3 most common cloud migration strategies

2021 11 26 · 3 min read

Traditionally well-established companies have their own data centres, or they rent servers from external providers. These companies typically host IT systems on-premises or in third-party data centre servers. For the last decade, companies ran their workloads on virtual machines and bare-metal hosting solutions.  

However, when purchased hardware becomes obsolete, it starts to fail and can lead to unexpected hardware outages. In such cases, it is necessary to decide what to do with the existing systems to ensure the reliability and uninterrupted functioning of the company’s operations.  

Instead of purchasing new hardware and migrating the existing systems to other on-premises hardware servers, most Baltic Amadeus customers are choosing to migrate to public cloud services. This blog post presents 3 migration ways and advice on how you should choose the most suitable way. 

Hybrid cloud solutions   

Companies that already have their hardware and infrastructure components choose a hybrid cloud approach. Such a strategy is not cost-effective to retire if the existing infrastructure is still supportable and stable. So, extending the existing on-premises infrastructure to the public cloud is usually the best-recommended solution.  

By creating a hybrid cloud connection, companies can easily take advantage of the benefits provided by the public cloud by ordering only the services they need at a particular time (on demand). The most popular scenarios to extend systems to the public cloud would be to establish the Disaster Recovery (DR) site, migrate development and/or testing environments, store backup files. 

Usually, a hybrid cloud solution starts with expanding a company’s intranet network to cloud services by configuring site-to-site or point-to-site VPN connections. After that, the IT department can use the public cloud resources to deploy new systems and service components by configuring the services using an intranet or extranet network connection.  

When a company does not have sufficient resources to create a DR site or development or testing environment, typically, it is cheaper to use the public cloud services than buy new hardware for their data centres. When it is necessary to create a high-reliability site, the DR must be in a separate data centre and on different hardware than the primary site environment. The public cloud can help to implement recent changes quickly and easily.  

In addition, when a company’s IT department wants to test innovative solutions or make a Proof of Concept (PoC), it is more rational to rent public cloud resources, which are usually cheaper, than buy hardware. The cloud providers offer flexible prices for test/dev resource pools, and it is easier and faster to implement a solution.   

Lift and Shift  

The following strategy is one of the easiest and most ineffective infrastructure hosting solutions to migrate existing IT systems or services to the cloud. Such a migration strategy is typically applied when a company wants to quickly migrate its existing on-premises resources to another data centre. It usually happens when the on-premises hardware fails, DR resources are not pre-arranged, or a company’s data centre lease runs out.  

When a public cloud data center is chosen, the company must migrate its entire IT systems and solutions to another vendor. Sometimes, companies want to try public cloud services with minimal system changes. Still, their legacy systems require additional time and a more significant investment to upgrade them for modern deployment.  

Typically, a Lift and Shift migration means virtualising the on-premises systems to virtual machines (VMs) and transferring them to the public cloud data centres. However, it is only the first step in migrating systems to the cloud. The systems also need to be modernised to reduce total costs of ownership (TCO) and become cost-effective, faster, and more reliable. Nonetheless, the Lift and Shift migration solution is often highly effective in moving cold backup files to cloud disk storage, which is usually cheaper than maintaining and expanding an on-premises disk storage system. It is necessary not to forget to calculate the data transfer costs, as data transfer to cloud storages usually are free of charge but transferring back to on-premises or other clouds – is not free. 

System modernisation  

It is always desirable to modernise the custom-developed systems during the migration to take advantage of all the benefits of the public cloud. The already mentioned – Lift and Shift migration method helps to migrate COTS systems or older monolith architecture systems deployed in virtual or physical machines. In this case, the modernisation approach helps upgrade the system to the newest versions and technologies and decouple its multi-tier architecture.  

Suppose that the current on-premises system is hosted as a monolith or is a legacy computer system or application program, which uses obsolete standards or technologies that are no longer supported and outdated by manufacturers but still in use. In that case, companies need the help of solution and technology architects to put together the modernisation path. System modernization project usually involves decoupling the monolithic system into components based on multi-tier and micro-services architecture.  

Then, the required system changes might follow, including: 

Sometimes, companies have an architectural principle to store all their data and files only on-premises. These companies can migrate the system’s presentation and application tier components to the cloud while all their data stays in the on-premises resources. However, many public cloud providers have data centres worldwide. The data can be stored securely and meet any regulatory requirements, even during the migration to the cloud.   

Many companies also choose containerisation solutions, meaning all the individual systems and operational components are brought to Kubernetes containers and then deployed to the cloud platform. Modernisation using containerisation allows to have a cloud-agnostic architecture of the system core components, and application PODs can be deployed to any cloud provider on an on-premises Kubernetes cluster.   

When a company already has its Kubernetes cluster, extending the existing solution to the public cloud by using hybrid cloud solutions is easy. It allows scaling the current system horizontally and vertically using public cloud resources. Part of the system’s container can then run on-premises, while the other is run on a public cloud platform. In this way, the company can free up its hardware resources or expand the system when an additional capacity is required, or if different system functionalities need to be deployed.    

Finalisation of the chosen migration strategy 

There are many ways a company can migrate its existing system to the public cloud. It all depends very much on the needs and requirements of the company.  

After the current system migrates to the public cloud, the system services should constantly be monitored, maintained, and given sufficient attention to select the appropriate plans and solutions to ensure cost-effectiveness.   

Why Baltic Amadeus? 

Baltic Amadeus has many years of experience in developing and implementing new systems for our clients, as well as in modernising and redesigning existing supported and legacy systems. For this reason, we are a great IT partner for your business.   

Our architects will use the Well-Architected Framework and the TOGAF ADM methodology to prepare the cloud solution, including architecture vision, system and technology architectures, migration, and modernization plans.  

Moreover, the DevOps team will help to migrate existing or set up new system environments using Infrastructure as Code technology.   

Lastly, the support team will take over the maintenance and support of public cloud and hybrid platforms, ensuring proactive infrastructure care and SLA on agreed terms.