More than 5,000 microservices and 30 million workloads migrated within three months with no customer-facing downtime: targeting 30% run-cost reduction and improved scalability.
A leading Indian value-commerce platform, in its hyper-growth phase, had to address an issue that was becoming harder to ignore: the cloud operating model, built on Amazon Web Services (AWS), was becoming increasingly expensive and complex to run. The e-commerce leader struggled to maintain unit economics as it handled more transactions, users, sellers, and orders.
The Google Cloud Platform (GCP) emerged as a viable option that supports operating margins. However, its AWS cloud estate was highly distributed: 5,000+ microservices, multiple data stores and variable demand that could shift sharply during sales. A successful migration would ensure no disruption during sale windows, predictable release and change management, and a measurable path to lower run-rate costs. Just as important, security and observability had to improve during the change, not after it.
The e-commerce leader commissioned Persistent, a Premier Google Cloud partner with an engineering-first mindset, to plan and execute the AWS-to-GCP migration under tight timeline and risk constraints.
Persistent’s Business-First Cutover Plan
Persistent approached the migration as a continuity program that keeps the business running while changing the underlying platform. We worked in a phased waves approach, mapping service dependencies, grouping workloads and defining the “go/no-go” criteria for each wave. Our battle-tested runbooks, rehearsals and coordinated change windows aligned with business-critical events to ensure cutover readiness, so the client could move quickly without relying on last-minute heroics.
We set up a dedicated team GCP engineers to worl alongside application, security and operations owners. We prioritized governance, focused on risk and decision speed by establishing a single migration backlog, a shared risk register and a regular stakeholder cadence for trade-offs and approvals. During key moves, a centralized “war room” model coordinated execution, incident response and business communications to ensure accountability and fast resolution.
We started the migration by setting up a solid foundation for economics and reliability – a landing zone. Here we brought in organizational nodes, shared VPCs, identity and access management, backup patterns, firewalling and migration tooling, along with baseline monitoring and security posture management. With those controls in place, we migrated workloads wave by wave over three months, moving 30 million ARR workloads and enabling containerized operations on GCP. The cutover was completed with no customer-facing downtime reported.
Data and streaming workloads were first-class dependencies. To maintain continuity for data engineering, streaming and data science, Persistent implemented Databricks and Confluent Cloud Kafka on GCP and transferred more than 100 PB of data with validation steps to confirm completeness and performance. Persistent also consolidated and migrated multiple database technologies (including MySQL, PostgreSQL, Redis, HBase, MongoDB and DynamoDB) into GCP-managed services to simplify operations and reduce variation across teams.
Lower Costs, Higher Agility, Stronger Controls
The GCP environment today supports approximately 30% lower operational cost, while improving scalability for a broad base of users and services. Equally important, a standardized operating baseline with identity, network controls, monitoring and security posture visibility reduces the cost of operating at scale and supports faster, safer change during peak events. Access to GCP data and AI/ML capabilities also supports analytics-driven use cases such as recommendations, customer profiling, seller verification and platform performance monitoring.
For the client, this migration was not only a change of provider; it was a reset of how the cloud is operated. Persistent set up the path for more standardization, clearer controls and a repeatable way to execute future changes with less risk. This is what enabled the team to migrate at scale within a compressed timeline while protecting customer experience.
Thank you, team, and sincere gratitude for us coming together. I’m grateful for everyone’s focus on ensuring the client had a successful weekend sale and for responding to the challenges.
Moe Abdula, VP, Customer Engineering, Google, APAC
Google Cloud recognized Persistent as the APAC Partner of the Year for Infrastructure Modernization at Google Cloud Next 2025, a testimony to Persistent’s delivery capability for complex modernization programs that combine platform foundation, security posture and large-scale workload migration under time and availability constraints.
