Bangalore-based startup ShareChat is a social networking app that enables users to share content in their own language, follow local celebrities, and expand their network. ShareChat was launched in 2015 by Ankush Sachdeva, Bhanu Singh and Farid Ahsan—three graduates from the Indian Institute of Technology (IIT) Kanpur, a prestigious center of science and engineering in India. Today, the app has around six million active users—a figure that is doubling every four months. ShareChat recently announced a US$18 million funding round, which increased its total funding to US$23.35 million.

The Challenge

ShareChat knew that scalability would be key to the success of its app. Bhanu Singh, chief technology officer and one of the founders of ShareChat says, “From the early growth in user numbers, we were sure that we would have to build for scale sooner than expected.” Besides scalability, Singh and his colleagues also saw the importance of reliability and performance to their venture. “As any smartphone user will tell you, if an app stalls regularly, you’re going to delete it,” says Singh. “So we have to deliver a great service to our customers.” Part of that great level of service included minimizing the impact of narrow bandwidths in some parts of India. Many ShareChat customers would be using the app across 2G networks. He says, “To make up for the bandwidth limitations, it was important that we cut network latency for our app as much as possible.”

The startup founders also wanted a solution that would help them reduce IT management overheads, enabling them to focus resources on product development. They decided that only a cloud infrastructure would meet their business requirements and looked for a cloud provider with the services and features to meet their goals. Singh says, “We wanted to integrate data analytics and eventually machine learning into our IT infrastructure to fully understand our customers and respond to their needs in a timely way.”

Why Amazon Web Services

ShareChat engaged with Amazon Web Services (AWS) and an AWS solution architect to build the IT infrastructure to support its app. Singh says, “AWS is a leading cloud-service provider in India that offers the scalability, reliability, and performance that we were looking for. AWS also provides the support we wanted in order to deliver our app to customers across multiple Internet bandwidths.”

To support its content servers, ShareChat uses Amazon Elastic Compute Cloud (Amazon EC2) instances. Because most of its customers access the service via 2G networks, the servers transfer data using MQTT, a machine-to-machine protocol that is designed for networks where bandwidth is at a premium. ShareChat also uses Auto Scaling to scale its Amazon EC2 instances up or down, depending on traffic. The company runs large parts of its code using AWS Lambda (Lambda), a serverless, event-driven computing service. Lambda starts running ShareChat’s code in response to service requests and gives additional scalability.

The profiles and posts of users are stored in Amazon DynamoDB, while Amazon Simple Storage Service (Amazon S3) is used for data archiving. ShareChat also relies on Amazon Redshift as a data warehouse for analysis of usage trends across the ShareChat service. Finally, the company uses Amazon CloudFront to distribute messaging traffic at low latency and with the highest data-transfer speed possible.

The Benefits

By using AWS, ShareChat gained the IT infrastructure its app required to support hundreds of millions of users if needed. “The scalability and reliability of AWS helps ensure that we can deliver a consistently high level of service to a rapidly growing user base,” says Singh. “We’ve grown our userbase ten-fold in the last 12 months without any scaling issues.”

The company’s ability to deliver this high-level service is also due to the low latency of the AWS Cloud and the performance of Amazon DynamoDB. Singh says, “The speed of the AWS network helps us overcome the Internet bandwidth restrictions of many customers. The fast, consistent performance of Amazon DynamoDB is important, too, because we see around 100,000 database requests per second when ShareChat users are active. With the impressive scalability of Amazon DynamoDB and the millisecond latency of the AWS network, we are able to grow at a breakneck speed and serve our tens of millions of users with an app response time of less than a second.”

Singh expects ShareChat to integrate additional AWS services in the future, and is developing use cases for various Amazon Machine Learning services. “Incorporating Amazon Machine Learning will simplify the process of building machine learning models without the need to run complex ML algorithms in-house,” he says. ShareChat is also working with AWS to augment its video transcoding and delivery pipeline because more users are sharing and watching videos on their social network.

Finally, because the AWS infrastructure doesn’t require any dedicated IT management, ShareChat can focus its resources on product development. “We want our personnel to dedicate as much time as possible to enhancing our product and delivering more value to our customers,” says Singh. “Where AWS is really important to us is that it has eliminated the headache of infrastructure management. We can now stay concentrated on innovation.”