Monica Baluna and Ehab Tawfik, Bloomberg
One of Bloomberg's flagship products, Instant Bloomberg (IB), is used by financial professionals around the globe for instant messaging. This system is powered by a multitude of microservices, databases and UIs that interact through synchronous or asynchronous API calls and queueing mechanisms.
We recently released Forums in IB. This new form of group chat introduced exciting features. With our clients needing increasingly larger group chats, we took the opportunity to ask how to make sure the new system and the existing one can scale up with the extra load without affecting the existing user workflows.
This talk explores the different load testing strategies we adopted while enabling support for chats ten times larger than before, while also migrating existing group chats to become Forums. We will focus on two elements: (i) creating a realistic representation of production traffic in a test environment, and (ii) how to efficiently gather insightful metrics.
Monica Baluna, Bloomberg
Monica Baluna is a software engineer at Bloomberg in London, where she has worked for the past six years. Her main interests include distributed systems, as well as building reliable software and robust APIs. She has had an opportunity to explore these interests, as her team manages a content sharing solution where performance and scalability are key features. Monica earned a bachelor's degree in computer science and engineering from Politehnica University of Bucharest.
Ehab Tawfik, Bloomberg
Ehab Tawfik is a software engineer who loves problem solving, technology, and business. He works in Core Products Engineering at Bloomberg in London. He is passionate about back-end systems and distributed computing. Ehab earned a bachelor's degree in computer science and engineering from Nile University in Egypt.
author = {Monica Baluna and Ehab Tawfik},
title = {Enabling Product Scalability through Load Testing},
year = {2024},
address = {Dublin},
publisher = {USENIX Association},
month = oct
}
