How Zoom Used DynamoDB To Grow From 10 Million To 300 Million Daily Users

In early 2020, Zoom encountered massive traffic spikes in user requests, due to the Covid-19 pandemic.

As users stayed home, they needed a way to continue meeting with colleagues and friends — and Zoom was the most popular choice for video conferencing.

At the time, Zoom was supporting a relatively meager 10 million daily meetings and in a matter of weeks, that number shot up to 300 million daily meetings. [1]

These spikes in traffic was challenging for Zoom and so they turned to DynamoDB to support their rapidly rising user base.

The Challenge

Zoom’s sudden growth in users required a database solution that could:

• Scale virtually infinitely to accommodate the 30x increase in user traffic.
• Maintain consistent performance under a high number of concurrent requests.
• Adapt to unpredictable traffic and user behavior.
• Minimize operational overhead during this growth period.

The Scalable Solution

Amazon DynamoDB proved to be the ideal solution that could answer all of these challenges and needs.

DynamoDB was able to offer Zoom the following benefits.

Scalability

DynamoDB allowed Zoom to scale its database capacity with the click of a button.

Using on-demand capacity, DynamoDB could support any scale with virtually infinite requests per second. This was key to enabling Zoom to grow from 10 million to 300 million users while offering the smoothest experience possible.

Adaptive Capacity

DynamoDB’s on-demand mode could support Zoom’s rapidly changing traffic patterns.

Zoom also didn’t need to manually intervene by changing capacity planning. There was also no risk of under or over-provisioning with on-demand mode.

On-demand throughput capacity simply adapted to Zoom’s massive scalability needs.

Fully Managed Service

The fact that DynamoDB is fully managed meant Zoom could eliminate database management and configurations during this critical growth phase.

This also meant that their engineers could be free to focus on product development, and user experience rather than infrastructure management.

Database Implementation

Here are 4 key elements Zoom must have used to reach that scalability and support its quickly growing user base:

1. Data modeling for scalability: concepts like the single table design and avoiding hot partitions are critical at such scales.
2. Primary Key design: High partition key cardinality and overloading sort keys are equally important for data to be efficiently queried and to avoid hot partitions.
3. Global Tables: With a global nature, Zoom had to make use of DynamoDB’s global tables service to support low-latency queries to their database across the world.
4. Auto Scaling: Zoom had to use auto-scaling to automatically adjust provisioned capacity based on their application’s traffic.

Results

Thanks to DynamoDB, Zoom was able to scale 30x higher without the need to re-architecture or over-provision hardware [2].

Zoom was able to maintain single-digit millisecond response times despite the massive traffic spikes they experienced in a small timeframe.

Since DynamoDB’s flexible scaling allowed Zoom to match capacity closely with demand and avoid over-provisioning resources, it helped them avoid unnecessary costs.

One of the highlight benefits was that DynamoDB allowed the engineering team at Zoom to leave database management to DynamoDB and focus on product design and making sure the user experience was smooth.

Zoom’s case study really shows the importance of choosing a scalable and flexible database for your application.

With a database solution like DynamoDB, you can achieve that scalability and performance easier.

All that is required is understanding how it works and how to model your data efficiently for scale.

You can check out this article for a beginner’s guide:

https://towardsaws.com/the-ultimate-guide-to-getting-started-with-dynamodb-6e6dbee83ad3

?? My name is Uriel Bitton and I hope you learned something in this edition of The Serverless Spotlight

?? You can share the article with your network to help others learn as well.

?? If you want to learn how to save money in the cloud you can subscribe to my brand new newsletter The Cloud Economist.

?? I hope to see you in next week's edition!