Today's the first day of Strata + Hadoop World at New York City's Jacob Javits Convention Center, and a slew of announcements are crossing the wire this morning.
Security at the source
Here's a big one: BlueTalon has announced technology implementing fine-grained security (filtering and data masking) right at the Hadoop Distributed File System (HDFS)-level. I haven't been briefed on this technology so I don't know exactly how it works, but the aim is clear: implement access controls on data at the file system-level so that the security controls propagate through to all Hadoop services which may attempt to read the data. How this reconciles with other security schemes -- some of them still emerging -- like Apache Sentry and Apache Ranger, is unclear. But if BlueTalon's solution works across the entire Hadoop stack, it will be a worthy contender.
Speaking of worthy contenders, consider MapR-DB, the Hadoop distribution vendor's own performance-optimized, HBase-compatible wide-column store NoSQL database. Turns out MapR-DB has been able to work as a key-value store as well. That capability has been there for a while, but today the company is announcing support for JSON data too, accessible to developers via an open source standard the company is putting forth called OJAI (Open JSON Application Interface). Want more details, check out my colleague Toby Wolpe's article.
How's my usage?
Have you heard of Attunity? It's a company that specializes in operational analytics focused on the actual usage of Hadoop clusters. Today the company is announcing version 7 of its Attunity Visibility product, which provides activity and usage data across Hadoop and enterprise data warehouse (EDW) installs. The company is also announcing Replicate Express for Hadoop, a downloadable version of the company's full Replicate product that's appropriate for test environments. Replicate Express supports both batch and incremental data loading, between Hadoop and relational databases (both OLTP systems and data warehouses) using change data capture (CDC) technology, and it's certified with multiple distros.
OLAP, Hadoop, data viz
One other bit of news this morning: Kyvos, an OLAP-on-Hadoop startup that emerged from stealth in June, is announcing a partnership with Tableau. Kyvos will exhibit at this year's Tableau Conference in Las Vegas and will join an array of companies in the industry highlighting the combination of their own technology and Tableau's popular data discovery and visualization offering.
But wait, there was more:
That's all of today's news (thus far), but it's worth mentioning several announcements from yesterday.
Pepperdata introduced a chargeback reporting system for Hadoop, allowing enterprises to make their Hadoop clusters multi-tenant, effectively facilitating Hadoop as a Service for the private cloud.
The Open Data Platform Initiative, spearheaded by Hortonworks, announced new members, progress on its spec and a governance tie-up with the Linux Foundation. Cloudera is still not part of this effort and continues to see it as counterproductive to the industry. Once again, my colleague Toby Wolpe has the details.
Impetus Technologies announced a new product for migrating databases and their associated workloads from EDW platforms to Hadoop. Impetus is calling the offering, logically enough, its Data Warehouse Workload Migration solution. While that rather long product name may cry out for an acronym of its own, the offering has even resulted in a related, three-letter acronym: BDW (Big Data Warehouse).
Microsoft launched its Azure Data Lake service, including components for both storage and analytics. It also announced developer tools for the new HDFS cloud-storage offering, including a new language it calls USQL. In addition, the company put the Linux-based flavor of its HDInsight cloud Hadoop offering into general availability.
Had enough news yet? Breakout sessions at Strata don't even start until tomorrow. So stay tuned to ZDNet for more.