San Jose-based Imanis Data, which makes a machine learning-driven data management platform, has announced a new platform release which improves its ransomware detection capabilities, added a data mirroring capability for disaster recovery for Hadoop, and implemented a new Recovery Sandbox for non-disruptive recovery testing.
Imanis Data was founded five years ago as Talena, and rebranded itself in 2017 to avoid confusion with a non-IT company which had the same name.
“The company was originally founded by Big Data folks from companies like HortonWorks and DataStax,” said Peter Smails, Imanis Data’s CMO, who recently joined the company from Datos IO, following its acquisition by Rubrik. “We have been making money on backup for Big Data and NoSQL, but our DNA isn’t as a backup and recovery company – it’s data management. Our objective has been to build the next generation of data management to accommodate the shift from a relational database world, and our vision of the company is machine learning-powered data management.”
The company underwent a major executive shakeup in March of this year.
“It coincided with our B series round,” Smails said. “The company had had some success in landing blue chip organizations as customers, but it was largely engineering-focused, and was not well known. John Mracek, a successful multi-time CEO, was brought in as CEO.” Between 2009 and 2017, Mracek led NetSeer, a machine-learning based concept advertising startup from pre-revenue to $21 million in sales, which led to its purchase by Inuvo. Nitin Donde, Imanis Data’s former CEO, became the Chief Operating Officer, where he heads up product and technology.
“We also brought in a new Chief Revenue Officer who knows how to scale a business,” Smails said. “I came in to lead an increased focus on marketing, to make the company better known. All this is aimed at scaling the business. The timing had become right for it.”
Smails indicated that a key part of his own role is to draw attention to Imanis Data’s unique differentiation.
“We differentiate ourselves as a data management company first and foremost – with backup being just a use case for us,” he said. “Customers don’t want just want backup. They want data management.”
Smails emphasized three differentiated features.
“First, we are built on a massive scalable distributed Big Data architecture, which is petabyte scale, and designed for Big Data,” he said. Other startups in this space, particularly on the hyperconverged side, also meet this test, but Smails said that their second differentiation is unique.
“We are data-aware,” he said. “Backup isn’t data aware. It’s about making a copy. We know exactly what type of data it is, where it was created and the schema in which it is stored. We can build masking directly into the product, because we understand the schema.”
Imanis Data’s machine learning technology is its third differentiation, Smails said.
“We are machine learning-based in our DNA and deliver it today,” he stated. “ThreatSense, our machine learning anomaly detection software, learns about backup patterns with no user intervention. Our machine learning library is built into our core platform, and we will continue to add use cases for it.”
A key part of Imanis Data’s strategy for expansion is expanding their reach by moving from a predominantly direct sales model, and adding a select value channel.
“We are today largely direct, but are strategically expanding,” Smails said. “We are partnered with other vendors on the data and platform side. We are looking at reseller channel partners, but are being very strategic about it, because data management is a C level conversation. That’s very different from backup. The resellers who we want to work with are the ones who get that.”
Three major enhancements to the Imanis Data Management Platform were announced.
“Our ThreatSense software has been enhanced, to double the number of metrics that we track to build the baseline data models, which increase its granularity, improve anomaly detection rates, and lower false positives,” Smails said. This will enhance its effectiveness against ransomware.
A new Recovery Sandbox feature has also been added, to facilitate periodic testing of backups without disruption to production environments.
“Many organizations don’t go through recovery testing because of this disruption,” Smails said. “This feature enables you to go through a full restore within the platform, without any impact on production server workloads. It’s not a simulation, but a full recovery test that doesn’t affect production environments.
Finally, a new data mirroring feature lets customers automatically create periodic and incremental copies of Hadoop databases from one data centre to another, enabling them to rapidly recover in the cloud in case of a data centre outage.
“It’s an option to create a copy of the data in a Hadoop database, and store it offsite,” Smails said.