MongoDB Atlas - eHarmony's New Message Store

PublishedJune 19, 2017

Speaker: Elan Srinivasan + Suneeta Puranik
Communication is an integral part of the dating process. The more users communicate on our platform, the higher the chances they will find someone they like.

By early August 2016 reports indicated a steep decline in communication rates. The multi-step, turn by turn communication process was no longer enticing to the users in the current world where real time messaging is a norm. The existing store and design was not conducive to moving in this direction given its rigid nature and slow response times. Change was imminent.

User communication being a high traffic feature, the store was the crux. The data store had to be robust, scalable and flexible enough to support current and future needs. Store performance could not be a bottleneck and its failure was not an option. Our team was lean and had no experience with MongoDB, let alone setting up a cluster and managing one. Also, given this was a new system, traffic patterns and capacity could not be predicted.

In this presentation, we will cover how eHarmony was able to leverage MongoDB Atlas to overcome the above challenges to build a highly successful real time communication system for its users.