Datanami recently reported on a new SQL on Hadoop benchmark from AtScale. SQL-on-Hadoop Test: Each Engine Has ‘Sweet Spots’ This is a benchmark using Tableau to generate the SQL – described as a BI for Hadoop benchmark. For 5 of 13 queries , Spark SQL 1.6 was faster than the latest version of Impala V2.3.…
Cloudera’s Feb 11 blog posting about new Impala benchmarks was quite interesting. http://blog.cloudera.com/blog/2016/02/new-sql-benchmarks-apache-impala-incubating-2-3-uniquely-delivers-analytic-database-performance/ It gave the impression that Cloudera published Apache Impala (incubating) performance results for 99 queries “derived” from TPC-DS. But if you read carefully you will find that Cloudera only published performance results for 47 of the 99 queries. For a single user.…
Earlier this year I blogged about Cloudera’s “benchmarketing” efforts which showed Impala running a 20 query subset of the industry standard TPC-DS benchmark. Well now its time to blog about IBM Big SQL running all 99 Hadoop-DS queries. What’s that? Hadoop-DS is inspired by, and is highly compliant with TPC-DS including full compliance of the…
Not so long ago, Jeff Kelly wrote on wikibon about customers shifting workloads from “traditional” DW to Hadoop, the role that SQL on Hadoop has, and that “some SQL on Hadoop offering have more SQL functionality than others” I think architecture also plays a role. I see three ways to look at vendor architectures – just…
IBM is extending the value of Big SQL with Big SQL 3 – coming soon in the next release of Big Insights – as discussed this week at the IBM Impact conference in Las Vegas. I had previously mentioned work being done on Big SQL, in the context of Cloudera Impala. By extending the value,…
Lots of bandwagon jumping onto the “Data Lake”. Which leads to the jokes about how to avoid the Data Swamp, or Data Cesspool. I’ve never quite understood why anyone would call it a Data Lake. Nobody knows what’s in a lake, you have to go fishing. Lakes get polluted. They freeze over and cause ice…
A recent blog posting by Merv Adrian of Gartner, “Hadoop is in the mind of the beholder” hit the nail on the head. While most commercial Hadoop distributions now include the same basic components (e.g. pig, hive, hbase, etc), each vendor is starting to add things that make their distribution different. Merv Adrian stated “The…
Update on April 8 Interview with cloudera CEO Tom Reilly clarified that cloudera is expecting to get approximately 60 percent of the overall $900m financing, reported on gigaom . So some money from Intel will go to cloudera to grow the business, but not all of it as originally implied. The financing is expected to…
This past Friday, March 28, IBM released BigInsights 2.1.2, the latest version of IBM’s Hadoop offering. This release bumps up the level of the Apache Hadoop open source components, and adds a few other goodies. Key enhancements include: BigR – end to end integration of R into BigInsights. It enables the use of R as…
Big Money for Cloudera and Hortonworks – What Does it Mean? Holy cow, there sure seems to be a lot of venture capital money being thrown at what are today – unprofitable – Hadoop distribution companies. In case you missed it, Cloudera recently announced $160m in new funding, and Hortonworks announced $100m in new funding.…
Sheffield View 