How Singlife used AWS Professional Services, AWS Managed Services and AWS Migration Programs to build and accomplish their Cloud Migration Plan

Singlife with Aviva or Singlife, is a Singaporean insurance company that announced in December 2022 they are going all in on AWS and will migrate their entire IT infrastructure to the cloud by end 2023. The depth and breadth of AWS services will enable Singlife to improve speed-to-market of new products and enhance the company’s digital workflows, powering all of its new user interfaces and elevating consumer experiences for its more than 1.5 million customers.

Three foundational pillars for successful migrations

A cloud migration initiative is often one of the key moments for many organizations to take an in-depth look at their on-premises environment. What they typically find is a complex web of applications with data interdependencies and licensing agreements that is very hard to fully understand. This makes every cloud journey uniquely challenging with many unknowns, assumptions that need to be proven and difficult decisions that need to be made. To help customers deal with the complexity and accelerate their cloud migration, AWS has developed a migration methodology based on their experience migrating thousands of enterprise customers to the AWS Cloud.

Figure 1. AWS Migration Acceleration Program.

The AWS Migration Acceleration Program (MAP) uses a proven three-phased framework (Assess, Mobilize, and Migrate and Modernize) to help customers build strong AWS Cloud foundations so they can accelerate their migration whilst reducing risk. MAP identifies 16 pillars that should be in place for a migration journey to be successful, with 3 pillars in particular that exist in every organization that has successfully delivered a cloud migration project.

1. Executive Sponsors
2. Cloud Operating Model and Skills Enablement
3. Migration Plan

Having Business, IT & Security Executive Sponsors for the cloud journey expedites the removal of blockers that inevitably arise and ensures the organization’s focus on migration goals. According to McKinsey, organizations with active CEO sponsors are 32 percent more likely than others to migrate on time and on budget.

High performing organizations also educate their employees from top to bottom on the cloud and build a Cloud Centre of Excellence (CCoE) early on in their migration journey. They are also 57% more likely to hire for advanced skill sets such as DevOps and FinOps.

The same study found that high performing organizations are 9% more likely to develop the full implementation road map, including the security and compliance framework, up front rather than funding a series of one-off initiatives. This post will provide insights into how Singlife built and accomplished their migration roadmap by working backwards from their business and technical drivers, and how AWS Professional Services (ProServe) leaned in with a comprehensive set of migration tools, best practices and service offerings.

Background to Singlife’s migration planning

It’s important to understand that migration planning is an iterative process that starts in the Assess phase and continues into the Migrate & Modernize phase with detailed wave plans. In the Assess phase, you build your directional business case, assess your organization’s readiness for cloud transformation and use the findings to build a high-level strategic roadmap. You also take an initial look at your portfolio, and confirm that the key stakeholders in your organization are aligned. For more information about the assess phase, see Evaluating migration readiness on the AWS Prescriptive Guidance website.

The migration plan and business case will become more detailed as the journey progresses and new information is uncovered, so it’s important to avoid overanalyzing at the start as this can stall the migration before it’s even begun. In the case of Singlife, they needed to exit 2 on-premises data centers before their contract renewal dates which dictated their timelines and business case for cloud. Working backwards from these fixed dates, they created a 3-year roadmap to transform to a cloud IT operating model, build cloud fluency at all levels of the organization and gain a deep understanding of their IT estate to plan and accomplish the migration to cloud. At the same time as IT was getting started on their cloud migration journey, the business was embarking on a merger that would eventually create the Singlife brand. The inherent challenges with mergers had the potential to derail the journey to cloud but Singlife built mitigation measures into the plan, as well as maintaining internal alignment and strong executive sponsorship to overcome the inevitable challenges that arise on multi-year large migrations.

The first foundational pillar – Executive Sponsorship

After building the directional business case for a cloud transformation and securing internal alignment to proceed, the next phase is mobilizing the organization to prepare for the migration. This is the Mobilize Phase and in this phase you typically build your AWS landing zone, conduct a more thorough portfolio assessment, build your security and operating model, and prepare teams for change. The Singlife migration team were concerned that the merger would delay the mobilize phase as internal teams would not be able to focus on key enablement activities like cloud training and operating model redesign.

Also, as they looked deeper into the portfolio, they realized that migrating away from some of their legacy hardware investments was going to be difficult – their monolithic core insurance system was running on AS/400’s which would have to be refactored for cloud and 80% of their application portfolio was dependent on a commercial database engine hosted on an on-premises appliance that was optimized for running databases. Whilst this database platform provided some of the highest levels of performance, it presented lock-in challenges as the license agreement did not include the right to host in a public cloud environment. This feature rich database platform also created complacency within the application teams who were using the full feature edition even though it was not required in some cases. As their migration timeline was fixed on the DC exit dates, Singlife had decided that only rehost and replatform migration strategies would be allowed with minimal re-architecting so they could migrate as fast as possible. If product teams wanted to refactor their applications, they would have to do so once they were on the AWS Cloud.

Rehost strategy is also known as lift and shift and it’s when you move your applications from your source environment to the AWS Cloud without making any changes to the application. Replatform strategy is also known as lift, tinker, and shift or lift and reshape. Using this migration strategy, you move the application to the cloud, and you introduce some level of optimization in order to operate the application efficiently, to reduce costs, or to take advantage of cloud capabilities.

However, migrating the commercial databases off the appliance using these migration strategies created additional challenges for Singlife, as they would have to procure new licenses with public cloud mobility rights in order to use the full feature edition on the AWS cloud. This unbudgeted procurement would require additional approvals that could have significantly delayed the migration.

When starting a large migration, it’s important to identify a single threaded technical leader who is 100 percent dedicated to the project and accountable. That leader is empowered to make decisions, help avoid silos, and streamline work-streams by maintaining consistent priorities. This is exactly what Singlife did and their migration program lead was able to create alignment between the various executive sponsors so they could overcome these challenges by removing roadblocks and staying focused on the migration success criteria.

The second foundational pillar – Cloud Operating Model and Skills Enablement

Figure 2. Transforming to the Cloud Operating Model.

During the readiness assessment in the Assess phase, Singlife identified the importance of transforming to a Cloud Operating Model, so one of the first mobilize activities on their roadmap was to establish a Cloud Centre of Excellence (CCoE) team. This team is foundational for a successful transformation as they start by providing knowledge and guidance on how to operate in the cloud but quickly become the driver of change across the enterprise.  Singlife staffed the team with tenured employees who had the right mix of technical and change management skills to lead the migration and prepare the organization for the new operating model. Realizing how critical this team would be to their success, Singlife engaged AWS Professional Services (ProServe) to accelerate the onboarding process and establish the most optimal team structure.

The CCoE offering from ProServe resulted in a cross-functional team of cloud experts that were focused on constructing the building blocks for the cloud migration, as well as developing their leadership skills so they could successfully own and execute the evolving migration plan. This new team set about enhancing the cloud strategy to the next level of detail, including the polices and standards to govern the organization’s usage of cloud, and how the AWS environment should be configured and managed on an ongoing basis.

The CCoE team knew they had to move fast which would be challenging in a newly merged organization that still needed time to stabilize, so they made a decision to use AWS Managed Services (AMS) to setup and operate a well-architected, multi-account AWS environment (Landing Zone). By using AMS, they were able to accelerate their Cloud Operating Model and quickly achieve operational excellence with advanced cloud operations meeting their policies and standards for monitoring, incident detection and management, security, patch, backup, and cost optimization. This allowed them to offload the undifferentiated but critical task of infrastructure management, so they could keep their people focused on strategic activities like skills development and migration planning.

The CCoE team made another important decision to accelerate their cloud operating model by mandating the use of Amazon Relational Database Service (Amazon RDS) for all their database workloads. Amazon RDS is a managed service that handles routine database tasks, such as provisioning, patching, backup, recovery, failure detection, and repair. By offloading the time-consuming database administration tasks, Singlife was able to free up their resources to focus on applications and the migration to cloud.

By accelerating the foundational layer of their cloud operating model, Singlife could turn their attention to enabling the rest of the organization with the essential cloud skills they would need to be successful when operating in the cloud. With the help of AWS Training & Certification team, Singlife, supported by their newly formed CCoE, crafted a bespoke training program to help the workforce upskill using AWS Jam sessions, AWS Builder Labs and AWS Cloud Quest. The organization was able to experience cloud learning in new ways and rapidly test and advance their knowledge with AWS Skill Builder.

The third foundational pillar – Detailed discovery for business case and migration wave plan

During the Assess Phase, Singlife had built a total cost of ownership (TCO) comparison of the on-premises costs and the projected run costs on cloud. This enabled them to build the business case for cloud and secure the executive sponsorship for the strategic roadmap. As is typical with these early stage TCO activities, infrastructure metadata from the data center inventory tools was used as a baseline and any unknowns were replaced with assumptions or benchmarks. This infrastructure centric approach can only produce high level cost models as the application portfolio assessment activity in the Mobilize phase is when the application scope and migration strategies start to take shape. Whilst this rapid assessment is essential to align internal stakeholders on the plan and build early momentum, it doesn’t go deep enough to produce an accurate budget for the migration execution. To do that, a detailed planning activity had to be undertaken so that the Singlife board could sign off on the migration plan and the associated costs for labor, platform and licensing.

To help accelerate this planning activity, the CCoE team again turned to ProServe to leverage their Application Portfolio Assessment offering. This offering builds a migration plan and business case by analyzing the in-scope application portfolio. It starts with a detailed discovery exercise employing specialist tools and workshops to establish a baseline for the application portfolio and underlying infrastructure. To manage the discovered metadata, ProServe established a single source of truth for migration data, which is an important best practice to keep the various teams aligned and enable data-driven decisions as the migration progresses. They selected Device42 as it could automate the infrastructure discovery and act as an extensible Configuration Management Database (CMDB) for all the migration metadata. The tool was a good fit for Singlife’s use case as it was CISO and regulator friendly (agentless, on-premise solution using hardened appliances) and also able to discover the critical workloads running on non-x86 systems, like the core insurance system on AS/400 and the commercial databases running on appliances. The tool was run for the duration of the migration journey to collect server metadata for planning and design purposes, such as resource utilization metrics (CPU, RAM, Disk etc.), installed software/databases per host and connections between hosts. ProServe used the discovered data during their workshops with the application teams and enriched it with non-discoverable metadata, such as application criticality, migration complexity, roadmaps, licensing agreements, HA/DR requirements and regulatory requirements to name just a few. This enabled them to build a complete picture of the application stacks and their inter-dependencies which is a key input to the planning process.

The collected data was uploaded to the Migration Portfolio Assessment (MPA) tool, offered by AWS, and available to both customers and partners. This tool analyzes a customer’s on-premise portfolio, generates a detailed business case for the migration and creates a migration wave plan based on application priority, dependencies and migration strategies. Wave planning uses the connectivity data collected by Device42 to group tightly coupled applications into dependency groups so they are migrated in the same waves, ensuring performance levels are maintained after cutover to the cloud.

The wave plan drives the execution of the migration in the Migrate & Modernize Phase, and it follows a migration factory approach based on the Agile methodology where you migrate your servers at scale in batches known as waves. A migration strategy is the approach used to migrate workloads to AWS Cloud. You would typically group applications with the same migration strategy in the same wave. In the case of Singlife, the CCoE had already limited the allowed strategies to Rehost and Replatform, therefore, simplifying this part of the planning process.

Optimizing commercial databases

The MPA tool helped ProServe estimate the level of effort and costs associated with migrating the application portfolio, as well as the estimated annual run rates for compute and storage resources on AWS post migration. With this, the CCoE team had almost everything they needed for budgetary board approval but they were still missing a deeper understanding of the costs associated with migrating their commercial database estate due to the licensing challenges. To help accelerate this activity, ProServe engaged Optima, an AWS specialist licensing partner to run an AWS Optimization and Licensing Assessment (AWS OLA), which is a free program for new and existing customers to assess and optimize commercial database estates.

Optima used their deep commercial database licensing expertise to identify the most optimal way to migrate to AWS by reviewing infrastructure performance data collected by Device42, as well as Singlife’s on-premises licensing agreements. Optima presented multiple deployment options for Singlife to consider, along with the cost models based on Singlife’s effective license position. However, the options were cost prohibitive as Singlife’s licensing agreement did not allow them to take their licenses to the AWS Cloud.

Optima and the AWS database specialist team stepped in to help Singlife create an optimized migration and licensing strategy that was not dependent on their existing license entitlements. They recommended an approach that stayed true to the fast migration goal, whilst introducing some low complexity changes to overcome this roadblock. They achieved this by assessing all the commercial databases from two different lenses. 1. They assessed the use of features to see which databases could be downgraded to a lower cost commercial database edition during the migration. 2. They used complimentary AWS database experts from the Amazon Database Migration Accelerator (DMA) team to analyze database schemas and application architectures to see which databases could be easily converted to open-source databases during the migration. Singlife made the decision that if both options were feasible, then it was preferable to convert to open source if the complexity was low enough, therefore avoiding significant unbudgeted costs and removing the roadblock.

The wave plan

With the database assessment complete, the CCoE finalized their business case and secured investment from the board of directors to migrate their application portfolio to AWS. They selected ProServe as their primary delivery partner for the Migrate and Modernize phase with an AWS Migration Competency Partner subcontracting under ProServe. This decision allowed the CCoE team to retain the portfolio knowledge that ProServe had acquired in the Mobilize phase, whilst leveraging the ProServe best practice agile delivery framework for large migrations. By using an AWS Partner from the AWS Partner Network (APN) they were able to scale the migration with multiple migration squads (Dojos) that could execute wave plan user stories in parallel delivery sprints. Thus, ensuring the migration would complete on time and on budget before the data center exit deadline.

With ProServe’s extensive experience helping enterprise customers migrate to the cloud, they knew they would need to keep iterating on the migration wave plan to accommodate shifting business priorities and availability of key resources in the application teams. To help them re-plan they first needed to baseline the sprint plan for each type of migration pattern. They did this by executing pilot migrations at the end of the Mobilize phase so they could apply the learnings to create repeatable runbooks and sequence the user stories into sprints. The pilots also provided the Singlife teams with ‘experience by doing’ which is an essential part of the preparation for the Migrate and Modernize phase and is a key tenet of the AWS Experience-Based Acceleration (EBA) methodology used by AWS teams to accelerate large migrations.

At the time of writing Singlife has successfully migrated 75% of their wave 1 applications on time and on budget with all the commercial databases either downgraded to lower cost commercial editions or converted to open-source databases on Amazon RDS.

Conclusion

In this post, you have gained insights into how Singlife built and accomplished their migration roadmap by working backwards from their business and technical drivers. This journey was guided by ProServe’s comprehensive set of migration tools, best practices and service offerings.

Contact us for any migration support you require in your cloud journey today.

About the authors: