MapReduce Improvements in the MapR Hadoop Distribution

Hi everyone,

Sorry for the late notice but we have scheduled the next Big Data meetup. See below for more information on the topic and the presenter. Please note that we have moved the meetup from our regular Tuesday to Thursday to accommodate the speaker. Future meetups will be on Tuesdays. Second, please note that we have changed our venue to the new Central Library. Matthew toured the library and the new meeting room (room 302) seats 125 people and has fast internet so we are excited about the new location and all future meetups will be anticipated to be at this location.

I hope to see everyone there!

Cheers, Pitt Fagan

MapR is the only Hadoop distribution that provides full data protection, no single points of failure, improved performance, and dramatic ease of use advantages. Adam will dive into the MapR features that provide improved performance and reliability over vanilla Hadoop. Basic outline:

• Hadoop/MapReduce background
• MapRFS vs. HDFS: FileSystem HA, Direct Access NFS, Volumes, Snapshots, Mirroring
• MapReduce improvements: DirectShuffle, JT HA, ExpressLane
• Beyond MapReduce: M7, YARN, etc.

Bio: Adam Bordelon, Senior Software Engineer, MapR Technologies
Adam is the lead developer on MapR's Hadoop/MapReduce team, with responsibilities including MapReduce optimizations and YARN integration. Prior to working at MapR, Adam was at Amazon developing recommendation algorithms on Hadoop and redesigning Amazon's distributed behavioral data service. Before Amazon, he re-engineered the National Instruments LabVIEW dataflow compiler for large-scale multicore systems. Adam earned a Master of Science in Computer Science degree from Rice University, with emphasis on performance optimization in High-Performance Computing (HPC) environments.