What is cloud migration?

Some of the biggest reasons to migrate to the cloud are to reduce costs and optimize infrastructure performance, utilization, resilience, and security. Migrating to the cloud can help organizations move spending from a capital expenditure model to an operational expenditure model, reducing the need to own, operate, maintain, and refresh expensive equipment in on-premises data centers.

In IT, capital expenditures (CapEx) are typically large, one-time investments in physical assets. For example, buying new servers, building a data center, or purchasing software licenses is considered CapEx. This is a depreciated capital expense, which means the asset loses value over time.

In contrast, operational expenditures (OpEx) are ongoing costs for running a business. With cloud computing, OpEx includes the flexible subscription fees for cloud services. This model allows for more predictable, ongoing costs rather than large, upfront investments. Migrating to the cloud can help companies improve performance and increase uptime. The cloud also helps companies modernize workloads, reducing burdensome licensing fees from legacy applications. The cloud can help organizations safeguard their data while increasing governance and compliance with regulators across the globe.

Organizations also migrate to the cloud to:

• Unlock data analytics: Migrating existing apps to the cloud can help businesses create more value out of tools they already use, such as CRM, SAP, marketing databases, and more. Getting these systems modernized in the cloud helps companies unlock their data and find new business opportunities while increasing organizational agility to address problems.
• Increase agility: The cloud provides organizations with on-demand IT resources, so they don’t have to wait weeks or months to build apps or install on-premises hardware. The cloud enables organizations to adapt to market changes and competitor actions much more quickly, increasing flexibility for go-to-market strategies.
• Consolidate and modernize data: By moving away from on-premises data centers, companies have the opportunity to consolidate their data and organize it in a universal data platform, opening up opportunities for more robust and sophisticated digital transformations.
• Leverage artificial intelligence: By moving organizational data and applications to the cloud, companies have close proximity to new and ever-evolving AI services and models, providing greater flexibility, agility, and performance for new AI applications and agents.

Organizations can choose from different types of cloud migrations, depending on their goals and objectives. 

A full data center exit migration is the process of moving all applications, services, and datasets from one or more data centers to a public cloud. Depending on the size of the organization, a full data center exit can be a long process requiring more than a year of planning, testing, and execution. 

An organization may want to move from one cloud vendor to another for a variety of reasons, including changes in service level agreements, better security practices, or for access to more advanced artificial intelligence and machine learning tools. This type of cloud-to-cloud migration is often used to avoid vendor lock-in or to access specific features on a new platform. Major vendors will typically have toolsets, services and lists of third-party vendors that can help an organization move from one public cloud provider to another.

Instead of a full data center exit, organizations may wish to migrate some of their software, services, or datasets from on-premises servers to the cloud where they can be more efficient and easily managed. Business intelligence, data analytics, customer resource management (CRM), machine learning and artificial intelligence apps, and datasets are typical types of applications that are migrated to public clouds.

More targeted migrations can consist of just a specific type of workload, which are resources running in the cloud that consume some type of resource, such as computing power or storage capacity. Examples include development environments, document creation and management, distributed databases, 3D modeling, or video encoding.

The most common migration strategies exist on a spectrum from a straight migration with little to no changes to apps and virtual machines (rehosting), to a complete rethinking of the entire digital infrastructure (re-architecting or replatforming).

• A rehosting cloud migration strategy aims to make as few changes during the migration as possible, redeploying applications to the cloud without making substantial changes to how they are configured.
• It is the most straightforward cloud migration strategy where administrators just “lift” their applications, workloads, virtual machines, and server operating systems and “shift” them to the new operating model in the public cloud. The same software that was running in the data center is now running in the cloud.
• The drawback to a rehosting strategy is that it often does not take advantage of all the cloud-native capabilities and efficiencies that make public clouds so powerful and useful.
• A rehosting migration strategy is best for organizations with predictable peak usage patterns (like the tax industry or food delivery), and is also commonly a ‘first step’ in a larger digital transformation project.

• A replatforming cloud strategy is the next step from a rehosting strategy, lifting the existing applications, workloads, and virtual machines and then optimizing them for the new cloud environment. For instance, a service may replatform a workload to the cloud to be able to take advantage of cloud-based microservice architecture, or containers like Kubernetes.
• The benefit of a replatforming cloud migration strategy is that the applications will have higher performance and more efficiency running in the cloud. The drawback is that replatforming takes more work to accomplish than a straight lift and shift. The new cloud platform will have a different underlying code base, which will mean several rounds of testing to make sure everything is running at its optimal level.
• A replatforming cloud migration strategy is best for organizations that want to take advantage of all the core competencies of the cloud such as elastic computing, redundancy, improved performance, and security.

• A refactoring cloud migration strategy means taking applications and re-engineering them to be cloud-native.
• Refactoring often means changing the code of an application without altering its front-end behavior or experience. For example, a refactored application may be broken up from long strings of code into more modular pieces that can better take advantage of cloud capabilities, thus improving the performance of the code.
• A refactoring strategy is often driven by the need to give existing applications new capabilities, such as increasing scalability, adding machine learning or artificial intelligence capabilities, or faster and more intuitive data analysis.
• Refactoring can be more complex than rehosting, but also allows an organization to set priorities on which applications are moved based on business needs, giving companies greater flexibility and the ability to keep up with the competitive ecosystem.

• A re-architecting cloud migration strategy is similar to refactoring but instead of restructuring how the application’s code works, it changes how that code functions in order to optimize the application and take advantage of cloud-native properties like scalability, security, and agility.
• One example of re-architecting an application is to take one large, monolithic application and turn it into several independent microservices.
• Re-architecting takes more time and effort than refactoring, increasing complexity while potentially introducing bugs or security issues into the new applications.

• A rebuilding cloud migration strategy is when an organization takes an application and rewrites it entirely for the cloud. It is often easier to build an application from scratch than it is to refactor its old code to work in a cloud environment.
• A rebuilding strategy allows an organization to plan from the ground up, choosing which cloud-native tools and capabilities to utilize from the beginning.
• The drawback in a rebuilding strategy is that it takes time and can introduce bugs or other performance issues that developers had not encountered in the legacy version of the app.

• A repurchasing cloud migration strategy is when an organization moves from a purchased on-premises application to cloud-hosted Software-as-a-Service equivalent. For instance, moving from on-premises collaboration software and local storage to Google Workspace’s cloud-hosted apps like Docs, Sheets, Slides, and Google Drive.
• From a resources perspective, a repurchasing strategy can be a lot easier than refactoring, rebuilding, or re-architecting.
• The drawback to a repurchasing strategy is that it can be a lot more expensive and organizations do not get the granular features of controlling their own cloud environments.

Enrich applications to utilize AI/ML and data analytics services for improving user experiences, extracting value out of data faster and more efficiently.

A full-scale migration can be a complicated, multi-year operation. Before beginning, organizations should have clear, attainable business objectives which includes understanding what the final cloud environment will look like, whether it's public, private, hybrid, or multicloud. Organizations should also to decide if they'll use third-party vendors or a cloud provider's tools, which apps will be SaaS, and how data will be organized in the new environment.

The skills required for cloud operations are different from those for on-premises environments. Cloud environments often rely on new operating models like DevOps, automation, and infrastructure as code. Organizations should either hire new employees or retrain existing staff to perform these functions. It's crucial to allocate time during the migration for employees to train and obtain new certifications to support the new infrastructure.

Security practices change when moving from an on-premises environment to the cloud. Identity and access management becomes more complex and critical. Organizations should understand new security tools and best practices, such as data loss prevention and zero-trust architectures. While cloud providers offer robust security tools, the organization is ultimately responsible for data governance and compliance policies when moving data.

Changing to a new operational model can make understanding all new costs and expenditures daunting. Cost management becomes an ongoing challenge, including new fee structures, billing cycles, and resource consumption. Organizations sometimes overprovision resources during the initial migration, leading to overspending. Fortunately, some cloud providers offer proactive recommendations to help adjust budgets and manage costs effectively.

Older or legacy systems may not be compatible with cloud environments and often require significant modification or replacement. This can be especially true for custom-built applications with outdated code, which can't take advantage of cloud-native features without extensive refactoring or rebuilding. Assessing application compatibility is a critical early step in migration planning to avoid unexpected technical hurdles.

Many organizations have complex networking topologies built over decades. Further some organizations require to retain their existing IP addresses to reduce risk and ensure application dependencies work.

Applications and datasets often have complex interdependencies. If these relationships aren't properly mapped before migration, there's a significant risk of data loss or corruption. Ensuring data integrity requires meticulous planning and testing to make sure all data is moved correctly and remains accessible and usable in the new environment. It's a key challenge that can undermine the success of an entire migration if not managed carefully.