Case Studies

Unlocking Time-Series Analytics in Snowflake with A Global Active Fixed Income Company

storage cost reduction

0 %

faster POC delivery than industry norms

0 x

Delivered proof of concept for ingestion of KDP time-series data to Snowflake.

Demonstrated the cost and compute advantages of utilizing Snowflake to deepen replicate time-series analytics.

Laid the foundation for an end-to-end data modernization project and sunsetting of on-premise legacy systems.

Challenge

The client, a global active fixed income leader, was mired in a costly, on-premise tick data swamp that included inefficient ETL pipelines, siloed data, and legacy technology debt. The customer needed to identify a more efficient, affordable, and scalable way to replicate its existing time-series analytics capabilities on the modern data stack, and needed a data partner that could prove out those capabilities.

Their objective for this POC was twofold: to unlock the analytical and competitive advantages of a modern data stack and to slash the considerable costs that accompanied their legacy architecture.

Solution

The customer selected Hakkoda to conduct a Proof-of-Concept (POC) that would demonstrate the ingestion and pipeline capabilities needed to migrate from KDB to Snowflake and to replicate time-series analytics using native Snowflake features.

Hakkoda’s successful delivery on this POC included the replication of the firm’s existing time-series data model and Refinitiv data set in Snowflake. Due to the deep industry knowledge the team brought to the engagement, we were able to deliver the completed POC in just under two months.

Once implemented, this Snowflake-native solution would give the client team access to a host of time-series functionality unique to the platform, including access to TIME_SLICE, ASOF Join, and Trend Analysis with Range-Based Sliding Window features that would greatly augment their existing trade cost analytics capabilities. Finally, the Hakkoda team provided storage and compute cost estimations that outperformed the existing, on-premise data estate.