Splunk goes native with machine learning, aims to speed up monitoring
Analytics biz eyes up fraud and IoT markets
Analytics firm Splunk is making machine learning central to the next generation of its enterprise solution, and claims it performs 20 times faster than before.
Splunk's platform drills into the machine-to-machine data generated by big companies to help them identify weak spots and boost operations and performance.
It has around 14,000 customers, which include healthcare providers, financial services firms like Nasdaq and Black Rock, and retailers such as John Lewis.
In a bid to woo more businesses, the latest release of its Enterprise solution, launched today at its annual conference in Washington, comes with added algorithms.
"We're now natively embedding machine learning into everything we're doing," Guillaume Aymé, IT markets evangelist, told The Register. "The machine learning package will help customers make better decisions."
The overall aim is to offer more insights into the data generated by companies' machines and help them boost performance, but Splunk also hopes that by making the tech native across all its products, it will appeal to smaller firms that might not usually go for the tech.
"We're trying to lower the bar of entry for organisations to use machine learning without having to hire data scientists," Aymé said. "They just want it to work, they don't need to know how it works underneath."
Machine learning will now be rolled into the latest versions of its products, with Splunk saying the solutions are focused on specific pain points like fraud and cloud monitoring.
For instance, User Behaviour Analytics helps detect insider trading and potential security breaches by searching for anomalies, such as someone logging on when they shouldn't be.
Meanwhile, its IT Service Intelligence product uses machine learning to distinguish between a "normal" anomaly – for examples banks might expect to see a surge in use at the end of the month – and one where performance needs to be improved.
The other big change for Splunk Enterprise 7.0 is improved metrics, which will allow customers to better analyse values, such as temperature. The firm claims the change will speed up monitoring and alerting by "at least" 20 times.
"If you think of Internet of Things devices, they're dumb, they're just sending out numbers. We're allowing customers to analyse that data much more easily," said Aymé. ®