Assess each workload’s suitability before moving it to the cloud or a different cloud platform
Moving to the public cloud is not an end in itself but rather the first step in an ongoing process. If you have mastered the key steps of a cloud migration for some of your workloads, you will likely decide to move more at a later stage. Or you may be prompted to migrate to a different cloud because of reduced costs, better functionality or security. But before you move a thing, you need to ask some key questions.
1. Do you have executive buy-in?
If you can’t answer “yes” to this question, your cloud migration will fail. Although most enterprises aim to have allocated 80% of their IT hosting to the cloud by 2024, few executives across the C-suite have complete confidence that their company’s cloud migration initiatives will deliver the results promised. That’s why it is so important to see cloud as far more than simply an IT program. You need to secure leadership buy-in for it as a platform to advance the kind of efficiency, innovation and growth that will ensure the company’s future success.
That kind of buy-in pivots on the persuasiveness of your migration plan, which needs to demonstrate a careful, systematic and strategic approach to what you will migrate where and – most importantly – what business challenge it will solve. Consider your company’s current growth, scale and pain points to present a convincing case for the genuine business value to be generated by the migration.
2. What does the business want to achieve with this migration?
Gartner predicts that the proportion of new digital workloads deployed on cloud-native platforms will have increased from 30% in 2021 to more than 95% in 2025. Before moving workloads to different clouds or shifting remaining on-premises workloads to the cloud, companies need to understand the reasons why they are proceeding with their specific migration.
Driving factors include cost savings, a reduced infrastructure burden, scalability, availability and an improved user experience. Perhaps growth is stretching the company’s on-premises systems and infrastructure to their limits, and new systems need to be established that scale in line with growth. Maybe the goal is to streamline procedures and processes for developers and users, make IT more efficient to deliver faster on customer requirements or increase agility to boost responsiveness to changes in the global marketplace.
Cloud migration helps to unlock innovation at scale and at pace, allowing businesses to attain milestones that are key to their success but that would take years to deliver without the cloud. Whatever the objective, business leaders need to be clear on the purpose of the migration to ensure progress toward its achievement can be tracked and measured.
3. Is this workload a suitable candidate for the cloud?
It makes sense to deploy many frontend applications to the public cloud because their tendency to change frequently will benefit from the flexibility the cloud enables. Workloads most suitable for the cloud tend to have short lifecycles, experience frequent spikes in usage and need quick infrastructure deployment.
But not every workload type is a good fit for the cloud. Applications that need sub-millisecond latency are generally best left on-premises. If you are running a financial trading system, for example, the cloud would not be the optimal solution for this element of the platform. Manufacturing is another vertical that is super sensitive to time lags, which is why some of the marquee car manufacturers use a system like AWS Outposts to build and run applications on their premises without the need to introduce relatively high latency to the cloud compared with on-premises.
Another situation where leaving workloads on-premises may be a better option – at least in the short or medium term – is if you have invested substantial Capex (capital expenditure) in your on-premises infrastructure and need to extract maximum returns from it before moving to the cloud Opex (operating expenditure) financial model.
However, contrary to common perception, compliance is generally not one of the reasons why a workload should be kept on-premises. In fact, the potential for increased visibility and tighter controls make the cloud an arguably superior location for these kinds of workloads. Take the EU’s General Data Protection Regulation (GDPR), for example: With careful attention to the regulations and a solid understanding of cloud, it is possible to deliver a superior return on investment with a cloud deployment rather than in a data center.
It’s useful to consult a cloud partner for a thorough cloud workload analysis to determine the optimal deployment model for your workloads. This will give you a clear picture of your current workloads’ technical landscape, challenges and dependencies; the value of migration against the effort involved and the optimal sequence of workload migration.
4. Why this cloud?
Although most organizations are using multiple clouds, they are not always using them efficiently or strategically. That’s partly because many public cloud initiatives are the result of shadow IT, with software, devices and services being bought without the knowledge or control of the IT department.
Workloads are also being allocated to clouds on the basis of expedience rather than analysis. Without accurate data about how infrastructure is performing for each workload, choosing the right cloud for each workload is a question of luck rather than informed decision-making.
Consider the reasons why you are moving your workloads to a specific cloud. The major cloud providers all have specific strengths that may suit your business use case. For example, many customers migrate to Amazon Web Services (AWS) for the breadth and depth of its services, existing Microsoft customers often choose Azure to innovate with services they already use and Google Cloud stands out for its data analytics capabilities.
Theoretically, you can run the same workload in multiple clouds, but you are catering to the lowest common denominator if you do. Each cloud provider has native services that simply aren’t available on rival clouds, so it is not realistic to expect your workloads to run seamlessly across any cloud provider. Leveraging public cloud for the innovation it enables means playing to the strengths of your chosen cloud when you choose where to deploy your workloads.
5. Which approach will you use for the migration?
Depending on your previous experience with cloud migration and your company’s access to the appropriate expertise, your migration approach will fall into one (or a combination) of five categories: rehost (lift and shift), refactor, replatform, rebuild, replace or retire:
- Rehosting is the most straightforward approach because it involves simply redeploying existing data and applications on cloud storage and compute resources without modification. It requires relatively little cloud expertise, but it does not leverage the cloud’s capabilities and is not suitable for all workloads.
- Refactoring is more complex because it involves re-architecting applications before moving them to a cloud environment, but it also delivers a high return on investment because it capitalizes on cloud-based features and their inherent flexibility and elasticity.
- Replatforming is a middle ground between rehosting and refactoring that involves some workload modification to leverage cloud-enabled benefits such as automation and improved scaling.
- Rebuilding essentially recreates the workload from scratch to leverage the cloud environment’s functionality to full effect. This requires a high degree of skill and commitment.
- Replacing workloads with a cloud-native application means you retire an existing application and migrate only the data required to run it. This could be a prudent option if it’s easier to use a utility the cloud vendor provides rather than attempting to deploy and use the same tools you run on-premises.
- Retiring workloads that have served their purpose is important to avoid wasting time and resources during your migration. However, before you retire anything it is important to identify any upstream dependencies.
The migration approach you choose will influence everything from your cloud costs to your architecture decisions.
6. What will it cost?
Budgeting for a cloud migration must factor in all stages of the process from pre-migration planning through to the post-migration operation of the workloads in the cloud. Even if leadership is united on the need for a migration, they must be aware of the costs involved in order to make decisions about how to proceed with the move and at what pace.
Cloud total cost of ownership (TCO) is a method used to calculate the various costs involved in operating workloads in the cloud over their lifetime. Every cloud deployment is different, but it is important to compare what it costs to run the same application on-premises and in the cloud. The application’s functionality must be considered in its entirety, particularly requirements such as security, other application dependencies and other areas that can contribute substantially to costs. Another important element of cloud TCO is the cost of addressing the skills gaps for bringing teams up to speed on the different cloud platforms associated with the migration.
7. What are the security and compliance implications of the migration?
The security controls and practices you have built for your on-premises environments will not work for the cloud. One of the primary differences you will have to accommodate is the shared responsibility model: The cloud provider bears most of the responsibility for infrastructure security, whereas the cloud customer is responsible for configurations and settings under their direct control – such as data and applications. This can also be defined as “of the cloud” versus “in the cloud,” with the vendor accountable for the cloud’s security and the customer responsible for securing what is in the cloud.
Another feature of cloud security is that the software-based nature of the cloud creates specific requirements for controls and processes that you may need to address with new technologies. You may also have to restructure governance workflows and alignments to make them more agile and continuous. Be prepared to involve representatives from a wide range of stakeholders and technical disciplines.
8. Do you have access to the resources required to perform the migration?
A dearth of skilled IT workers is slowing the adoption of cloud-related technologies. In its 2021-2023 Emerging Technology Roadmap, Gartner revealed that IT executives consider the talent shortage the biggest barrier to deploying emerging technologies – including databases, machine learning, advanced storage and analytics – all of which are cloud-enabled.
Once you have the relevant skills in place, you need to ensure that your team operates as one cohesive unit: The process falls apart if your data engineers work in isolation from your networking experts, for example, and everybody considers the job done once their specific piece is complete. Encourage active participation by everybody and instill the ethos that nobody is finished until everybody is finished. This is a primary reason for senior stakeholder engagement and buy-in to the process to ensure that there are no blockers when it comes to building holistic teams charged with the delivery of the cloud journey. Otherwise, the migration project will stall as roadblocks mount.
If you lack the in-house skills to implement your cloud migration, you could consider training and upskilling your team to take on the task. Bear in mind that this process will take time and will also involve taking people away from the valuable work they are already doing. Engaging with a cloud partner who has the expertise and experience to guide you through the migration could be a prudent move and potentially accelerate your journey with sound, experience-led advice from seasoned architects.
9. What should you do next?
Having asked all these questions, you may be left wondering whether you have the necessary skills and resources to carry out a cloud migration successfully. However, there is an alternative: By engaging with a cloud partner like DoiT, you have access to migration experts who can guide you through the entire process at no additional cost.
A premier partner of AWS and Google Cloud, we offer multicloud expertise supported by valuable optimization, analytics and governance technology. Get help gathering requirements and setting business objectives, building the appropriate cloud architecture and scaling your infrastructure. With support like this, your next question should be “where do I start?”