any interest ?

[rjr3]

anyone .

[Gravis]

i think the forum should be renamed MapReduce because Apache Hadoop is just the name of an implementation of MapReduce. Hadoop is written in Java and there are other implementations of MapReduce like and which are both in C++. i would go with Phoenix++ because it doesn't have any boost dependencies.

an Epiphany asm port would be really nice so that you could get highest performance possible. i know would be interested in it.

[9600]

[mackrorysd]

I've also been curious to know what Hadoop interest there may be. My interest in Parallella is more for other languages and frameworks, but I work primarily with Apache Bigtop (integration and distribution for the Hadoop ecosystem). If there is interest in using Hadoop on Parallella clusters I would consider volunteering some time to get better ARM support in Bigtop. The packages for Hadoop itself are architecture-dependent and would need a bit of work, but ZooKeeper (cluster coordination service) should work out of the box already.

[RJoshi]

I am interested in hadoop/hbase. Only concern is limited memory.

[afbrandt]

I'm interested in contributing to/starting up a port of mapreduce. I'm actually a student and was looking for a way to combine my capstone project with some open source work. I've started looking over the Epiphany sdk and the referenced MapReduce implementations, and I'd like to know where this project stands. Thanks!

[9600]

[mokshafaced]

I would love to contribute, still waiting for my two cards... I'd also like to work on a process to parallelize R and Clojure tasks/programs across the cores.

[pegasis]

will Raspbian wheezy from raspberry pi run on Parallella?

if so this might be interesting.. these guys are running Hadop on raspberry Pi(s)


http://blog.ittoby.com/2013/08/starting ... mpute.html


use the Parallella to do the compute and the 2-3 PIs with attached SSD drives for storage


another way:

http://www.michael-noll.com/tutorials/r ... e-cluster/

any feedback

[9600]