Migrating and Modernizing Windows Workloads on
AWS
blogs.infosys.com/infosys-cobalt/public-cloud/migrating-and-modernizing-windows-workloads-on-aws.html
Joint blog post by Nirmal Singh Tomar, Principal Consultant at Infosys, Mandar Shete,
Principal – Cloud Practice at Infosys and Ashok Sundaram, Principal Solutions Architect
at Amazon Web Services (AWS)
Many enterprises have Windows workloads that need to be transformed to become more
flexible, scalable, and cost-effective. They typically choose to migrate these applications
to the cloud due to its flexibility, robustness, ease of deployment & management, and
scalability. Migrating and modernizing applications on AWS enables businesses to cut
costs, improve efficiency, and become nimbler, while maximizing their existing
investments.
If enterprises focus only on a ‘Lift & Shift’ approach to migrate their Windows workload to
the cloud, and choose computing comparable to on-premises servers, they frequently end
up with inefficient resource provisioning in the cloud, resulting in a sub-optimal TCO.
This blog provides an overview of common approaches and migration patterns, and
selection criteria that can help organizations modernize and migrate their Windows
workloads to the AWS cloud in an optimal manner.
Strategies for Migrating Windows Workloads to AWS
There can be various disposition strategies for transforming a traditional legacy
architecture to a cloud-based architecture, including Re-host, Re-platform, Re-locate, Re-
purchase, refactor, retire, and retain. Generally, enterprises select three strategies for
Windows workload migration – rehosting, replatforming, and refactoring. The diagram
below depicts the gradual stages and paths that enterprises may take to move their
existing Windows workloads to the AWS cloud.
The best migration strategy for enterprises depends on various factors like type of
application, data center sunset, license portability, and time to migrate. Using a “one size
fits all” approach like moving all applications to Platform-as-a-Service (PaaS) or designing
a cloud-native application architecture for all applications using Container-as-a-Service
(CaaS) may not always be optimal. Usually, we recommended a multi-path migration
approach to migrate Windows workloads to the cloud.
1/5
Level 3
Re-Factor/Re-Architect
(Adopt E2E Cloud Native
Level 2 Capabilities)
• Modernize application to run it on
Level 1 Re-Platform
Linuxcontainers on Amazon
(Leverage PaaS/CaaS
without ECSorAmazon EKS including
Existing ApplicationRe-host– (Lift and Shift) changing core architecture) AWS Fargate
(Migrate the Application • Migrate application to
• Refactor short
-lived tasks that can
and Database as is ) AWS Elastic Beanstalk run onAWS Lambda/Batch
On-Prem Windows platform
• Migrate the legacy • Migrate Microsoft SQL Server
• Legacy Monolith application toWindows • Migrate application to database toAmazon Aurora and
applications EC2instancesor self- Windows containers on access same usingBabelfish
• Legacy Database managed Docker based Amazon ECS orAmazon
Windows containers EKS • Modernize Microsoft SQL Server
database to new age databases
• Migrate Microsoft SQL like Amazon DynamoDB, Amazon
• Migrate Microsoft SQL DocumentDB, Amazon
Server database on self- Server database to
managed EC2 instances ElastiCacheand may more
Amazon RDS
Figure: Multi-path approach for cloud migration
Re-host:
The quickest and easiest way to migrate applications is through “re-hosting” or “lifting-
and-shifting”. By rehosting, enterprises can move their Windows workloads to the cloud
without any code changes. The primary purpose of this strategy is to reduce the server
management burden from on-premises resources, so that they can focus more on
building/optimize applications.
Re-host Options:
Amazon EC2: Suitable when you want complete control over the underlying
Windows server, storage, and operating system.
Self-managed Windows containers on Amazon EC2: If your workload is
container-ready, you can select a self-managed, container orchestration platform
running on EC2 to host Docker Windows containers. And it will also allow you to
have full control and access over the underlying server. It is important to note that
the Windows container images are bigger than the Linux container images.
Self-managed databases on Amazon EC2: Suitable when you want complete
administrative control and flexibility over Microsoft SQL Server relational database.
Re-platform
In this path, you optimize workloads by upgrading their components (such as the
operating system, storage, or database) without changing its architecture and with
minimum code or configuration changes. The primary purpose is to take advantage of
cloud features like scalability, elasticity, better management, and resilience without
significantly changing applications.
Re-platform Options:
2/5
AWS Elastic Beanstalk: Elastic Beanstalk reduces management complexity for
applications, making it an excellent solution if you are new to AWS or need to
launch a web application as rapidly as feasible.
Windows Containers Using Amazon ECS or Amazon EKS: ECS or EKS can be
used when your workload is container ready and you want to offload management
of underlying infrastructure to AWS. EKS brings an industry-standard, Kubernetes
platform with the associated tooling, ecosystem, flexibility and should be the
platform of choice if multi-cloud or hybrid-cloud deployment is involved. On the other
hand, ECS offers more in-depth integration with AWS services and can be chosen if
you are looking to work solely in the AWS cloud. AWS Fargate can also be an
option with ECS/EKS as serverless compute backend. At present only ECS
supports running Window containers on Fargate. With EKS, you will have to deploy
Windows containers on self-managed nodes.
Amazon RDS: Amazon RDS for SQL Server can be used when you want to host a
full-featured SQL Server relational database while outsourcing database
maintenance.
Re-factor/Re-architect
The primary purpose of this strategy is to optimize applications when there are compelling
business reasons to make significant changes such as adding features, increasing
performance, or enabling scalability that are difficult to achieve without significantly re-
architecting the application. For example, you can save costs by porting .NET
applications to .NET Core / .NET 5 and running them on Linux containers.
Re-factor/Re-architect Options:
Linux Containers Using Amazon ECS or Amazon EKS: For modernizing long-
running applications, you can refactor .NET standard applications to .NET core
microservices and containerize them to Linux containers on ECS/EKS, with EC2 or
Fargate option for hosting.
AWS Lambda: You can select the serverless path using Lambda for short-lived
tasks that can be triggered, based on an event or a scheduler. Lambda provides
two options for.NET 5 workloads: .NET 5 Lambda functions that use Lambda
container image and .NET 5 runtimes that use a custom runtime with Lambda.
Amazon Aurora: While modernizing applications, migrating from a commercial
database like SQL Server or Oracle to Aurora can help you save database licensing
costs. You can also use Babelfish, which enables Aurora PostgreSQL to understand
commands from applications written for Microsoft SQL Server, to reduce the
application changes needed.
New-age Databases: While modernizing applications, enterprise can consider new
age purpose built databases such as Amazon DynamoDB, Amazon DocumentDB,
Amazon ElastiCache and many more, which provides high performance, scalability
and agility for various use cases.
Infosys & AWS Services for Windows Workload Migration:
3/5
Infosys and AWS have a wide range of services for different migration strategies. Here
we will talk about some key ones.
Infosys Modernization Suite: Infosys Modernization Suite, part of Infosys Cobalt,
simplifies and accelerates application migration and modernization to the cloud, including
Windows workloads. It covers all phases like assessment, application code remediation,
cloud native development, environment provisioning and deployment. It complements the
AWS services mentioned below and enables enterprises to start their projects in a matter
of minutes, with up to 40% reduced effort. Key features include:
Cloud assessment and disposition for applications
Automated environment provisioning using infrastructure-as-code (IaaS / PaaS /
Kubernetes)
App containerization and deployment
Modernization of monolithic applications to microservices (including mainframe)
Accelerated cloud native development including microservices, APIs and batches
Accelerated adoption of serverless architecture
Proprietary database migration to open source or cloud managed RDBMS or
NoSQL
AWS services used with various migration patterns:
Re-hosting
AWS Application Migration Service: Migrates applications/databases from
physical infrastructure/VMware vSphere, Microsoft Hyper-V to supported versions of
Windows.
Re-platforming
AWS App2Container: Reduces manual effort required to containerize applications
using automated discovery, dependency extraction, docker image creation, and
creation of ECS/EKS task/pod definitions for deployment.
Windows to Linux replatforming assistant for Microsoft SQL Server
Databases: Automates the steps needed to migrate databases from Windows OS
to Linux via backup and restore.
In-place upgrades using AWS Systems Manager: Automated upgrades instances
running for Windows Server 2008 R2, Windows Server 2012 R2, Windows Server
2016, and SQL Server 2008 R2 with Service Pack 3.
Re-factor/Re-architecting
AWS Microservice Extractor for .NET: A new service introduced at re:Invent
2021 that can help enterprises in extracting efficient microservices from complex
ASP and .NET monolithic applications to improve agility, uptime, and scalability.
AWS Migration Hub Refactor Spaces: Another new service introduced at
re:Invent 2021 which can help enterprises combine existing applications and
microservices into a single application. It also allows different approaches for
architecture and technology, team alignment, and the process between parts.
4/5
Porting Assistant for .NET: Modernizing from .NET to .NET Core can be time
consuming, as applications can require remediation by moving to alternative
supported packages or APIs. Porting Assistant for .NET helps in estimating the
effort and application changes required.
AWS Schema Conversion Tool (SCT): Migrating from a commercial engine like
SQL Server to an open source engine like PostgreSQL or MySQL can be time
consuming due to manual effort required for conversion of the database schema.
SCT makes heterogeneous database migrations easy by automatically converting
most database objects to a format compatible with the target database.
Data migration
AWS Database Migration Service (DMS): Supports data migration between both
homogeneous (e.g. SQL Server on-premises to Amazon RDS for SQL Server) and
heterogeneous (e.g. SQL Server to Aurora PostgreSQL) databases. It includes both
one-time migration and continuous replication to keep the source and target
databases in sync.
Conclusion
Enterprises have growing expectations for functionality, performance, mobile access,
ongoing data expansion, cost reduction and a high user experience. This necessitates
new methods of providing Windows workloads, such as hosting them on the public cloud.
Earlier, cloud migrators were driven largely by IT cost savings through lifting and shifting
their workload to the cloud. However, today enterprises are realizing that modernizing
their Windows workloads faster on AWS can provide them more value via efficiency,
performance, and agility at a lower cost. Infosys and AWS work closely with global
enterprises at every stage of workload migration journey. With our proven tools,
methodologies, and expertise, we ensure an accelerated and hassle-free migration to the
cloud.
Author Details
Nirmal Singh Tomar
Nirmal has 16 years of development experience and has been working on Public
Cloud Architecture and Transformation services for more than seven years. Currently he
is a part of the Infosys ADMCOE Cloud practice and leading various initiatives with AWS
and GCP, along with solutions. Nirmal has a keen interest in industry and cloud trends,
and working with hyperscalers to build differentiated offerings. He is extremely passionate
about utilizing technology in helping customers in their cloud transformation journey.
5/5
