We started off the week by announcing our new product offering – MapR M7. M7 brings enterprise grade features to HBase™ applications including instant node recovery, snapshots and mirroring while delivering unprecedented scale and performance to its database users.
The announcement was followed by a joint workshop with Google. The Google Compute Engine team first provided an overview of the service and the value Google Compute Engine brings to users. MapR then announced and provided a demonstration of the new Hadoop TeraSort World Record. MapR completed a 1 TB TeraSort in 54 seconds on 1003 instances of the Google Compute Engine. This record was not only faster than the previous record of 62 seconds, it was set with 46% less servers, 1/3 the cores and 1/6 the disks. More interestingly, the Google resources required amounted to $9 utilizing nodes over Google Compute Engine compared to over $5M in on-premise servers for the previous record.
Back at the conference, during his keynote speech, MapR CEO John Schroeder used the term “Datacosm” to describe the Big Data phenomenon. This description builds on George Gilder’s concepts of Microcosm and Telecosm. The keynote also introduced M7 to the participants along with a product demo. At the end, MapR’s cloud partner Google was on stage congratulating MapR on creating the new TeraSort world record over Google Compute Engine.
The MapR Booth at the conference saw a steady influx of interested and intrigued visitors who wanted to understand more about MapR and our latest innovation. Topics ranged from enterprise grade features to elimination of region servers and compactions. Work on Apache Drill was another popular topic among visitors.
The Strata conference was yet another example that further validates that Big Data and Hadoop are a reality for the enterprise and are here to stay. MapR is excited to be part of the Datacosm and we look forward to many more such events.