Abstract :
Ensuring low replication latency of database events is business-critical but challenging with big data. A proposed capacity-planning model helps achieve this goal by forecasting future traffic rates, predicting replication latency, and determining required replication capacity. The Web extra at http://youtu.be/ZupPlrS8dGA is a video of in which author Zhenyun Zhuang demonstrates Naarad, an open-source performance analysis tool (https://github.com/linkedin/naarad) written in python that analyzes various metrics (gc, sar, Jmeter etc), evaluates SLAs and generates a user friendly report to aid in performance analysis and investigations.
Keywords :
Big Data; database management systems; public domain software; software metrics; software performance evaluation; Naarad; SLA; big data events; capacity-planning model; database events; future traffic rates; low replication latency; open-source performance analysis tool; python; replication capacity; replication latency; user friendly report; Big data; Capacity planning; Data models; Predictive models; Soci; ARIMA; Internet/Web technologies; LinkedIn; SPARQL; autoregressive integrated moving average; big data; capacity planning; database replication; graph databases; high-performance computing; multithreading; replication latency; time-series decomposition RDF databases; traffic rate forecasting;
