Simplified Cloud Control Using Dimension Reduction
Automated management of complex information technology applications such as cloud systems requires dynamic configuration of both application-level and system-level parameters. The existence of large number of tunable parameters makes it difficult to design a feedback controller that adjusts these parameters effectively in order to achieve the application-level performance targets. In this talk, we will summarize our recent work which introduces a new approach for simplified control architecture of large-scale complex systems based on dimension reduction techniques. It combines the online selection of critical control knobs through LASSO -- a powerful L1 -constrained fitting method, and Compressive Sensing (CS)-- a L1-optimization method, and the design of adaptive control of the identified knobs. We use evaluation results to demonstrate the effectiveness of this new approach.