The number of choices to make in a cloud migration process is certainly plentiful. In that process, it’s easy to focus on what’s immediately tangible, but it’s equally important to understand when those design and architecture decisions might be too short-sighted or what their ultimate impact is on the user experience - arguably the most important metric for any organisation attempting to succeed in today’s digital economy. 

Cloud migration has traditionally been - and still is - a very infrastructure-centric conversation. It is often based on a desire to exit on-premises data centres, perhaps for sustainability reasons or to reduce management overhead and costs. The headline direction is sometimes set with little consideration for how it’s to be executed.

For the teams charged with execution, the immediate challenge is understanding the performance of different pieces of cloud infrastructure to support making some appropriate migration decisions. 

Modern applications are composed of many different application parts or services that are variously distributed across clouds, data centres, regions, and availability zones (AZs). Which clouds, regions and AZs house different components and how they perform relative to the organisation’s user base, is of critical concern.

This was a key objective of last year’s ThousandEyes Cloud Performance Report, which examined how different elements of cloud infrastructure influences, and alternatively impacts, application design and cloud service performance. 

Understanding and being continuously aware of the performance characteristics of the underlying cloud infrastructure puts organisations in a better position to have some influence and control over the delivery of digitised, cloud-based user experiences.

But application design also goes beyond infrastructure. An application’s characteristics are determined well before the discussion on where and how to re-host it. 

The role of users in determining those characteristics needs to be taken into account in the cloud migration process. Doing so could lead to a very different set of infrastructure requirements for each application.

This is arguably the lens through which all cloud migration programs should be viewed.

Making users top-of-mind

User- or human-centred design (UCD/HCD) has conceptually been around for decades but is only now gaining critical mass. Gartner, for example, estimates that 60% of governments “will integrate HCD techniques into their digital service design process” by this year. One would expect private sector usage of these application design techniques to be even higher.

The growing adoption of UCD/HCD is on the money insofar as it can change the course of application development and cloud hosting decisions. 

It can also solve common problems in application design that have a flow-on impact on hosting decisions: where applications are over-engineered compared to user needs, and that over-engineering leads to the over-provisioning and over-specification of cloud infrastructure to support the application.

Over-specification of cloud instances is a common place organisations go looking for financial optimisation; rightsizing of instances is an obvious place to look, and it’s common for web applications to be able to run efficiently with fewer compute resources.

Without a lens that looks at the ultimate delivery of a digital experience, applications are often designed by developers to be as robust as possible inside of a given budget envelope. 

Cloud infrastructure decisions are then made to meet the perceived user requirement without having specific regard to or knowledge of what users can actually live with. 

There’s nothing nefarious on the part of developers and operations teams here; they’re simply working to meet cloud migration targets. 

But with better knowledge of users and expected performance, it may be that users can accept a slightly lower level of performance consistency or higher latency when making API calls to the application compared to the way it’s been specified. That could open the door to hosting it in a different AZ, cloud region or instance size, or to configure the application’s underlying infrastructure differently altogether. 

Organisations that have continuous visibility into the various cloud ecosystems are able to consistently make infrastructure choices that are aligned with user needs and to ensure these choices remain optimal, even as ecosystems evolve and new options become available.

The point is: this needs to be a core part of the cloud migration strategy. Doing so changes the complexion and, more importantly, the outcome - of the exercise.

Decisions change when users come first 

What should be apparent is the need to align the execution of cloud migration with the end goal. If the end goal is to create better digital experiences and improve application performance for users, then a volumetric application migration strategy is not the optimum strategy to achieve that outcome.

When applications are migrated in batches or tranches, specific consideration of individual application design risk is an afterthought. Individual application characteristics may not be considered until after the initial lift-and-shift is completed or until it comes time to migrate a specific application, and suitability issues are then encountered. 

At this point, a range of alternatives - from workarounds to refactoring to the ‘too-hard basket’ - may need to be considered. But urgency is an overriding factor: there will be little appetite to hold up the migration, so the easiest option is likely to be chosen. This may not be the option that best fits user requirements, but that simply enables the organisation to continue moving forward with the migration. Specific optimisations for users may be revisited later.

By contrast, approaching the delivery of digital experience on a more individualised basis from the outset of the migration will allow for the right foundational choices to help power-optimised performance now and in the future.