Scaled backend 10x with GKE Autopilot in a 14-day sprint
Maintained 100% uptime with Apigee API Management
Spanner was a key component in the payment and subscription architecture, supporting reliable and consistent payment data operations
Zee 5 scales its backend 10x on Google Cloud to stream the FIFA World Cup in 14 days, driving a reliable subscriber experience.
Zee 5 scales its backend 10x on Google Cloud to stream the FIFA World Cup in 14 days, driving a reliable subscriber experience.
Conquering a 10- to-14-day sprint to scale live streaming experience for paid subscribers
When Zee Entertainment Enterprises Limited acquired media rights for the FIFA World Cup, the most watched sporting event in the world, its digital streaming platform, Zee 5, faced an extraordinary engineering mandate: prepare its entire backend ecosystem to handle a 10x traffic surge in just 10 to 14 days. To meet this aggressive timeline, Zee 5 reached out to its cloud partner, Google Cloud, to support the infrastructure scaling required for the tournament.
Live sports broadcasting tests every layer of a digital architecture. Unlike free ad-supported streaming events, Zee 5 hosted the tournament behind a subscription paywall (SVOD). Minutes before kick-off, millions of fans would land simultaneously on the app, loading content rails and thumbnails, logging in, buying plans, and initiating playback. A delay of just a few seconds could impact subscriber acquisition and user experience.
The Zee 5 technology team stress-tested and hardened the architecture and optimized performance across 100+ microservices, event streaming systems, databases, APIs, and critical user journeys, such as login, subscription payments, and video playback. In parallel, Google Cloud helped scale the infrastructure and address key capacity requirements across GKE Autopilot, Apigee API Management, and the service mesh. Together, this preparation enabled Zee 5 to handle extreme traffic spikes and deliver a smooth, uninterrupted viewing experience for millions of fans.
Engineering microservices, load testing, SOS switches, and joint war room operations
Realistic traffic was generated to validate the platform against the target TPS/QPS. Controlled load testing was conducted in the production environment during low-peak hours. This helped save time, identify real-world bottlenecks, and uncover issues that could not have been fully replicated in non-production environments.
The team built strategic SOS switches across the frontend and backend, so that recovery could happen in seconds, not minutes. Pre-defined and planned SOS switches were designed to temporarily disable specific features—such as Recommendations, Continue Watching, and Watch List—during peak events to protect the overall viewing experience.
Likewise, on-demand and panic/crisis SOS switches were introduced to disable specific functionality when required, while keeping core viewing uninterrupted. For example, ABR capping on mobile devices could be enabled to manage bandwidth congestion and maintain overall streaming stability. Fortunately, this switch was not required during the season.
While SOS switches provided manual controls at the application layer, protecting the platform also required automated traffic management at the API gateway. Apigee Spike Arrest was used to control traffic on selected APIs during extreme spikes, protecting critical user journeys and overall platform stability. Google Cloud engineering teams helped address this on priority, enabling Spike Arrest to support higher QPS. This allowed controlled load shedding when required, ensuring that a surge on selected APIs did not impact the overall platform.
And since autoscaling alone is not enough for live events, where traffic can increase sharply within minutes, critical GKE Autopilot services were pre-warmed before kick-off by scaling up the required pods in advance, ensuring capacity was available before the surge occurred. After the match, the infrastructure was scaled down based on demand to optimize resource utilization.
GKE Autopilot, essentially, enabled the team to independently scale critical high-demand microservices like video playback without over-provisioning secondary services, controlling overall infrastructure costs.
The biggest optimizations often come not from code changes, but from understanding the system and making the right architectural decisions.
Throughout every match, Google Cloud engineers operated on-site in a joint ‘war room’ at the Zee 5 premises, backed by global Google Cloud support and product teams conducting real-time monitoring and rapid issue resolution to safeguard the fan experience.
Unlocking next-gen content discovery and developer speed with Gemini
Following the successful delivery of the FIFA World Cup streaming experience, Zee 5 expanded its strategic collaboration with Google Cloud on various transformative AI solutions and capabilities to drive increased developer productivity, long-term viewer engagement and enhanced user experience.
To maintain high operational velocity, Zee 5 developers actively leverage Gemini Code Assist across daily workflows for microservice refactoring, code review, unit testing, and automated code generation.
In collaboration with Google Cloud, Zee 5 launched the “Clips” feature on its Zee 5 app. Clips is built on an automated vertical video generation pipeline driven by Gemini. The system automatically identifies key highlight scenes within horizontal movies and shows, extracts clips, and formats them into engaging vertical short feeds on the Zee 5 app. This enables users to discover new content interactively from Zee 5’s rich content library, engage with full-length titles using direct calls-to-action, and helps Zee 5 create new vertical ad inventory for revenue growth.
Simultaneously, Zee 5 is deploying Gemini to generate granular, scene-level metadata across its library. This foundation drives an AI conversational search feature, allowing viewers to execute natural language queries across multiple Indian languages, such as searching for specific actors in action sequences, recommendations based on intent, while helping internal editorial teams quickly locate specific scenes for previews and trailers.
By pairing resilient cloud infrastructure with ongoing generative AI innovation, Zee 5 and Google Cloud are setting a new operational benchmark for digital broadcasting, transforming the platform into an interactive, AI-driven, next-generation entertainment hub on Google Cloud for millions of its viewers. These joint AI innovations accelerate time-to-market for new platform features while continuously elevating the daily viewer experience.
Zee 5 is a leading video-on-demand platform operated by Zee Entertainment Enterprises Limited, offering thousands of hours of multilingual movies, series, news, TV shows, and live TV and sports to millions of viewers.
Industry: Media and Entertainment
Location: India
Products: GKE Autopilot, Apigee API Management, Spanner, Google Cloud Armor, Security Command Centre, Gemini, Gemini Code Assist
