-Peregrine: A new map reduce framework for iterative/pipelined jobs.
I'm pleased to announce Peregrine 0.5.0 - a new map reduce framework
for iterative and pipelined map reduce jobs.
This originally started off with some internal work at Spinn3r to build a
and efficient Pagerank implementation. We realized that what we wanted was
runtime optimized for this type of work which differs radically from the
traditional Hadoop design.
Peregrine implements a partitioned distributed filesystem where key/value
are routed to defined partitions. This enables work to be joined against
previous iterations or different units of work by the same key on the same
Peregrine is optimized for ETL jobs where the primary data storage system
external database such as Cassandra, Hbase, MySQL, etc. Jobs are then run
Extract, Transform and Load stages with intermediate data being stored in
We enable features such as Map/Reduce/Merge as well as some additional
functionality like ExtractMap and ReduceLoad (in ETL parlance).
A key innovation here is a partitioning layout algorithm that can support
many to many recovery similar to HDFS but still support partitioned
with deterministic key placement.
We've also tried to optimize for single instance performance and use modern
primitives as much as possible. This includes NOT shying away from
specific features such as mlock, fadvise, fallocate, etc.
There is still a bit more work I want to do before I am ready to benchmark
against Hadoop. Instead of implementing a synthetic benchmark we wanted to
a production ready version first which would allow people to port existing
applications and see what the before / after performance numbers looked
the real world.
For more information please see:
As well as our design documentation:
Founder/CEO Spinn3r.com <http://spinn3r.com/>
Location: *San Francisco, CA*
Skype-in: *(415) 871-0687*