Signal: Accelerating Large Scale Business Processes

Speaker: Jason Yanowitz, Vice President of Engineering, Signal

Read the case study

Signal is an identity resolution platform. So, for marketing folks who have a huge amount of data, some of it’s online, some it’s in their CRMs, and what they call offline sources, we allow them to bring that all onboard onto our platform and use it to target their existing customers all over the internet.

The main business challenge we were looking to address with Aerospike was to replace our existing data store. One of the biggest problems we were running into was latency. And it was slowing down every element of our business processes.

Compared to the other solutions that we were evaluating, the main drivers that made Aerospike so attractive was its total cost of ownership was far lower than the competitive offerings that we had evaluated. The engineering team during the pre-sales process was so engaged with us, and clearly understood our problems, that it gave us a tremendous level of confidence about migrating.

During our initial deployment to Aerospike and our shaking of it out, we pushed the limit up to 8 million transactions per second and saw the p50 at 10 microseconds. Which was absolutely stunning to us, and almost a thousand times faster than what we were seeing before that. When we switched over to using Aerospike, we saw immediate improvements on a bunch of axes. One was, the data was more reliable, and what we were putting in there was actually what would come back out. Secondly, we saw huge performance improvements. Our p99s went from 3,900 milliseconds to 23 milliseconds.

And probably most importantly, the way Aerospike has helped us is for our large-scale business processes, they’ve gotten much faster. So we have some processes that used to take six days that now take 14 hours. Things that took three hours take three minutes now. And so, across the spectrum it’s been much better.

Because of the performance of Aerospike, even while we’re live and taking active traffic, we’re able to completely take our dataset and stick it into a data warehouse for ongoing analysis. Which was a much more difficult and time-consuming processing with Cassandra.

Before Aerospike, we were spending more and more of our time on the care and feeding of Cassandra, and less and less time on the building of new product offerings. With Aerospike, we’ve now cleared the roadmap and we’re just focused on adding new functionality to our platform for our customers.