Hadi is a mobile live video stream trivia app that took Turkey by storm and rapidly became one of the hottest app store downloads. The mobile app delivers the traditional TV trivia competitions to the screens of millions of smart phones through interactive video. Every day, users log on to compete for cash prizes by responding correctly to 12 questions within seconds.
The service was launched in February 2018 and grew rapidly. Within weeks after the launch, the number of users grew to 60,000 and reached the peak of 1.2 million live users on New Year’s Eve in 2018. Hadi community comes together for daily contests for an opportunity to win cash prizes that are distributed among all users who correctly responded to the entire series of questions.
Processing Millions of Answers in Real-Time
Success of the app on iOS and Android means that Hadi was under pressure to ensure a smooth experience to the millions of users who respond to the quiz in real-time. Here are some peak moments that give a sense of the scale of the app: 10 million push notifications are delivered within 3 minutes and 700,000 answers are processed in under 400 milliseconds. Overall, the app has had to process 1.5 billion answers.
To deliver this level of performance, Hadi partnered with Lena Software, a global solution provider with extensive expertise in developing high-performance and scalable real-time video streaming mobile technologies. “Hadi came to us with a challenging project that required cutting-edge infrastructure and extensive expertise,” said Serkan Köse, Managing Partner at Lena Software. “We were excited to have the opportunity to come up with an architecture that displays our expertise in real-time video streaming, data processing, and mobile application development.”
Lena Software built the entire backend architecture for real-time interactive data processing and streaming video delivery. The team focused on responsiveness as the solution needs to process millions of responses submitted by users in real-time to ensure a seamless experience with low latency. The daily competitions are scheduled in advance, which means that the infrastructure needs to accommodate regular daily peaks in traffic. While other apps may have more users and more traffic overall, Hadi needs to respond to rapidly occurring extreme peaks as thousands of users log on and take part in the quiz.
Focus on CPU Performance and Scalability
The required performance was achieved by Lena Software’s team that used Node.js, Redis, elasticsearch, and .NET core technologies to develop the solution. The team used Node.js to improve efficiency and then coupled that technology with a data serializer model they built from ground up for this project. This combination delivers faster in-CPU data processing. “Node.js turned out to be the best fit for the project’s technology needs,” explained Serkan Köse. “At the same time, a proprietary data serializer and CPU optimizations allowed us to focus on CPU capacity and deliver the performance required to process hundreds of thousands of requests and responses sent by Hadi users.”
The required scalability is delivered by a hybrid approach to infrastructure that mixes on-premise and cloud infrastructure to deliver the performance needed during peak times. This allows Hadi to stay responsive in the face of sudden surges in the number of requests and responses.
Hadi now operates on a reliable and high-performance infrastructure that delivers a fun and exciting service to millions of users and can be easily scaled to support the company’s global expansion plans. Lena Software, on the other hand, turned the know-how and experience gained with its data processing product – Maestro into Hadi’s main structure.
