PROJECT TITLE :
Analyzing Enterprise Storage Workloads With Graph Modeling and Clustering
Utilizing graph analysis models and algorithms to exploit advanced interactions over a network of entities is rising as an attractive network analytic technology. During this paper, we have a tendency to show that traditional column or row-based mostly trace analysis may not be effective in deriving deep insights hidden in the storage traces collected over complicated storage applications, like complex spatial and temporal patterns, hotspots and their movement patterns. We tend to propose a completely unique graph analytics framework, GraphLens, for mining and analyzing universe storage traces with three distinctive options. 1st, we model storage traces as heterogeneous trace graphs so as to capture multiple complex and heterogeneous factors, like diverse spatial/temporal access data and their relationships, into a unified analytic framework. Second, we use and develop an innovative graph clustering technique that employs two levels of clustering abstractions on storage trace analysis. We discover interesting spatial access patterns and establish vital temporal correlations among spatial access patterns. This allows us to better characterize vital hotspots and understand hotspot movement patterns. Third, at every level of abstraction, we tend to style a unified weighted similarity live through an iterative dynamic weight learning algorithm. With an optimal weight assignment scheme, we have a tendency to can efficiently mix the correlation data for each type of storage access patterns, like random versus sequential, scan versus write, to spot interesting spatial/temporal correlations hidden in the traces. Some optimization techniques on matrix computation are proposed to additional improve the efficiency of our clustering algorithm on large trace datasets. In depth evaluation on real storage traces shows GraphLens can offer broad and deep trace analysis for higher storage strategy coming up with and efficient knowledge placement guidance. GraphLens will be applied to both a single LAPTOP with multiple disks and a di- tributed network across a cluster of compute nodes to offer a few opportunities for optimization of storage performance.
Did you like this research project?
To get this research project Guidelines, Training and Code... Click Here