» VIEW ALL POSTS
Apr 2 2009 4:48PM GMT
Posted by: John M. Willis
ap reduce
In 140 characters tell me why map reduce is important?
Posted by: John M. Willis
ap reduce
This morning Lance Weatherby, a local Atlanta venture catalyst, asked a simple question on twitter – “In 140 characters tell me why map reduce is important please.” When I saw the question it made me think about the answer. All of us who cover the cloud-o-sphere take terms like map reduce for granted. However, unless you are working directly with cloud computing or you have a BS from Stanford, you probably won’t have a clue why every one is so excited about these terms. My initial one word twitter answer was parallelism. Then I added a second word - multi-core. However, I guess my answers were lazy ones compared to others. Here is a list of great answers posted this morning to Lance’s and my re-tweet question…
-
keithmcgreggor: @lance example: count all of the unique words in a huge set of documents by simultaneously counting in each document, then adding the result -
keithmcgreggor: @lance mapreduce is applicable to problems which admit parallelism readily. Difficult part is deciding how to carve up your data.
-
talltodd65: @Lance MapReduce is ideal for processing very large data sets on a cluster of computers -
pst555: @botchagalupe @lance As data is ever increasing I need to find a way to to see the trend I base my decision on. (proposal) -
chriscurtin: @Lance Data analytics for $20/run in EC2 vs. $2MM in TerraData? or unlimited ‘databases’ on cheap disk->point logic to data vs. BN rows RDBM -
-
ewantoo: @Lance @botchagalupe MapReduce is an cheap + fast way of breaking down 1 large process into 1000s of small ones which then run in parallel -
botchagalupe: rt @Lance In 140 characters tell me why map reduce is important please. ( Thoughts? )
-
-
vincezappa: @Lance MapReduce allows one to cost effectively process and analyze large amounts of data very quickly. Many applications of this. -
-
-
-
botchagalupe: @Lance Oh heck two words parallelism and multi-core
-
botchagalupe: @Lance one word .. parallelism
-
-
-
pst555: @botchagalupe @lance As data is ever increasing I need to find a way to to see the trend I base my decision on. (proposal)
-
rjurney: @botchagalupe With map/reduce we’re going to start making productive use of all that data lying around, informing nothing.
-
jrep: @botchagalupe Map-Reduce is a really powerful way to massively parallelize a really limited set of possibly important tasks -
ewantoo: @Lance @botchagalupe MapReduce is an cheap + fast way of breaking down 1 large process into 1000s of small ones which then run in parallel
-
botchagalupe: rt @Lance In 140 characters tell me why map reduce is important please. ( Thoughts? )
-
tjsweeney: @botchagalupe happy five oh -
botchagalupe: @Lance Oh heck two words parallelism and multi-core
-
botchagalupe: @Lance one word .. parallelism
-
Comment
RSS Feed
Email a friend
You must be logged-in to post a comment. Log-in/Register