed in infrastructure costs
real time. by 85% ›

Improve With a pay-for-value pricing model, See the proof:

TCO modern applications reduce the cost Save up to 80%
of over-provisioning or paying for idle on app maintenance
resources. Additional maintenance with Lambda ›
savings can be achieved by offloading
infrastructure management.

28
Learn more about modern application development

Re-platform to managed Build new modern applications with Transform to a Modern

container services serverless technologies and tools DevOps model
More than 80 percent of organizations today Save up to 80 percent of the time spent on By 2027, AI will dramatically increase developer
are using containers.⁹ More containerized maintenance and approximately 70 percent velocity by automatically generating code to
applications running in the cloud run on AWS on development when adopting a serverless meet functional business requirements for 80
than anywhere else.10 strategy for building modern applications.11 percent of new digital solutions in development
and early deployment.12

Resources Resources Resources

Amazon ECS workshop Innovator Island – Serverless web The Amazon Builders’ Library
Amazon EKS workshop application development workshop
Build a serverless app video tutorial

Recommended training (Classroom) Recommended training (Classroom) Recommended training (Classroom)

Running Containers on Amazon EKS Advanced Developing on AWS DevOps Engineering on AWS

Recommended training (Online) Recommended training (Online) Recommended training (Online)

Amazon ECS Primer Architecting Serverless Solutions Getting started with DevOps on AWS

9
“ CNCF Annual Survey ”, Cloud Native Computing Foundation, 2022 11
“uG idebook: Containers and Kubernetes on WS A ”, Nucleus Research, 12
“ IDC FutureScape e W bcast: o
W rldwide Developer and DevOps 2023
10
“ 2023 State of the Cloud Report ”, Flexera, 2023 October 2019 Predictions ”, IDC, December 2022

29
Need more help?

Learn more about building modern applications on AWS ›

Talk to an expert to implement the best practices of modern application development ›

Connect with an AWS Partner to accelerate your modernization projects ›

© 2023, Amazon Web Services, Inc. or its affiliates. All rights reserved.
30