Rockset Sets Industry Standard in Real-Time Analytics Performance
Rockset is up to 9.4x faster than the alternative when measured against the Star Schema Benchmark for query performance
Rockset, the real-time indexing database company, announced it has published new benchmark results showing millisecond-latency query performance against the Star Schema Benchmark (SSB). When coupled with RockBench, a benchmark measuring data latency, Rockset is the only real-time analytics solution to publish benchmarks showing it can execute queries up to 9.4x faster than the alternative, and also ingest one billion events a day with one second data latency.
The performance of real-time analytics solutions can be measured along two dimensions: data latency and query latency. As the need for real-time analytics becomes more critical, technology stacks require a database that is capable of both high write rates and low-latency queries, enabling applications to operate in real-time. Once data is made queryable, applications must take action quickly to deliver immediate insight to end users. Failure to meet latency requirements can result in missed opportunities, failure to detect threats, or a poor user experience. Low latency is a key requirement across the broad range of use cases that now rely on real-time analytics, including commerce sites, real-time supply chain logistics and delivery tracking systems, gaming leaderboards, fraud detection systems, health and fitness trackers, social media newsfeeds, and more.
A widely recognized industry-standard benchmark, the SSB is designed to measure database performance for analytical applications, yielding valuable insight into Rockset’s query performance on a range of common analytics queries. The findings on the SSB show Rockset executed every query with sub-second latency, with a median response time across all queries of 254 milliseconds. Rockset was on average 1.5x faster than an open-source alternative, with one query executing 9.4x faster, and most queries executing 2x, 3x, or 4x faster. SSB’s suite of 13 queries was completed in 4,146 milliseconds.
Rockset’s key competitive differentiator is that it indexes all fields, including nested fields, in a Converged Index™ which combines an inverted index, a columnar index, and a row index. The SQL optimizer uses these indexes in parallel, exploiting selective query patterns and accelerating aggregations over large numbers of records, to achieve millisecond latencies at significantly lower compute costs. Rockset’s execution engine is built from the ground up using vectorization techniques, which offer order of magnitude speedup over the traditional row-based engines. Additionally, the new release supports column-based clustering, where users can colocate data according to a clustering key they specify. This maximizes the opportunity for sequential access and reduces the amount of data that needs to be scanned for each query.
Recommended AI News: Dream Advisors Acquires Additional German Logistics Assets
While these findings prove Rockset’s capabilities for delivering low-latency queries needed for real-time analytics, in September 2020 the company also released RockBench: a benchmark that measures time from when data is produced to when it can be queried, showing an application’s data latency — a factor that is rarely prioritized in most current database benchmarks. Measuring Rockset against RockBench showed that a Rockset 4XLarge Virtual Instance can support one billion events flowing in every day, while keeping the data latency to under one second.
“Real-time analytics demands fast queries on fresh data. We are on a mission to make real-time analytics faster, more flexible, and easier than current alternatives like Elasticsearch and Apache Druid,” said Venkat Venkataramani, CEO and co-founder of Rockset. “Modern data applications today demand both speed and scale. As a massively distributed system, Rockset is designed for cloud scale. Our SSB results prove that Rockset is one of the fastest real-time analytics solutions in the market today.”
Recommended AI News: YouTube’s Top Earner Only 9 Years Old – Almost $30 Million in 2020