E-commerce giant simplifies its architecture and accelerates growth
About Flipkart
Founded in 2007, Flipkart has grown into one of the largest e-commerce platforms in the world. It offers products across electronics, fashion, home essentials, and groceries to over 200 million users and is known for its Big Billion Days sale and services like Flipkart Plus and SuperCoins. In 2018, Walmart acquired a 77% stake in the company for $16 billion, one of the largest e-commerce acquisitions in history. With a strong focus on technology and logistics, Flipkart consistently innovates to enhance its customer experience.
Inefficient virtual machine-based infrastructure hindered performance across 40+ Aerospike deployments
Flipkart’s extensive operations demanded a scalable and efficient database infrastructure to handle diverse workloads, including caching, data storage, and user applications like search, product pages, and recommendations.
Initially, Flipkart managed over 40 Community Edition Aerospike deployments using virtual machines (VMs), each managed by individual teams. This decentralized approach led to several challenges:
Scattered deployments
VM-based clusters were difficult to standardize and optimize, leading to inconsistent performance and high operational costs.
Resource inefficiency
Virtualization overhead and the challenge of optimizing VMs for diverse workloads resulted in wasted resources and inefficient scaling.
High operational complexity
Managing VMs required significant manual effort, especially for routine tasks like hot edits and hardware management, making it difficult to maintain performance consistency across clusters.
Centralized, automated, and scalable database management with Kubernetes and Aerospike
Flipkart created a centralized team to manage Aerospike as a shared service across the company, consolidating all Aerospike deployments under one umbrella. This transition, along with an upgrade to Aerospike Enterprise Edition and the implementation of Kubernetes, allowed Flipkart to efficiently manage over 170 clusters, automating tasks and optimizing performance at scale.
The Aerospike Kubernetes Operator (AKO) played a crucial role in scaling and automating database operations, enabling seamless management across environments.
Unified management
Centralized Aerospike operations as a shared service, standardizing deployments and improving resource efficiency across multiple teams.
Efficient scaling
KO enabled effortless horizontal and vertical scaling, ensuring infrastructure could handle growing demand with minimal manual intervention.
Automation and self-service
Automated deployment, scaling, and monitoring processes reduced operational overhead and allowed teams to focus on core tasks.
Optimized performance
Running Aerospike on bare metal through Kubernetes improved resource utilization, bypassing virtualization for better overall performance.
Streamlined operations, enhanced scalability, and cost savings
With Aerospike and Kubernetes, Flipkart significantly improved its data management infrastructure, enhancing performance and delivering a seamless customer experience.
Boosted performance
Achieved 95 million transactions per second (TPS) during its Big Billion Days sale, one of India’s largest retail events.
Increased scalability
With AKO, Flipkart’s infrastructure scaled automatically and efficiently, ensuring 100% availability and uninterrupted service.
Cost savings
AKO’s automation reduced operational overhead and infrastructure costs by minimizing the need for overprovisioning.
Testimonials
Since the time we went live with the Kubernetes infra, we have seen 100% availability with the operator. There hasn't been any incident where we had a production issue because of the operator. It has been working very well for us.
Sharath B.P.
FlipkartSoftware ArchitectTech standardization was a charter for the (shared services effort, which was) not just limited to Aerospike. There were multiple databases, and Aerospike was clearly the winner.
Sandesh Kumar Gupta
FlipkartSoftware Development EngineerAdditional resources
For a deeper understanding and more insights, explore these additional resources.
Get started with Aerospike
For high-performance, scalable data management and ultra-low latency, ideal for handling massive datasets and real-time applications.
