Venture-backed data processing startup StarTree Inc. today launched its first product, a cloud analytics platform called StarTree Cloud that it says can run more than 100,000 queries a second.

The platform is launching less than two months after StarTree raised a $24 million funding round from a group of investors that included Microsoft Corp.’s LinkedIn unit. Several prominent executives from the enterprise software segment participated as well.

StarTree Cloud is designed to power analytics applications that have thousands to millions of users. One example of such an application is the analytics tool Uber Technologies Inc. provides as part of its UberEats service to help restaurant operators measure sales. A large retailer, meanwhile, could apply the same concept to create a demand forecasting dashboard for each of the thousands of store managers at its brick-and-mortar locations.

Building such applications is highly challenging because the more users interact with an analytics system, the more data it has to crunch at once. The volume of data and requests can quickly become difficult to manage. Another dimension of the challenge is that many analytics applications are expected to generate results in close to real time, which is technically complicated to implement even when there aren’t thousands of users running queries at once.

StarTree is positioning StarTree Cloud as a way to simplify such projects. The platform is a commercial version of Apache Pinot, an open-source data store developed by StarTree’s founders. It’s specifically built to power real-time analytics applications with larger user bases.

The data store incorporates another open-source tool, Apache Helix, that allows it to scale across a large number of servers to accommodate applications with a significant number of users. The data that the applications use for analytics, in turn, is kept in a so-called columnar store to speed up queries.

Traditional databases take a so-called row-oriented approach to managing information. In such systems, a record such as a purchase log can be represented as a row of data fields stacked one atop another, with each field containing a different piece of information such as the size  of the purchase and when it was made. If a user wishes to access the information inside the fourth data field of a ten-field row, the database would first have to check the first three fields that are located above it.

A columnar store doesn’t have to check the first three fields. Instead, it can jump directly to the level of the row containing the requested information, which allows queries to run considerably faster.

StarTree Cloud can process up to 170,000 queries per second, StarTree claims, which it says makes the platform several times speedier and more scalable than competing products. The platform is capable of analyzing historical data from sources such as Amazon S3 and Google Cloud Storage. StarTree Cloud can also ingest real-time information such as stock price changes thanks to an integration with Apache Kafka, a popular platform for real-time data processing.

As a cloud service, StarTree Cloud allows companies to use Apache Pinot’s high-speed analytics features without worrying about managing their deployments. The platform automates maintenance tasks such as rolling out updates in a bid to free up time for administrators. 

By reducing the amount of manual maintenance work required, StarTree Cloud potentially also lowers the risk of technical issues caused by human error. That might emerge as a major selling point given the type of use cases StarTree targets.

An outage in an internal analytics system that supports a small business intelligence team doesn’t necessarily cause a major business disruption, but if the outage hits an analytics feature of a food delivery app used by millions of users, it can have a much more significant impact. As a result, reducing the risk of disruptions is likely to be a key priority for enterprise adopters. 

StarTree is launching StarTree Cloud into a market with significant growth potential. The startup says companies are increasingly building analytics features into their user-facing applications to stand out from rivals.

“Companies have a remarkable opportunity to provide insights to their users to help drive decision making and behavior,” said Kishore Gopalakrishna, StarTree’s co-founder and chief executive. “The next revolution in analytics is the recognition that every user is a decision maker in their own right, and that companies can add value by helping users make smart decisions.”

StarTree is backed by a total of $28 million in venture funding. 

Photo: Unsplash