Rockset and the Real-Time Cloud Data Stack
Four years ago, we partnered with Venkat Venkataramani, Dhruba Borthakur and team on a new platform for building data-driven applications. After joining forces with our friends at Greylock on both the Seed and Series A, we are delighted to triple-down on this tremendous team by leading their Series B.
One of the defining trends of the past five years has been the rise of ELT (Extract, Load, Transform). In the beginning of the “big data” era, enterprise data was shuffled from place to place via complex, brittle data pipelines. In the best case, these systems led to hours or days-old data and months of development time. In the worst case, these pipelines would randomly break and your applications would be operating on stale data for days or weeks.
Tools like Fivetran, Snowflake and DBT have ushered in a new model in the data warehouse space. Instead of a complex pipeline, the ELT model connects all of your data directly to your data warehouse (e.g., Snowflake) and then centralizes all of your transformations in a single system like DBT.
This model is transformative for warehouse-style use cases and enables queries that take seconds on data that is hours- or minutes-old.
What do you do if you need something even faster? What if you need application-speed queries (e.g., tens of milliseconds) on always-fresh data that is spread across multiple systems?
That’s where Rockset comes in. Similar to ELT systems, it connects to your data stores or data lakes, indexes them in real-time and then enables fast, scalable, complex SQL queries (with JOINs) across those data sets.
One common use case for Rockset is real-time analytics — aggregating data from multiple sources and enabling real-time metrics and dashboards across that data. Unlike others in this space, Rockset shines by also enabling rich, full-fledged applications that span these data sets. One e-commerce customer is using Rockset to power their personalization algorithm to ensure that their recommendations adapt in real-time as the consumer browses. Another customer is using it to track a complex logistics operation to ensure that they know the status of every asset in real-time.
In our references, the most common thing we heard from prospects was skepticism around whether this was technically possible. The most common thing we heard from active customers was that it was “magic.” For one customer, it took a six-month roadmap item and compressed it into an afternoon.
We think the speed, scale and simplicity of Rockset will be game-changing for the cloud data stack. We are thrilled to strengthen our partnership and excited for the road ahead.
