ScyllaDB NoSQL for
AdTech and MarTech

The largest privately held software company in the world, SAS Institute’s Customer Intelligence platform supports billions of advertising decisions monthly, helping digital marketers and web publishers with A/B testing, product recommendations, and customer engagement tracking. Seeking lower TCO and predictable low-latency, SAS replaced DataStax Enterprise with ScyllaDB, migrating production workloads in under two weeks with no downtime or lost data.

Numberly’s multi-platform, multi-device, multi-format solutions encompass real-time bidding, targeted display advertising, customer relationship management, and an advanced data management platform. By moving to ScyllaDB, Numberly reduced their infrastructure footprint by 80% and achieved cost savings, ease of operation, and production reliability.

Steelhouse offers a television advertising service for major networks, including NBC, Universal, and ABC. Steelhouse seamlessly migrated mission-critical production workloads from Apache Cassandra to ScyllaDB and saw an instant improvement in stability with far fewer timeouts and 100% SLA compliance. After switching to ScyllaDB, Steelhouse experienced their smoothest Black Friday and Cyber Monday ever.

AdGear delivers targeted television ads and uses ScyllaDB as a user data store to support extremely low latency, and high availability requirements. AdGear moved from Apache Cassandra to ScyllaDB as a better database platform for their real-time bidding and ad retargeting use cases. AdGear not only benefits from much lower and more predictable read latencies but also saved money on hardware. While Cassandra bottlenecked AdGear to 1 in every 7 bid requests, ScyllaDB enabled AdGear to bid 95% of the time.

Using the latest in deep machine learning technology, GumGum’s contextual intelligence platform provides sophisticated in-image and in-video contextual targeting for ads. With Apache Cassandra, GumGum experienced high operational overhead and costs. And Cassandra simply did not perform well under high loads. GumGum moved to ScyllaDB Cloud with no major changes to their systems or code, and significantly improved the price-performance while virtually eliminating operational overhead.

Persosa helps media and entertainment companies create connected experiences for their viewers. Persosa chose ScyllaDB because consistently low-latency reads were the most important factor for their application. With ScyllaDB, requests went down from 80 milliseconds to about 9 milliseconds on average, primarily attributed to ScyllaDB’s consistently fast reads.

Zeotap’s Customer Intelligence Platform hosts one of the world’s largest identity graphs, consisting of nearly 16 billion nodes, 8 billion edges, and around 3.6 billion identities. Zeotap benchmarked JanusGraph on ScyllaDB, OrientDB, and DGraph. By adopting ScyllaDB, Zeotap reduced P95 SLAs from 10 hours to 2 hours and slashed the job failure rate from 20% per day to 2% per day.

mParticle provides a data platform that helps their customers, including NBC, Turner, Airbnb, Venmo, and Spotify manage and control the data they send to third parties, segment that data, and enrich it with identity information. To do so, mParticle processes more than 50 billion messages, 100 billion events, and 150 terabytes of raw data every month with very strict SLAs. mParticle adopted ScyllaDB for its raw performance, ease-of-configuration and low operational overhead.