Overcoming process impediments in cloud migration

Large global enterprises are in various stages of cloud adoption, to propel their digital transformation and gain competitive advantages from cloud. Yet the medium enterprises have consistently demonstrated higher migration velocity to cloud compared to larger enterprises including the global 2000.  Though larger enterprises have access to the required technical resources to migrate at scale, they encounter challenges navigating enterprise processes and internal operational structures.

Enterprises have complex IT delivery process to operate at scale to ensure consistent software lifecycle processes and meet compliance requirements laid out by the enterprise architecture teams, regulators like APRA, EBA, etc., industry standards like HIPAA, SOX, etc., geo/ country specific standards like GDPR, PDPA, etc. This multiple towers IT organization structure with fine-grained and well-defined boundaries is required for bringing efficiency in the “run the organization” side of the business and is acceptable for slow changes. However, this becomes an impediment for large scale “change the organization” initiatives like cloud migration. While the improvements across people, process and technology are required to put the organization on an accelerated path, we shall examine the process challenges in this blog, and address the nuances of organization structure in a separate blog. .

Enterprises typically have a multi-stage gating process for IT project implementation that start from ‘business need definition’, business case, implementation roadmap, architecture, cloud exit plan, high-level design, security, risk and compliance assessments, network assessments, testing, build, migration, cutover, operational readiness, etc. In any large organization, this would require 20-50 documents per application across the software lifecycle to ensure approvals from key controllers and stakeholders. This methodology in the enterprises is rigid and not scalable for cloud migration which typically have more than a 100  applications to be migrated in a given quarter.

Why is the process complex now?

Considering the age of the applications in the enterprise, the drift from the initial design and enterprise standards is high due changes over the past several years for the applications. Unfortunately the changes are not documented adequately. This leads to a major effort in fresh documentation for the cloud migration rather than  a straightforward update.

The software development process was defined for the slow-changing systems and suited for ‘waterfall’ models for the IT delivery where the business requirements are well-defined and technology stack finalized early in the project lifecycle. However, today the business requirements are continuously evolving to meet the changing market needs and technology advances are much faster in the cloud. This leads to multiple iterations between the migration team and respective approving authorities in areas like security, network, business continuity, compliance, cost allocation, testing, etc.

What can be done to overcome the challenges?

Cloud migration, irrespective of disposition like simple rehost (lift & shift) or remediate (or replatform) to address technology debt or incompatibilities, are being executed in factory model to get the desired velocity, aided with the right migration tools. Yet, impediments with process remains.

We observe that adopting architecture standardization and automation can help overcome the process complexity and address the concerns of all the stakeholders in an effective manner.

  1. Standardization – Define architecture patterns and their deployment architecture on cloud for each of the application stereotypes. Map on-premises applications to one of the defined patterns when migrating to cloud. This simplifies the major decision making in cloud migration like cloud disposition of applications and eases the documentation burden, as the applications are mapped to the defined patterns. Working with the defined patterns gives confidence to various stakeholders and reviewers like architecture review board, security team, business continuity teams that the applications in target state will meet the performance objectives, remain compliant to organization standards. Design approvals are needed only when there are deviations to address an unique requirements of the applications, which cannot be met with the defined patterns.
  2. Automation – Adopting ‘As-a-code’ for infrastructure, security, policy and even documentation, ensures accurate implementation of the patterns with no deviation, backed up a robust mechanism to identify deviations (if any) and to take corrective actions. This reduces reviews at every stage in the development process and reduces the testing effort as well.

Combining the ‘standardization’ and ‘automation’ gives tremendous speed and agility, which will accelerate the migration by eliminating the process steps and/ or reduce the time taken to execute the remaining process steps. Take a scenario like network design with firewall ports opening, under the new approach, when the application is mapped to defined application patterns, architecture and security review may not be required under most circumstances. As a result, the automation to create all the cloud resources as per the pattern, lengthy and high touch activities like firewall port are executed in minutes.

How to bring in the change?

While architecture and technology are at the core to bring in acceleration in the migration, few process and organization changes would be ideal.

  • Establish a solution factory to define the standard cloud architecture patterns for the application stereotypes in the enterprises. Restrict the application stereotypes within manageable range like 15-25, as standardization would simplify migration and reduce cost of management of the target state.
  • Create a cloud platform engineering squad, to create the automation for cloud resource provisioning by entirely adopting ‘as-a-code’ model and adopting DevOps to the extent possible.
  • Critically evaluate the existing SDLC processes and improvise the process for high velocity cloud migration by eliminating/combining the steps due to standardization and implementation of automation.
  • Establish a cross functional technical governance team which convenes at least once a fortnight to remove any impediments.
  • Foster the culture of collaboration across units by establishing ‘community of practices’ for common interest topics with required executive support.

In conclusion, the process impediments can be overcome with technology, architecture, tools and right team structure along with suitable adjustments to the process to achieve the larger objective of faster cloud, thus enabling the IT organization to be the go-to partner for businesses in technology led innovation and growth.

Author Details

Madhan Raj Jeyapragasam

Cloud Solution Strategist in Enterprise Cloud Ecosystem at Infosys and helps enterprises transform with cloud to deliver on their business outcomes. Earlier as an IT strategist and enterprise architect, he has designed business-aligned IT solutions for large and complex business systems for Fortune 1000 companies.

Leave a Comment

Your email address will not be published.