Key enhancements include a new RPO-driven backup capability and the first automated Any-Point-In-Time Recovery capability for NoSQL databases.
Peter Smails, Imanis Data’s CMO
San Jose-based Imanis Data has announced the 4.0 version of their machine learning-driven Data Management Platform, which is focused on Hadoop and NoSQL. This release extends the machine learning automation to new functionality, particularly through SmartPolicies, which are a new RPO-driven backup capability. They are also announcing ThreatSense Intelligence Augmentation, which improve anomaly detection, and which has special value as an anti-ransomware capability. Recovery capabilities to any point in time have been added. Finally, enhanced management capabilities include expanded database support, enhanced Fastfind search capabilities and simplification of system maintenance.
“We are focused on the blue ocean of Hadoop and NOSQL,” said Peter Smails, Imanis Data’s Chief Marketing Officer. “We aren’t looking to make a better mousetrap for Oracle databases. We are focused on an underserved market, and from a channel standpoint, it creates an enormous opportunity. We have been building out our channel very strategically. We are looking to work with partners who get it when it comes to these newer technologies, and who are making the investment in them.”
Imanis Data was founded by Big Data veterans from companies like HortonWorks and DataStax, and their focus since their inception five years ago [originally as Talena before a 2017 rebranding] has been on building the next generation of data management that would address the enormous expansion beyond the traditional relational database world using machine learning-powered data management.
Jay Desai, Imanis Data’s VP of Product
“SmartPolicies are an RPO-driven backup,” said Jay Desai, Imanis Data’s VP of Product. “Before, to do backup, a human would do the calculations manually and schedule them. SmartPolicies use machine learning to take backup out of human hands. It removes the possibility of manual error. It saves them time that would otherwise be spent on doing this. And it provides flexibility for business requirements that you don’t get with a static four-hour or six-hour backup cycle. The machine learning predicts how long a backup will take, depending on a given RPO. It also allows for the flexibility that comes from seasonality, or something as simple as weekdays versus weekends.”
SmartPolicies also have a predictive capability.
“It can do predictions, based on the machine learning model, on what’s likely to happen in the next 30 days, for example, with recommendations on what to do if the capabilities fall short of the requirements,” Desai said.
The Imanis Data platform had anomaly detection before, but it has been upgraded with this release.
“ThreatSense Intelligence Augmentation provide enhancements to the anomaly detection, to defend against threats like ransomware,” Desai said. “We have integrated human feedback with the machine learning-based anomaly detection. That’s because machine learning can create false positives on things that in certain cases can be normal behaviour. Humans can give feedback, and the machine learning uses it to eliminate the false positives.”
Another innovation is automated Any-Point-In-Time Recovery [APITR] for NoSQL, a first in the industry.
“Any-Point-In-Time has been sadly lacking in the world of NoSQL,” Smails said.
“Relational databases all have this capability,’ Desai said. However, some of the NoSQL databases like Couchbase have none, and others, like Mongo and Cassandra, are rudimentary and not as sophisticated compared to the relational databases.”
That kind of lag in development isn’t something abnormal.
“In the early days of Oracle and SAP, there was a 7-8 year lag in supporting data management capabilities,” Smails said. “Their job number one was to provide a good database, not the supporting data management capabilities. This is the same thing.”
“In providing the fully automated APITR recovery for the different NoSQL databases, we have taken the uniqueness out of each type out to be able to handle all of them with our software, even though they each have their own unique approach,” Desai said. “They can restore data to any specific point in time. All the user needs to do is to give us a date and time. It’s also very simple to do the restore.”
The 4.0 adds additional platform support for DataStax, Couchbase, HDInsight, HortonWorks and Vertica, joining Apache Cassandra, Apache HBASE, Cloudera, Hadoop, Microsoft ADLS, and MongoDB.
“We have also enhanced our FastFinds search capability, with a job tag listing and the ability to browse as well as search, Desai said. “And we have simplified system maintenance by making it single-click.”
