Brandon Salmon

Introduction    

I am currently a Ph.D. candidate at Carnegie Mellon University.  I work in the Parallel Data Lab with my advisor  Greg Ganger.

My research interests include distributed file systems, information dissemination mechanisms and user centered design. I am interested in tuning the architectures of such systems so that they are easy for users to understand, and fit well into social systems. I have focused recently on distributed storage systems for the home.

My work involves a combination of such system design and then ethnographically-inspired explorations of the affects of architecture changes on the use of such systems in deployment. While the user design communities and system architecture communities have traditionally been far from one another, I feel this research is critical. Too frequently usability experts are stuck fixing architectures that are fundamentally difficult to understand and manipulate, because system architects are primarily concerned with performance.

Contact Info

  • Office: CIC 2220A
  • Phone: 412-268-5854
  • Fax: 412-268-677
  • Mailing address:
    ECE - Carnegie Mellon Univ.
    5000 Forbes Ave.
    Pittsburgh, PA 15213
  • Email: bsalmon@ece.cmu.edu

Resume

Current Projects

Perspective

The home digital environment is an area of growth and change in the world of computing.  Digital music players, DVDs, cameras and digital video recorders have converted much of the remaining analog media in the home into the digital realm.  Managing this burgeoning amount of data and devices will be difficult and time consuming.  However, home users should not be expected to spend time dealing with this kind of management.  My work focuses on automating the replication and migration of data, so that it will be both reliable and available at devices where it is needed without requiring much user input.

Self-*

Administration is currently a large part of the total cost of ownership of enterprise storage.  The Self-* Storage project focuses on designing storage systems that decrease the need for administrator work in data management.  We have built a versatile brick-based storage infrastructure to support automation of various administrative tasks.  For example, one planning task is choice of data encoding, which heavily affects performance and reliability.  I am building a mechanism to automatically choose an appropriate encoding when an object is first created in the system.  I have also researched methods for administrators to specify high-level objectives to the system, to allow the system make mechanism and policy decisions.

Previous Projects

Continuous Reorganization

The way data is laid out on disk has a huge impact on system performance.  For this reason, researchers have proposed various heuristics to optimize for particular data access patterns.  Each of these heuristics works well for some data, but may be useless or detrimental to other data.  As a result, few are used in practice.  To enable their safe use for data-specific layout optimization, we proposed a two-tiered system that takes a variety of heuristics and adaptively combines them using simulation and optimization methods to place each piece of data using a correct heuristic.  Robustly combining heuristics proved to be difficult, providing us insight into the complexity of even relatively small automation problems, such as this one.

Freeblock Scheduling

To perform adaptation like layout optimization, a system must be able to efficiently service background tasks.  Freeblock scheduling allows a system to perform background disk operations without cost to a foreground workload, even if the disk is currently 100% busy.  It obtains this extra bandwidth by scheduling background operations in the rotational latency of the disk, which would otherwise be wasted.  I worked specifically on modifying background tasks, such as data migration, to work well with the freeblock system.

Papers

Learning to Share: A Study of Sharing Among Home Storage Devices. Brandon Salmon, Frank Hady, Jay Melican. Carnegie Mellon University Parallel Data Lab Technical Report CMU-PDL-07-107, October, 2007. Abstract / PDF

Putting Home Storage Management into Perspective. Brandon Salmon, Steven W. Schlosser, Lily B. Mummert, Gregory R. Ganger. Carnegie Mellon University Parallel Data Lab Technical Report CMU-PDL-06-110, September 2006. Abstract / PDF

Early Experiences on the Journey Towards Self-* Storage.. 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 29(3). Special issue on self-managing database systems. September 2006.Abstract / PDF

Relative fitness models for storage. Michael Mesnier, Matthew Wachs, Brandon Salmon, Gregory R. Ganger. SIGMETRICS Performance Evaluation Review 33(4): 23-28 (2006). Abstract / PDF

End-to-end tracing in a cluster-based storage system: a feasibility study. Eno Thereska, Brandon Salmon, John Strunk, Matthew Wachs, Michael Abd-El-Malek, Julio Lopez, Gregory R. Ganger. Proceedings of the Joint International Conference on Measurement and Modeling of Computer Systems (SIGMETRICS'06), Saint-Malo, France. June 26th-30th 2006. Abstract / PDF

Ursa Minor: versatile cluster-based storage. Michael Abd-El-Malek, William V. Courtright, Chuck Cranor, Gregory R. Ganger, James Hendricks, Andrew J. Klosterman, Michael Mesnier, Manish Prasad, Brandon Salmon, Raja R. Sambasivan, Shafeeq Sinnamohideen, John Strunk, Eno Thereska, Matthew Wachs, Jay J. Wylie. Proceedings of the 4th USENIX Conference on File and Storage Technologies (FAST '05), San Francisco, CA. December 13-16 2005. Abstract / PDF  [490K] Best paper award

Challenges in Building a Two-Tiered Learning Architecture for Disk Layout. Brandon Salmon, Eno Thereska, Craig A.N. Soules, John D. Strunk, Gregory R. Ganger. Carnegie Mellon University Parallel Data Laboratory Technical Report CMU-PDL-04-109. August, 2004. Abstract / Postscript [6.8M] / PDF [344K]

Design and Implementation of a Freeblock Subsystem. Eno Thereska, Jiri Schindler, Christopher R. Lumb, John Bucy, Brandon Salmon, Gregory R. Ganger. Carnegie Mellon University Technical Report CMU-PDL-03-107. December 2003. PDF [160K]  Best student paper award

A Framework for Building Unobtrusive Disk Maintenance Applications. Eno Thereska, Jiri Schindler, John Bucy, Brandon Salmon, Christopher R. Lumb, Gregory R. Ganger. Proceedings of the 3rd USENIX Conference on File and Storage Technologies (FAST '04). San Francisco, CA. March 31, 2004. Abstract / Postscript [5.1M] / PDF [148K]

A Two-Tiered Software Architecture for Automated Tuning of Disk Layouts. Brandon Salmon, Eno Thereska, Craig A.N. Soules, Gregory R. Ganger. First Workshop on Algorithms and Architectures for Self-Managing Systems. In conjunction with Federated Computing Research Conference (FCRC). San Diego, CA. June 11, 2003. Also published as Carnegie Mellon University SCS Technical Report CMU-CS-03-130.  Abstract / Postscript [718K] / PDF [405K]

Classes

  • Valuations
  • Accounting
  • Engineering Entrepreneurship
  • Optimization
  • Topics in Systems and Control: Large Scale Distributed Systems
  • Graduate Algorithms
  • File System Survey
  • Advanced Operating and Distributed Systems
  • Special Topics in Cryptography and Network Security