HEAVY.AI has released a limited release of its latest version, HEAVY.AI v6. It is due to be generally available next month. The key feature is HeavyConnect. The company says that this will allow users to “immediately analyze and visualize their data wherever it currently exists, without the need to export or import data or duplicate storage.”
Jon Kondo, CEO of HEAVY.AI said, “Enterprises today recognize the importance of data to guide business decisions. Yet, organizations are challenged when it comes to managing big data and making sense of rapidly growing, increasingly complex datasets. As a result, they often make critical decisions based on incomplete and old data.
“HEAVY.AI 6.0 enables enterprises to quickly unify and analyze all data as they collect it, regardless of how complex or messy it may be. With timely information from real-time data, these organizations will be able to uncover new opportunities and risks to support better business results.”
Five key features for HEAVY.AI v6
HEAVY.AI has announced five key features for v6. They all focus on faster data analysis and making it easier for users to work with ever-larger data sets.
HeavyConnect: This feature allows users to work with data wherever it resides. The key here is no export, no import and no data duplication. It means no risk to their data for organisations that have invested in data quality.
HeavyDB: There is a range of new SQL capabilities and performance improvements. Among them are real-time string operations that can run on both GPUs and CPUs. Other new features include more comprehensive window function support and system and user-defined table function support.
Heavy Immerse: The browser implementation has a new administration panel. It delivers a comprehensive view of existing deployments. Administrators can now see how many tables and dashboards a system has, who owns those assets, the amount of storage and memory they’re using and more.
Performance and Scalability Improvements: This includes 10X faster initial joins, parallel query compilation, and dramatically faster data egress to Immerse and other clients.
Interactive ML and Custom Algorithms: Users of machine learning (ML) models can integrate those models into analytics workflows. It is done through SQL and Heavy Immerse. Importantly, it gives users real-time visual feedback on how model accuracy is impacted by changes to input data and model parameters.
Making data analysis more powerful
The five features above add some significant new power to HEAVY.AI. The HeavyConnect will be welcomed across the board from IT departments to data scientists and users. Not having to spend time exporting, importing, normalising and making data fit for analytics is a big bonus. The saving in time, compute power and storage will be significant.
Organisations that have invested in data quality and enrichment will welcome this release. Often, this effort gets broken down when users start making duplicate data sets for analysis. That leads to dirty data and poor results.
The other standout here is the ability to work with ML models. To bring that data into your analytics tools is good news. It will give a greater return on the investment in those models as they become more widely used. Seeing how data changes impact models will also be of interest outside of data teams. Audit teams are increasingly looking at how data is used and its impact. They can now identify what changes are likely to diminish the effectiveness of data.
Of the other new features, speed, performance and easier administration are the order of the day.
Enterprise Times: What does this mean?
Users spend a lot of time and effort trying to get the answers from the data that organisations hold. The problem is that many of the mechanisms they use can be unproductive. Data gets dirty, duplicated, changes are made for modelling, and eventually, the data becomes untrusted.
What HEAVY.AI is doing here is offering a set of tools that will reduce those problems and, for some customers, eliminate them. It also makes it easier to work with those ultra-large data sets that many organisations have, but struggle to get the most out of. They end up working with just sections of the data due to the size of the data set and the time required to export and duplicate it. Now they can work with all their data.
It will be interesting to see how quickly HEAVY.AI customers move across to the new version of the software.