From overshooting the budget and missing deadlines to finding that things ‘just don’t work’, cloud migration failure comes in various forms. But all these issues can be avoided or rectified with pragmatic cloud migration management.
So, what are the main reasons that cloud migrations fail, and how can risks be averted?
1. The strategy falls short
Proper planning is the cornerstone of successful cloud migration. But that doesn’t mean you need to prepare for every eventuality upfront. It’s about having sufficient understanding of the existing infrastructure and the target environment.
Conducting a cloud readiness assessment and building a business case for the move is essential. This provides enough information to predict how much work is needed to ensure existing workloads will function well in the cloud. Once that’s established, likely costs, timescales and returns can be ascertained, so all stakeholders know what they’re signing up to before migration begins.
Once the initial assessments are complete, strategy development can begin in earnest. Key decisions – such as whether to modernise workloads before, during or after migration – can be taken in confidence, based on clearly defined goals agreed in the early stages. It also becomes easier to prioritise tasks effectively, according to the organisation’s needs.
One key element of the strategy should be ensuring the ‘migration bubble’ is kept to a minimum. During a largescale migration, dual running costs for the cloud and datacentre can lead to an eye-watering spike in short-term spend. To keep this under control, it may be more beneficial to evolve legacy applications than opt for a full rewrite. Sometimes, it’s best to go for the most straightforward rehosting option, a lift and shift. Just make sure any applications moved in this way are earmarked for modernisation as soon as possible.
Top tip: conduct a detailed assessment of the existing infrastructure, then consider any improvements needed to ensure good performance in the cloud.
Check out our whitepaper, Supercharge your cloud migration with DevOps.
2. Inadequate skills or experience
Largescale cloud migration usually only happens once in any organisation, so it stands to reason that many existing members of the team won’t have prior experience. This can result in a combined lack of confidence and capability, which is a recipe for slow, bumpy progress. Worse still, it can result in serious oversights that might compromise security or reliability.
It’s not just the migration itself that’s at risk here. Ongoing management in the cloud environment is also likely to suffer, leading to poor cost control and an inability to leverage cloud benefits such as agility and scalability.
Acknowledging that new, different skills will be needed is the first step to overcoming this challenge. Simply hiring more experienced staff is one option, but it’s rarely enough in itself. They lack the in-depth knowledge of the existing infrastructure that’s needed to facilitate a smooth transition to the cloud.
Investing in and supporting the current team is usually the best way forward. Partnering with a trusted third party that has a proven track record handling migrations to the target cloud provider can help secure better outcomes. Ideally, they’ll be well-versed in modern ways of working too. If so, they can get you started with advanced infrastructure engineering methods such as containerisation, Infrastructure as Code and microservices.
For ongoing cloud management, ensure different members of the team receive relevant training in cloud engineering or DevOps and Agile ways of working. Longer-term mentoring, coaching or operations engineering support also pay dividends.
Top tip: look for third party support from organisations that are certified by your target cloud provider and consider a training needs assessment to pinpoint in-house skills gaps.
Download our whitepaper Modernise your workforce for further guidance on how to support teams’ continual learning.
3. Doing too much too soon
Tempting as it may be to prioritise valuable, business-critical applications for the early stages of migration, this is rarely advisable. Even with third party support and training in place, your team needs time to get to grips with the migration process and day-to-day working in the cloud.
Much of the time, critical applications are also the most complex. Successful migration is likely to require focused attention to address any technical debt and prime the infrastructure for the cloud environment. What’s more, if the migration goes wrong, the repercussions could be hugely detrimental to the business, and take a long time to rectify.
It’s often better to start with an application that’s relatively simple to move but offers quick wins once it’s in the cloud. This is a great way to build the team’s confidence and capabilities, steadily progressing to more complex applications that reinforce recent learnings. Small steps which accumulate and gather momentum over time are a quicker route to success than an overambitious approach that ultimately implodes.
Taking the time to build proofs of concept or migration blueprints is another good move. It enables potential problems to be identified and rectified ahead of live migration, enabling secure and steady progress.
Top tip: slow and steady wins the race, especially for complex cloud migrations. Rushing the process is more likely to result in serious issues such as security vulnerabilities.
Prepare your strategy, prepare your team, take it steady
Thinking holistically about cloud migration – from the mechanics of the move, to managing short and long-term costs, to upskilling staff – is the secret of success. There is no single right way to migrate. But spending time upfront assessing the existing infrastructure and understanding the extent of any skills gaps is definitely the best way to get started.
Our whitepaper Move to the cloud and modernise workloads takes a more detailed look at some of the themes covered in this blog.
Here at DevOpsGroup we’ve got lots of experience handling AWS and Azure cloud migrations as well as providing cloud managed services. What’s more, DevOpsGroup Academy offers practitioner-led training to help teams get up to speed with modern ways of working and cloud engineering best practice.