Migration to cloud is a marathon, not a sprint. Here’s why and how to get started

When cloud computing went mainstream a decade ago, early adopters quickly realized business value by migrating “easy” workloads to cloud and adopting a cloud-first strategy for new systems. Fast forward to today, when the top IT priority is to modernize existing, mission-critical applications on hybrid cloud for better business agility and faster access to data insights. To fuel that effort, cloud spending is growing at six times the rate of overall IT spending.

Yet, surprisingly, only about 20% of enterprise workloads have migrated to the cloud so far. What is slowing down the remaining 80% of workload migration? What are the obstacles on the migration path? And what steps can organizations take to realize the agility, innovation and digital transformation that is only possible on the cloud?

Why migration to cloud is a multi-year journey

Let’s look at some of the major obstacles to enterprise cloud migration.

Organizations often encounter unexpected and discouraging costs and complexity, even during routine lift and shift migrations.
Developing new apps across multiple vendors and clouds, each with their own propriety tools and management systems, has led to challenges in terms of security, compliance and governance.
The larger the organization, the older the application suite. According to IDC[1] 47% of existing applications in Canada are more than five years old, which means they were probably not built with containers or microservices in mind. Migrating these legacy applications can be painstakingly slow because they first must be evaluated based on suitability, cost and feasibility.

All of this is going on while enterprises continue to make significant investments in traditional IT.

The reality is, for all but the newest companies, migration will be an incremental, multi-year journey. You won’t get there in one step. Modernized cloud native and legacy apps will coexist for the next 10-plus years, with business and IT value realized as each step is taken as part of a comprehensive digital transformation strategy.

Migration to cloud is a marathon, not a sprint. The question is, how can you accelerate and simplify that journey?

First, why migrate (and why not)?

Why migrate? Don’t move to the cloud because everyone else is, or you believe it could “automagically” save on infrastructure costs. However, if cloud is part of an overall change in how IT operates – creating more flexibility, agility, and the ability to respond quickly to changing business circumstances – you are on the right track.

Many organizations begin their migration efforts by rehosting an easy workload, but the highest value applications are in the functional areas where you want the most innovation and are the most amenable to change, such as client facing or internal business systems – or both.

Remember, not every application is a candidate for migration. Consider:

Large, stable, existing applications are probably a lower priority, until or unless you need to refresh hardware or the underlying infrastructure.
Security issues. Today’s hybrid, multi-cloud reality has IT leaders wrestling with how to move and manage workloads and apps between clouds efficiently without creating security risks. Sometimes this is more fear than reality, but in some cases, private cloud is the answer.
The impact of network performance on the migrated application. If the application is not designed as cloud native, consider network latency (lag) when you’re designing the system. If you are not located in a larger metropolitan area, complex tasks may take long to be received or processed. Applications that use overlapping data probably should be moved together.

Know your technical debt
One of the major challenges in moving to cloud is dealing with your organization’s ‘technical debt.’ If you’ve ever put off a project with the intent to finish it later, you have incurred technical debt. If you are behind on software versions, you have technical debt. And if you took short-cuts like hard-coding variables to get code to work ‘for now’; you definitely have technical debt!

Today, many enterprise workloads are drowning in technical debt. Migrating or refactoring workloads and applications to the cloud is probably the best way to pay off that technical debt – while improving flexibility and agility. Understanding software currency, release levels and requirements for underlying middleware, database and operating system are crucial before you begin.

Lift and shift, refactor or modernization?

Business growth in a digital world requires innovation and the ability to leverage data for competitive advantage. Cloud makes that possible by removing infrastructure costs and transforming enterprise processes. There are several approaches to take:

Replace: When considering the “build vs purchase” question, decide if you really need to build a proprietary, cloud-based solution – for things like CRM or ERP — or if you can leverage a “born on the cloud” packaged solution? (If a package does most of what you want, why re-invent the wheel?)
Rehost – Lift and Shift: This is the easiest way to migrate to the cloud. Recently developed applications are often “cloud ready” and can be easily moved to the cloud using a container or virtual machine with no modifications required.
Refactor: Refactoring (or redeveloping) an existing application to take advantage of cloud-native functionality can make it more efficient, scalable, maintainable or reusable. Or you can make minimal changes by removing some dependencies on mainframe while avoiding the cost of completely rewriting the program.
Rebuild – Application modernization using microservices: Leverage microservices and containers to build a cloud solution based on an existing application.

The American Airlines modernization story

The reality is, you will use all of these techniques, depending on your application mix and how you intend to use them going forward.

Take American Airlines, who used this combined approach to save money and improve customer service. Most of us have endured a flight cancellation due to weather events. IBM helped American Airlines build a mobile Dynamic Rebooking app on the IBM Cloud to vastly improve the customer experience. Behind the scenes in parallel, using the agile, product-focused Garage method, American Airlines migrated core VMware applications from its legacy environment to IBM IaaS in just four months. The result? Major capital cost savings while ensuring secure integration with back-end systems.

This is a great example of a large enterprise leveraging cloud to digitally transform their business quickly, through a mix of cloud-native innovation development, lift-and-shift migration and modernization of important business functions with microservices.

How to evaluate application readiness

To determine the complexity of moving legacy applications to cloud, ask questions such as:

Is it mission critical?
How complex and time-consuming will the move to cloud be?
What is the value generated to the business by being able to scale the application? The greater the need to scale, the higher the migration priority.
How frequently are you changing or innovating in this business area? The greater the innovation required, the more value there is in moving to cloud.
How secure is the data and where should that data reside? If your application is linked to other high priority applications, it is advisable to move it as a bundle.

Determining migration priorities

IBM doesn’t use a one-size-fits-all approach to cloud migration. With a focus on enterprise innovation, data privacy, security, and your specific business priorities, IBM uses a matrix that evaluates complexity, time to value and frequency of updates to create a cost-benefit analysis showing which applications should move to cloud first, be retired, or kept on traditional IT (for now).

The evaluation of migration priorities is stored in a database that can be queried, revisited and revised on a regular basis. As the centre of “data gravity” shifts in your environment, priorities will also shift, so data and database dependencies are identified to ensure that all complimentary tools/datasets are migrated to the same cloud to optimize integration.

Take the next step in the journey

One of the main reasons that companies accrue technical debt is that their IT departments are already overwhelmed with everyday tasks. An experienced cloud migration partner can accelerate your switch to cloud and help you avoid costly mistakes along the way.

If your company’s application portfolio is more than five years old, not built for the cloud, nor optimized for containers, yet digital transformation is your top priority, it’s time to think about taking the next step in the multi-year journey of migration to the cloud – and the path from complexity to simplicity.

Watch the video Tips for a Successful Cloud Migration

Read the blog Cloud migration checklist: 5 steps to a successful journey

Explore the benefits of Moving your applications to the cloud

Experience the Garage

[1] IDC Canada’s PaaS Ecosystem Research Program’s 2018 survey on PaaS Adoption in Canada, January 2018),