Scaling Hadoop for MapReduce Applications Webinar
Source: Sun Microsystems
Hadoop is typically scaled on a large pool of commodity system nodes. However, by using multicore, multithreaded processors, one can achieve the same scale with fewer machines. The presenter of this webcast will discuss how Sun's chip multithreading (CMT) technology-based UltraSPARC T2 Plus processor can process up to 256 tasks in parallel within a single node. The presenter will also share with how sun has evaluated CPU and I/O throughput, memory size, and task counts to extract maximal parallelism per single node, as well as an evaluation of the performance.