Tradelab, a programmatic marketing platform company based in France, uses MemSQL to process and analyze real-time bidding data for hundreds of customers.
Challenge: NoSQL Data Latency
The Tradelab real-time ad serving platform requires a heavy mixed read/write workload, and the NoSQL database they had in place was introducing unnecessary data latency into the ad-bidding process. The company began searching for a replacement – a true real-time data management solution with two key criteria:
- Capacity to store more data in-memory, granting real-time performance
- Ability to analyze relational and JSON data together to power the Tradelab platform
Solution: High Performance Database with Horizontal Scale
As a fully relational, distributed SQL database, MemSQL was the perfect solution to these latency challenges.
MemSQL scales out horizontally on commodity hardware, allowing Tradelab to add nodes as it grows. Tradelab can also scale while supporting mixed reads for ad serving, and writes for impression and campaign tracking. In addition, MemSQL increases system performance by 30x, delivering millisecond response times, and reducing overall latency and cost.
Tradelab was able to replace its existing solution with MemSQL in a snap – MemSQL is fully wire protocol compatible with MySQL, integrating nicely into the existing Tradelab technology stack.
“We use MemSQL as an in-memory solution to analyze large amounts of data without sacrificing our competitive performance,” said Vincent Mady, CTO, Tradelab. “The effectiveness of our platform has increased significantly: we’re now able to keep pace with increasing data volume and provide faster insights for our customers. Ultimately, this is a better experience for their users.”
The Tradelab real-time analytics and bidding software now runs on MemSQL, which has dramatically increased the efficiency of its media traders and the accuracy of ad campaigns.
To learn more about how MemSQL provides Tradelab customers a competitive edge with instantaneous ad recommendations, read the official News Release.