Notable DBMS presentation: Yao Yu on how Redis is used at Twitter

by Matthias Gelbmann, 25 September 2014
Tags: Redis

Building the caching layer at Twitter is not an everyday challenge. Yao Yu presents why and how they are using Redis for that task.

From time to time we at DB-Engines come across presentations that we find to be outstanding in providing theoretical or practical insights into database management systems or their application. On these occasions, we want to contribute to spread the word by presenting them in our blog as "notable DBMS presentation".

Yao Yu from the Core Infrastructure team at Twitter gave such a presentation entitled "Scaling Redis at Twitter". Yao Yu not only shows the technical competence to give a notable presentation, she is also able to explain very well their problems and decisions to an audience that is not familiar with the problem at hand.

She explains in 75 minutes why Redis is a better choice than Memcached for their use case; where they use Redis and where they rely on other systems; how they managed to get stuck at Redis 2.4; how they built a Redis cluster and in which directions they would love Redis to evolve.

Of course, most of us are rarely faced with systems at Twitter scale. Therefore some of the aspects of this presentation will not be suitable for direct application into most people's daily work. Nevertheless, if you were wondering whether this Redis-toy is really ready for production, and if it's worth the trouble introducing a new system into your landscape and what exactly it might be used for, then this presentation might help you to come to a decision.

Share this page