Comments on: Vertica’s version of MapReduce integration http://www.dbms2.com/2009/08/04/verticas-version-of-mapreduce-integration/ Choices in data management and analysis Thu, 09 Feb 2012 16:57:09 +0000 hourly 1 http://wordpress.org/?v=3.0.3 By: Will a Shotgun Marriage Avert Squabbles among the Data Clans : Beyond Search http://www.dbms2.com/2009/08/04/verticas-version-of-mapreduce-integration/#comment-153106 Will a Shotgun Marriage Avert Squabbles among the Data Clans : Beyond Search Thu, 17 Dec 2009 06:03:24 +0000 http://www.dbms2.com/?p=858#comment-153106 [...] TechWorld.com ran a story that I thought was interesting and closer to the truth about the relational databases and big data. “Sybase Embraces Google MapReduce” runs down a number of data management companies expressing interest in one of Google’s earlier innovations. One comment worth noting in my opinion was: Relational database pioneer Michael Stonebraker co-authored a paper earlier this year contending that that SQL technology still beats MapReduce in most cases. But that conclusion didn’t stop Vertica Systems, the startup where he serves as CTO, from adding Hadoop functionality to its new Vertica 3.5 database. [...] [...] TechWorld.com ran a story that I thought was interesting and closer to the truth about the relational databases and big data. “Sybase Embraces Google MapReduce” runs down a number of data management companies expressing interest in one of Google’s earlier innovations. One comment worth noting in my opinion was: Relational database pioneer Michael Stonebraker co-authored a paper earlier this year contending that that SQL technology still beats MapReduce in most cases. But that conclusion didn’t stop Vertica Systems, the startup where he serves as CTO, from adding Hadoop functionality to its new Vertica 3.5 database. [...]

]]> By: How 30+ enterprises are using Hadoop | DBMS2 -- DataBase Management System Services http://www.dbms2.com/2009/08/04/verticas-version-of-mapreduce-integration/#comment-152416 How 30+ enterprises are using Hadoop | DBMS2 -- DataBase Management System Services Sat, 12 Dec 2009 03:26:02 +0000 http://www.dbms2.com/?p=858#comment-152416 [...] (Vertica recently made its 100th sale, and of course not all those buyers are in production yet.) Vertica/Hadoop usage seems to have started in Vertica’s financial services stronghold — specifically [...] [...] (Vertica recently made its 100th sale, and of course not all those buyers are in production yet.) Vertica/Hadoop usage seems to have started in Vertica’s financial services stronghold — specifically [...]

]]> By: Vertica Projects Leadership, Embraces MapReduce (Sorta) « Market Strategies for IT Suppliers http://www.dbms2.com/2009/08/04/verticas-version-of-mapreduce-integration/#comment-134809 Vertica Projects Leadership, Embraces MapReduce (Sorta) « Market Strategies for IT Suppliers Wed, 12 Aug 2009 02:02:55 +0000 http://www.dbms2.com/?p=858#comment-134809 [...] MapReduce support, but with a difference. Unlike Greenplum and Aster, who are bringing it into the database itself, Vertica is providing a streaming connection to Hadoop instances (the open source implementation of MapReduce; Vertica is contributing the adapter to the community). This architecture mirrors usage patterns we’ve seen, and which Vertica asserts its customers have told them they want. One scenario: use your ADBMS to retrieve stored data, pass it to Hadoop for analysis by staff with different skill sets from the typical ADBMS users, and then bring result sets back. A separate hardware for the Hadoop sandbox is fairly typical among early adopters today, and via a Cloudera partnership, Vertica can offer a deployment architecture that doesn’t break the bank. Curt Monash does the usual excellent summary of Hadoop issues in his blog. [...] [...] MapReduce support, but with a difference. Unlike Greenplum and Aster, who are bringing it into the database itself, Vertica is providing a streaming connection to Hadoop instances (the open source implementation of MapReduce; Vertica is contributing the adapter to the community). This architecture mirrors usage patterns we’ve seen, and which Vertica asserts its customers have told them they want. One scenario: use your ADBMS to retrieve stored data, pass it to Hadoop for analysis by staff with different skill sets from the typical ADBMS users, and then bring result sets back. A separate hardware for the Hadoop sandbox is fairly typical among early adopters today, and via a Cloudera partnership, Vertica can offer a deployment architecture that doesn’t break the bank. Curt Monash does the usual excellent summary of Hadoop issues in his blog. [...]

]]> By: Daniel Abadi http://www.dbms2.com/2009/08/04/verticas-version-of-mapreduce-integration/#comment-133794 Daniel Abadi Tue, 04 Aug 2009 15:01:12 +0000 http://www.dbms2.com/?p=858#comment-133794 I agree with Omer's clarification. Also, just for the record, it's probably giving me too much credit to say that C-Store was my PhD thesis. My thesis involved research behind building the query execution engine for C-Store, but the C-Store project was much bigger than just the work that I did. I agree with Omer’s clarification.

Also, just for the record, it’s probably giving me too much credit to say that C-Store was my PhD thesis. My thesis involved research behind building the query execution engine for C-Store, but the C-Store project was much bigger than just the work that I did.

]]> By: Omer Trajman http://www.dbms2.com/2009/08/04/verticas-version-of-mapreduce-integration/#comment-133777 Omer Trajman Tue, 04 Aug 2009 12:12:59 +0000 http://www.dbms2.com/?p=858#comment-133777 One clarification regarding compute/data locality. MR necessarily has a data re-distribution phase prior to reduce (unless data is distributed by map key on load). When pushing the map down to Vertica there is no more data shuffling beyond what any other MR requires. You do get the added flexibility of being able to reduce on a different collection of nodes. One clarification regarding compute/data locality. MR necessarily has a data re-distribution phase prior to reduce (unless data is distributed by map key on load). When pushing the map down to Vertica there is no more data shuffling beyond what any other MR requires. You do get the added flexibility of being able to reduce on a different collection of nodes.

]]>