MapR Technologies has updated its M7 edition to improve HBase application performance with throughput that is 4-10x faster while eliminating latency spikes.
NOTE: HBase is a column-oriented DBMS built to run on top of the Hadoop Distributed File System (HDFS) — according to IBM it is "well suited" for sparse data sets, which are common in many big data use cases.
IBM further explains, "Unlike relational database systems, HBase does not support a structured query language like SQL; in fact, HBase isn't a relational data store at all. HBase applications are written in Java much like a typical MapReduce application."
The 20 Millisecond Factor
HBase applications can now benefit from MapR's platform to address one of the major issues for online applications, consistent read latencies in the "less than 20 millisecond" range, as they exist across varying workloads.
"Our customers are moving Hadoop from pilot adoption and project use to mainstream enterprise deployments," said John Schroeder, CEO and cofounder, MapR Technologies. "MapR customers are experiencing the same reliability and enterprise-level performance with our distribution as they have seen with the Oracle platform at a fraction of the cost."
The firm says that customers have been migrating to M7 from Oracle, MySQL, and other NoSQL databases.
Differentiated features here include architecture that persists table structure at the filesystem layer; no compactions (I/O storms) for HBase applications; workload-aware splits for HBase applications; direct writes to disk (vs. writing to an external filesystem); disk and network compression; and C++ implementation that does not suffer from garbage collection problems seen with Java applications.