Comments on: LexisNexis open-sources its Hadoop killer

By: Neuromancer

Neuromancer — Wed, 22 Jun 2011 14:26:12 +0000

Looks very cool am looking at using this to replace mahourt as the mahout devs seem unable to get a flipping stable relese out there.

By: Chris Albee

Chris Albee — Fri, 17 Jun 2011 14:21:30 +0000

Davor, there is an ECL Programmers Guide available at http://www.lexisnexis.com/risk/about/guides/program-guide.html .

By: George L'Heureux

George L'Heureux — Fri, 17 Jun 2011 13:51:14 +0000

Sorry, wrong link. This is the one you want to check out. http://en.wikipedia.org/wiki/ECL,_data-centric_programming_language_for_Big_Data

By: George L'Heureux

George L'Heureux — Fri, 17 Jun 2011 13:46:38 +0000

There’s a Wikipedia article available here: http://en.wikipedia.org/wiki/ECL_programming_language

By: Tim Pham

Tim Pham — Thu, 16 Jun 2011 16:24:30 +0000

You have no idea what’s going with JVM in current state. It can actually beat C++ on creating new object with just 10 CPU instructions.

5000 transactions/second is not a lot, but it surely depends on context.

This HPCC System is not a Hadoop Killer in any mean if it can’t be easily extended and modified, like Hadoop. That’s what truly good open-source system is about.
But anyway, We’re always welcomed any competitive solution in the open-source world.

By: Matt Benati

Matt Benati — Thu, 16 Jun 2011 14:15:37 +0000

Derrick,

You point out a critical issue facing the analytics market – complexity. ECL and other technologies will move the market forward and allow analytics to become more pervasive across the enterprise. Simplicity is key for wide-spread adoption, especially by the business users.

Matt Benati
Senior Director, Analytics
IBM Netezza

By: Jonathan Burger

Jonathan Burger — Thu, 16 Jun 2011 12:41:35 +0000

I think you’ll find the differences between the enterprise and community versions here. http://www.hpccsystems.com/products-and-services/products/ee-ce-comparison

By: Mark

Mark — Thu, 16 Jun 2011 06:35:46 +0000

Another open source competitor worth keeping an eye on is Nokia’s Disco, which has an Erlang framework (ideal for dealing with huge numbers of parallel processes), while the user provides the map and reduce functions in Python. See: http://discoproject.org/

By: Mihaela Mihaljevic Jakic

Mihaela Mihaljevic Jakic — Thu, 16 Jun 2011 01:27:05 +0000

“HPCC is written in C++”

I think that this is a killer feature over Hadoop. A major one.

By: Richard Chapman

Richard Chapman — Wed, 15 Jun 2011 21:58:29 +0000

Yes, Roxie uses the same ECL language.