April 16, 2010
Introduction to Datameer
Elder care issues have flared up with a vengeance, so I’m not going to be blogging much for a while, and surely not at any length. That said, my first post about Datameer was never going to be very long, so lets get right to it:
- Datameer offers a business intelligence and analytics stack that runs on any distribution of Hadoop.
- Datameer is still building a lot of features that it talks about, for target release in (I think) the fall.
- Datameer’s pride and joy is its user interface. Very laudably for a software start-up, Datameer claims to have spent considerable time with professional user interface designers.
- Datameer’s core user interface metaphor is formula definition via a spreadsheet.
- Datameer includes 124 functions one can use in these formulae, ranging from math stuff to text tokenization.
- Datameer does some straight BI, with 4 kinds of “visualization” headed for 20 kinds later. But if you want to do hard-core BI, use Datameer to dump data into an RDBMS and then use the BI tool of your choice. (Datameer’s messaging does tend to obscure or even contradict that point.)
- Rather, Datameer seems to be designed for the classic MapReduce use cases of ETL and heavy data crunching.
- Datameer’s messaging includes a bit about “Datameer is real-time, even though Hadoop is generally thought of as batch.” So far as I can tell, what that boils down to is …
- … Datameer will let you examine sample and/or partial query results before a full Hadoop run is over. Apparently, there are three different ways Datameer lets you do this:
- You can truly query against a sample of the data set.
- You can query against intermediate results, when only some stages of the Hadoop process have already been run.
- You can drill down into a “distributed index,” whatever the heck that means when Datameer says it.
- Datameer will let you import data from 15 or so different kinds of sources, SQL, NoSQL, and file system alike.
Categories: Analytic technologies, Business intelligence, Datameer, EAI, EII, ETL, ELT, ETLT, Hadoop, MapReduce
Subscribe to our complete feed!
Comments
3 Responses to “Introduction to Datameer”
Leave a Reply
BI reporting tools can connect to Hadoop through Hive (which has SQL access to the Hadoop). In theory, of course. So what will happen when SQL/ODBC/JDBC layer become standard in Hadoop? Users will get access to numerous open-source BI solutions like Pentaho and Jaspersoft. I think this is the future of Hadoop BI analytics – not proprietary Excel spreadsheets.
[…] previously noted, I’ve briefly cut back on blogging (and research) due to some family health issues. The first casualty was a post about memcached. One […]
[…] Approximate query results, aka “data sharpening”. Zoomdata simulates high-speed query by first serving you approximate query results, ala Datameer. […]