Taking a Machine-First Approach to Cloud Migration

Many companies have put cloud computing at the center of their digital transformations. Yet a number are struggling to shift their systems from on-premises technology to public clouds. Several challenges are slowing them down, but they can be overcome with the right cloud strategy.

An early challenge – getting companies to risk putting their IT infrastructure, applications and data in the hands of a public cloud company – has been diminishing rapidly. Executives in many industries have come to recognize the importance of cloud computing in creating new business models and new, compute-intensive business processes. They increasingly rely on cloud providers like Amazon Web Services (AWS), Google and Microsoft, as well as niche cloud players that are building deep competencies. Netflix turned to AWS starting in 2008 to stream movie and TV programming content.1 Now many enterprises deliver their digital offerings through public cloud providers. CapitalOne now depends on AWS too to handle its ever-increasing customer contact center inquiries.2

More firms are leaning on cloud providers to deliver digital customer experiences, information and other online offerings for customers in minutes or hours, not days. Meeting such demands requires unprecedented amounts of computer power. And with companies facing constant pressure to reduce data processing costs while complying with laws like the General Data Protection Regulation (GDPR), the cloud offers an increasingly attractive alternative to maintaining their own data centers.

Once companies embrace the cloud, however, they find the move is far more than flicking an on-off switch. Many firms overlook the intricacies in planning and shifting to the cloud, which can lead to inefficiencies and slow their digital transformations. A 2018 survey of nearly 1,000 IT professionals found that 35% of cloud spending was wasted because of inefficient use of computing resources.3 These firms miss out on opportunities to take full advantage of a Business 4.0 world in which cloud compute-intensive analytics and AI technologies are increasingly necessary.

Why Traditional Cloud Migrations Fall Short

In spite of the aforementioned inefficiencies, cloud computing has become a staple of the modern IT budget. IDC expected spending on cloud services worldwide to reach $160 billion in 2018, and that by 2020 more than 90% of enterprises would use multiple cloud services and platforms.4

The benefits are inarguable. The cloud offers affordable, plentiful computing power that can reduce or eliminate the need to invest in data centers. Cloud resources are available on-demand, rising to meet peak surges and falling when demand subsides. Companies don’t have to maintain IT staff trained to manage loads and infrastructure. Cloud services host advanced technologies that start as native-cloud systems, making more applications available. Cloud providers offer robust security and enable regulatory compliance.

In the past, many CIOs took an IT-driven approach to cloud migration that emphasized IT cost reductions rather than building new digital capabilities that would provide superior products and processes. For example, packaged-software vendors drove the migration of certain applications, such as enterprise resource planning (ERP) and customer relationship management (CRM) software suites. Often, simply deploying cloud-based versions of existing systems (a technique known as “lift-and-shift”) was IT’s preferred solution for managing expensive and difficult-to-maintain legacy systems.

But this approach only produces IT cost reductions. It doesn’t enable a company to create strategic digital capabilities that provide competitive advantage in areas such as marketing, sales, service, manufacturing, R&D and procurement through application modernization, digital transformation and data monetization. These capabilities can require compute power that goes beyond what most companies can provide in their data centers.

Many companies attempt to modernize their systems by retaining their legacy mainframe and building an API-based platform on top of it. While they may connect some of their applications to the cloud, they are also accepting the computational limitations of the underlying system. In contrast, one major U.S. healthcare business is updating its sales and marketing applications by replacing its mainframe with a series of Java-based microservice applications in the cloud. This architecture (when completed) will free the company from the constraints of its legacy systems. It will open it to all the cloud has to offer.

A Machine-First Approach to the Cloud Enables Digital Transformation

Companies need a new approach to digital transformation, one that puts cloud computing front and center. We call this a machine-first approach. By “machine first,” we do not mean replacing every possible human worker with artificial intelligence and other technologies that automate labor. Instead, the machine-first approach asks leaders to determine how to use three key digital technologies – cloud computing, artificial intelligence and analytics – in every aspect of their business: in the products and services they offer, how they develop their offerings, how they create supply and demand, how they support their customers, and more.

Using a machine-first approach, a growing number of companies have become leaders in their sectors: digital natives like Amazon in ecommerce and cloud services, Netflix in streaming movies, Uber in taxi service, and Peloton in fitness equipment; and long-established firms, like CVS Health in pharmacies and health clinics, ABN AMRO in financial services and Cummins in truck engines.

The same principle applies to two phases of the cloud migration process. The first is the migration assessment and developing a migration strategy before migration occurs. The second is executing that strategy.

The pre-migration assessment phase. A machine-first approach in this phase supports the company preparing its overall migration roadmap and business case for specific cloud migration choices. A company must evaluate its existing IT landscape, including platforms, business applications and data in its data centers.

A machine-first approach automates this work. Software assesses the IT landscape details, including applications and infrastructure. Based on pre-built mapping rules, it recommends the future architecture and migration roadmap. This work includes:

  • Identifying the best “future landing zone” for a company’s existing technology components (application, infrastructure and data layers).
  • Mapping the existing application technology layers to identify which are best suited to infrastructure-as-a-service (IaaS) and platform-as-a-service (PaaS) – and which providers are the best fit based on the company’s needs and marketplace of cloud vendors.
  • Recommending the best choices for technology architecture on cloud platforms (whether it’s a hybrid, or some other configuration based on the analysis).
  • Recommending the sequence and chronology of moving various applications.
  • Estimating the costs.

Automating these steps has four benefits. It enables a company to rapidly assess different cloud migration paths and jumpstart the migration execution phase. Such meticulous planning can reduce risks and optimize potential cost savings in the migration. It enables a company to align its migration roadmap with its business context, both constraints and opportunities. And it provides information required to strengthen a business case for the migration to justify the investment.

The migration execution phase. A machine-first approach makes it possible to migrate IT systems to the cloud using an automated factory model. This model establishes an automated system for reliably and securely deploying workloads to the appropriate cloud service (IaaS and PaaS) that is like an assembly line. The phase includes:

  • Migrating applications based on the specific requirement of each. Some are “low touch” and require re-hosting without any software modifications. Others are “high touch” and require updating the software to work properly in the cloud host environment (re-platforming), or more involved coding changes to take advantage of cloud systems performance (re-factoring).
  • Managing the migration of workflows and business processes associated with the migration of applications to the cloud.
  • Taking advantage of AI-enabled robotic process automation software scripts to automate the process.
  • An agile approach to execution, using a continuous integration and continuous delivery (CI / CD) tool chain built within each “assembly line.”

At the outset, a machine-first approach enables a company to accelerate data discovery – what data is handled by which applications – and to build the target cloud architecture of the IT landscape including applications, infrastructure and data. This step identifies interdependencies in the IT landscape, and clarifies a sequence for migrating these elements to the cloud. A machine-first approach enables companies to automate the actual migration of their existing technology components, from applications, infrastructure and data layers, to the cloud. This approach has several benefits:

  • It lets them monitor migration progress.
  • It accelerates the migration process, from data discovery to identifying target cloud architectures and a migration roadmap.
  • It enables automated testing and automated release management processes, which reduce human errors that can occur in a more manual process.
  • Automated testing and release management execute tasks that would otherwise require a wide array of technology skills from a large staff performing ongoing technical assessments during a migration to the cloud.

These moves lay the foundation for a company’s cloud-enabled digital transformation.

The Machine-First Approach in Action

The experiences of two large European enterprises, a postal services company and an airline, show the machine-first approach in action.

Cloud migration lays groundwork for innovation. When its cloud migration project began, the postal services company faced a looming deadline: in 11 months, its contract with the firm running its data center was set to expire, putting stable maintenance for more than 500 applications at stake.

The company wanted to move all applications to the public cloud while creating a better IT environment that would give it control, scalability for its applications, and put it in a position to take advantage of future innovations. There was an urgent need to assess and plan for cloud migration to leave time for its move to the cloud.

What happened: The company used a machine-first approach to assess its IT environment, analyzing its applications and data. It mapped its existing applications to appropriate cloud-based applications. The recommendation was to migrate its applications to a hybrid cloud model, in which some are hosted by public cloud services and others reside in private clouds.

The benefits: The assessment was completed in four weeks and included an inventory of the company’s IT landscape, analysis of its data and recommended cloud migration roadmap. In addition, the assessment identified applications with high operational costs that could be migrated first to the cloud for immediate ROI. The recommendations showed that the company could achieve a 30% cost savings within a year of its cloud migration.

Cloud migration a catalyst for digital transformation. The European airline faced the challenge of owning a very complex IT landscape, including more than 1,450 applications and 7,500  servers. The company wanted to use a cloud migration program to start its digital transformation journey, harnessing cloud capabilities such as elastic computing power, faster response times for customer-facing applications and cloud-based applications that promote organizational agility.

What happened: The company used a machine-first approach to assess its existing assets and the optimal migration path. It validated and mapped its data usage to appropriate public cloud services and detailed how each application could be migrated to the cloud, depending on its specific needs (including which ones required reengineering). The recommendation: build the company’s future technology architecture on Microsoft Azure cloud, while retaining some on-premise applications due to legal constraints and other requirements. The assessment included a roadmap for executing the migration and integration with on-premise systems.

The benefits: The company completed its assessment in six weeks and determined that 89% of IT components could be moved to Azure. In total, the assessment showed that 80% of components migrated to cloud would be modernized, giving the company more functionality at lower cost.

Machine-First Approach Realizes Cloud Migration Potential

Cloud migration is critical to digital transformation. But the process takes a great deal of effort to ensure that a company is managing its data properly, accommodating all its business needs while taking advantage of the flexibility, innovation and cost savings that the cloud offers. A machine-first approach enables a company to automate the key step of assessing the company’s current IT environment and mapping its best path forward – all while speeding the migration to the cloud.


About the author(s)

Suranjan Chatterjee

Suranjan Chatterjee is the Global Head for TCS' Cloud Apps, Microservices & API unit. Suranjan is currently responsible for driving business growth for the TCS' Cloud Apps Practice, delivering digital transformation services covering PaaS/API innovations to global customers, and leading the strategic initiatives to develop next-gen application modernization & integration service offerings on Cloud.

Continue your journey


  • Realizing the Automation Potential: A Machine-First Approach to Achieve Cognitive Operations


    Enterprises often focus their digital transformations initiatives on operations that directly touch customers. They build mobile apps for customers to purchase their products and services, or answer their inquiries and complaints through social media. While these efforts are important, we…

  • Why Digital Transformation Depends on a Machine-First Approach to Analytics, and How to Achieve It


    Many companies are not retrieving the value they expect from their digital initiatives and they don’t fully understand why. In general, enterprises have been slow to make progress on their digital transformations and, in particular, their analytics-led efforts. A recent…