<span class="gmail_quote"><span class="e" id="q_1149dbc436a2040b_0"><span class="gmail_quote">On 8/25/07, <b class="gmail_sendername">Isaac Levy</b> <<a href="mailto:ike@lesmuug.org" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">
ike@lesmuug.org</a>> wrote:</span><blockquote class="gmail_quote" style="margin-top: 0; margin-right: 0; margin-bottom: 0; margin-left: 0; margin-left: 0.80ex; border-left-color: #cccccc; border-left-width: 1px; border-left-style: solid; padding-left: 1ex">
Hey All,<br><br>This *may* be a thread which is aimed at the wrong list, but I<br>thought it's appropriate. Feel free to yell at me to take the thread<br>elsewhere.<br><br>--<br>Pretext:<br>I've been working on a personal project lately which has landed me
<br>with some home-scale monster number-crunching tasks, as well as some<br>quickly scaling massive storage requirements.<br>(Fun fun fun, I'm trying to scan and OCR my personal book collection,<br>and I'm getting scared out of my mind now that I'm making some
<br>headway :)<br><br>Anyhow, I've been looking really closely at Google's MapReduce system/<br>algorithm spec, which seems to be at the heart of how they make their<br>massive clusters work. This seems to be the current hot topic in
<br>macho computing.<br><a href="http://en.wikipedia.org/wiki/MapReduce" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">http://en.wikipedia.org/wiki/MapReduce</a><br><br>Fun for UNIX folks, Rob Pike made an awk-like utility/language called
<br>'Sawzall' which uses Google's internal MapReduce API- I think it's <br>pretty interesting.<br><a href="http://labs.google.com/papers/sawzall.html" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">
http://labs.google.com/papers/sawzall.html</a><br><br>With that, I've also found that Yahoo is putting massive support into<br>an implementation of the MapReduce idea, Open Source as a part of the <br>Apache Project:<br>
<a href="http://lucene.apache.org/hadoop/" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">http://lucene.apache.org/hadoop/</a><br><br>There's other implementations cropping up all over, it seems. Like I
<br>said, it's the current buzz... <br><br>With all that, many of us have noticed that Google is good at scaling<br>patterns in computing- to bastardize their whole tech. operation,<br>it's their big trick. When I say scaling patterns, I mean: applying
<br>classical computing paradigms and methodology at wild scales. (E.G.<br>with the Google Filesystem, it's simply a filesystem where the disk<br>blocks exist as network resources, etc... You see what I mean with<br>
scale?)<br><br><br>--<br>Question:<br><br>Anyhow, I'm looking for more patterns in this MapReduce stuff,<br>because I'm simply not one to dive headfirst into 'buzz-tech'. With<br>that, aside from the map, and reduce, functions found in many
<br>programming languages,<br><a href="http://en.wikipedia.org/wiki/Map_%28higher-order_function%29" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">http://en.wikipedia.org/wiki/Map_%28higher-order_function%29
</a><br><a href="http://en.wikipedia.org/wiki/Fold_%28higher-order_function%29" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)"> http://en.wikipedia.org/wiki/Fold_%28higher-order_function%29</a><br>
<br>can anyone shed some light on similar prior works in distributed<br>computing and RPC systems which are 'old classics' in UNIX? These<br>distributed computing problems simply can't be new. <br><br>To be really straight, what I'm getting at, is why is this more or
<br>less useful than intelligently piping commands through ssh? What<br>about older UNIX rpc mechanisms? Aren't there patterns in even<br> kernel source code which match this work, or are even computationally<br>more sophisticated and advanced?
<br><br> From kernel to userland to network, I'm dying to find similar works,<br>any help is much appreciated!<br><br>Rocket-<br>.ike<br><br><br>---<br>p.s.<br>If anyone is interested in book-scanning stuff, Google happens to
<br>currently host the Open Source 'tessarect' project, a very nice OCR<br>software- a very clean command-line application for OCR processing. <br>(I was actually inspired to start all this stuff based on how much
<br>fun I had screwing around with tessarect). Apache licence, /me shrugs.<br><a href="http://code.google.com/p/tesseract-ocr/" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">http://code.google.com/p/tesseract-ocr/
</a><br><br><br>_______________________________________________<br>% NYC*BUG talk mailing list<br><a href="http://lists.nycbug.org/mailman/listinfo/talk" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">
http://lists.nycbug.org/mailman/listinfo/talk</a><br>%Be sure to check out our Jobs and NYCBUG-announce lists <br>%We meet the first Wednesday of the month<br></blockquote><br></span></span>You might want to look into Starfish, which is a MapReduce implementation for Ruby.
<br>( <a href="http://rufy.com/starfish/doc/" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">http://rufy.com/starfish/doc/ </a> ) Should be greatly simpler than dealing with Hadoop<br>(but thats just my personal opinion).
<br>