Highlights:

  • Rockset’s key feature is compute-compute separation, curbing performance risks in real-time analytics.
  • The company asserts that its database can handle the ingestion of tens of millions of data points every second.

Rockset Inc. recently announced that it has raised USD 44 million in funding to support growth initiatives, which Icon Ventures led. Several other institutional investors joined the venture capital firm, including Sequoia Capital, Greylock Partners, and others. Rockset now has 105 million USD in outside funding due to the investment.

Industrial equipment often has sensors built into it by manufacturers to look for problems. Data from a sensor that gathers information about potential impending failure must be immediately analyzed to speed up troubleshooting. Other use cases also call for the capacity to analyze data as soon as it is generated, covering fields like cybersecurity and e-commerce.

Implementing such use cases is made simpler by the Rockset database. According to the company, the database can quickly and efficiently ingest data from sources like industrial sensors and make it available for analysis. Applications can begin running queries within two seconds of receiving a record.

Co-founder and Chief Executive Officer of Rockset, Venkat Venkataramani, wrote in a blog post recently, “We believe modern data apps should operate on data in real time. The best apps are the ones that serve as a better windshield for your business and your customers, and not be a glorious rear-view mirror.”

Compute-compute separation is one of Rockset’s most prominent features. The company claims it lessens the chance of real-time analytics applications experiencing performance drops. Sudden slowdowns can significantly impact user productivity because real-time applications must, by definition, process data quickly.

A typical database processes and analyzes new records using two different software modules. The processor resources for those two modules are typically pooled together. If one module uses more processor resources than usual, the other will have less computing power available, causing it to operate more slowly.

In real life, a sudden increase in the number of records a database ingests can cause analytics queries to be sluggish. Conversely, a complicated query might make a database ingest new records more slowly.

According to Rockset, these performance problems are avoided by its compute-compute separation feature. According to the company, the feature allocates two pools of processor capacity to the module responsible for ingesting new records and the module responsible for their analysis. As a result, a spike in usage that affects one module won’t slow down the other.

Rockset added embedding support to its database’s feature set in April. A mathematical structure called embedding is what artificial intelligence models use to store the processed data. Due to the April update, Customers can now use Rockset to power their AI applications more easily.

According to the company, its database can take in tens of millions of data points each second. Additionally, Rockset asserts that the platform outperforms Elasticsearch, a well-liked open-source substitute, in many instances regarding cost-effectiveness.

Companies are interested in Rockset’s value proposition. The company announced that its revenue and customer base have doubled each of the last two years in conjunction with its most recent funding round. Rockset intends to increase the size of its go-to-market team and spend money on new database features in order to grow its installed base.