StarRocks introduces cloud-native SQL engine for faster real-time analytics

California based StarRocks announced a cloud-native version of its SQL engine to help enterprises deliver faster and more affordably real-time analytics†

With data exploding several times each year, companies are looking for ways to extract value from the growing amount of information in their backyard. Part of the effort has focused on leveraging real-time data — information about events as they happen. It’s the new holy grail, with a number of organizations processing and responding to event streams for use cases like outage detection.

However, when it comes to analytics, where patterns in data are discovered, working with real-time streams can be a major challenge. After all, you can’t run analytical queries if information is constantly being inserted, updated, and even deleted. The problem becomes multiple when dozens of people try to request data at the same time.

StarRocks’ unified engine

StarRocks solves these problems with a solution that unifies real-time analytics, an OLAP database, and data lake analytics in a single engine, with a single data pipeline.

“This is the industry’s first analytic database to address the critical technical challenges in real-time analytics, such as the need to denormalize data, the inability to handle updates, and the challenge of supporting large numbers of concurrent users. We did this by creating a brand new query engine with many groundbreaking technologies,” Li Kang, VP of strategy at StarRocks, told Venturebeat.

The engine is purposely designed to support real-time data and a large number of concurrent users, with join queries across multiple tables. According to the company, it can ingest data from streaming stores at 100MB/s per node and perform more than 10,000 searches per second. Ultimately, this allows enterprises to combine their latest streaming transaction data with historical records for effective recommendations and decision making.

More than 500 companies, including Airbnb, Trip.com and Lenovo, have already adopted the solution.

Cloud offering

With the new cloud-native version, available as a fully managed SaaS platform, StarRocks makes its product more lucrative for enterprises.

In short, StarRocks Cloud enables organizations to integrate existing data infrastructure into the cloud and do away with the regular engineering and administrative tasks required for real-time analytics, from setting up the servers/virtual machines to deploying the software.

In addition, cloud support also brings various cloud-specific features such as separation of compute and storage, automated resource management, etc. for end-user insight.

Contest

Currently, multiple organizations are looking at the real-time analytics space, including ClickHouse, imply (Apache Druid), StarTree (Apache Pinot) and rock set (RocksDB). However, StarRocks claims to offer a much better price-performance ratio in comparison.

“StarRocks outperforms – we published benchmark test results showing we have 3x to 5x performance advantages over our competitors… In addition, StarRocks is more cost effective – with our industry-leading query engine we are able to get better query performance without complex data pipelines and heavy indexing processes, resulting in a much better value for money. Finally, StarRocks is more flexible. Our unique design can handle frequent updates of past transactions, while maintaining high query performance in real time. This allows us to analyze use cases support previously unsuitable for real-time analytics,” said Kang.

StarRocks Cloud will be generally available in Q3 2022 on Amazon AWS, with support for Google Cloud Platform later.