Visualizing request-flow comparison to aid performance diagnosis in distributed systems

Conference papersJournal papers
Raja R. Sambasivan, Ilari Shafer, Michelle Mazurek, Gregory R. Ganger
IEEE Transactions on Visualization and Computer Graphics (Proc. Information Visualization 2013), Vol. 19, no. 12, Dec. 2013
Publication year: 2013

Distributed systems are complex to develop and administer, and performance problem diagnosis is particularly challenging. When performance degrades, the problem might be in any of the system’s many components or could be a result of poor interactions among them. Recent research efforts have created tools that automatically localize the problem to a small number of potential culprits, but research is needed to understand what visualization techniques work best for helping distributed systems developers understand and explore their results.  This paper compares the relative merits of three well-known visualization approaches (side-by-side, diff, and animation) in the context of presenting the results of one proven automated localization technique called {\it request-flow comparison}.  Via a 26-person user study, which included real distributed systems developers, we identify the unique benefits that each approach provides for different problem types and usage modes.

Relative fitness modeling

Journal papers
Michael P. Mesnier, Matthew Wachs, Raja R. Sambasivan, Alice X. Zheng, Gregory R. Ganger
Communications of the ACM 52, 4 (April 2009), 91-96
Publication year: 2009

Relative fitness is a new approach to modeling the performance of storage devices (e.g., disks and RAID arrays). In contrast to a conventional model, which predicts the per- formance of an application’s I/O on a given device, a relative fitness model predicts performance differences between devices. The result is significantly more accurate predictions.

Early experiences on the journey to self-* storage

Journal papers
Michael Abd-El-Malek, William V. Courtright II, Chuck Cranor, Gregory R. Ganger, James Hendricks, Andrew J. Klosterman, Michael Mesnier, Manish Prasad, Brandon Salmon, Raja R. Sambasivan, Shafeeq Sinnamohideen, John D. Strunk, Eno Thereska, Matthew Wachs, Jay J. Wylie
IEEE Data Engineering Bulletin Volume 29, Number 3, September 2006
Publication year: 2006

Self-* systems are self-organizing, self-configuring, self-healing, self-tuning and, in general, self- managing. Ursa Minor is a large-scale storage infrastructure being designed and deployed at Carnegie Mellon University, with the goal of taking steps towards the self-* ideal. This paper discusses our early experiences with one specific aspect of storage management: performance tuning and projection. Ursa Minor uses self-monitoring and rudimentary system modeling to support analysis of how system changes would affect performance, exposing simple What…if query interfaces to administrators and tuning agents. We find that most performance predictions are sufficiently accurate (within 10-20%) and that the associated performance overhead is less than 6%. Such embedded support for What…if queries simplifies tuning automation and reduces the administrator expertise needed to make acquisition decisions.