HSLA: Heterogeneous storage-tier log analyzer over hadoop

Hadoop ecosystem processes extremely large datasets in a parallel computing environment. The Hadoop Distributed File System (HDFS) manages operational aspects of processed, unprocessed and log archives. Recently, HDFS has adopted heterogeneous environment, that enables file system to cope with storage-tier data processing. This increases the functional utilization of storage devices and distributes node capacity to storage-tier unevenly. Thus, a job having high priority is affected with delay latency and storage devices i.e. Disk, SSD and RAM consumes individual time overhead to release a non-priority job data. To analyze the complexity of storage-tier, we present Heterogeneous Storage-tier Log Analyzer (HSLA) strategy, that collects control and data flow events to a central repository and performs an analysis over log datasets. The analytics metrics consists of pre-emptive measures observed through events traces. The experimental results depict that, HSLA presents a broad aspect of storage-tier contingency problem and proposes a node computing capacity share strategy to balance functional processing of HDFS blocks over storage-tier.