Daniel Myers
Profile
I am a fourth year graduate student studying analytic performance modeling and system design. My advisor is Mary Vernon.
I was raised in Kingsport, TN and Miami, FL.
I studied computer engineering at the University of Florida and earned my bachelor's and Master's degrees in 2004 and 2005. I conducted research on acoustics, remote sensing, and morphological neural networks.
From 2006 to 2008, I served as a senior member of technical staff at Sandia National Laboratories in Albuquerque, NM. I worked with a group that applied customized computer vision and signal processing algorithms to national security problems, with an emphasis on satellite imagery. My projects included target detection, image stitching, satellite data compression, and validation of reflective particle tags, a technology used to seal and secure nuclear monitoring equipment.
I left Sandia and came to UW in the Fall of 2008.
My hobbies include old-time and bluegrass guitar, shape note singing, and weight lifting. I won the 2007 New Mexico state flatpicking guitar championship and my playing was featured in the now-defunct Frets magazine in 2006.
I am interested in using analytic performance models to build better computer systems.
Modeling techniques provide a fast and cost-efficient way to identify the most important elements driving system performance. Validated models can be used to explore tradeoffs in the design process, search for bugs in a system, or evaluate proposed design changes. Analytic modeling techniques provide an alternative to time-consuming simulations and complicated benchmarking experiments.
With Mary Vernon, I am working on applying modeling techniques to the design of next-generation storage systems, with an emphasis on datacenter workloads and cloud computing. This project has three major elements:
- Identifying the characteristics of modern large-scale datacenter workloads
- Developing models that accurately predict the performance of existing storage systems on these workloads
- Using models to design new systems that achieve superior performance with lower cost and energy consumption
In 2011, I worked as an intern at Google, where I developed analytic models for the low-level components in their storage infrastructure.
I have also done some work on the theory of analytic modeling, developing new approximation techniques for queue length probability distributions and bounds for tail latencies in queueing networks.
I will be teaching CS 547: Computer System Modeling Fundamentals in Spring 2012. This course is an introduction to analytic performance modeling. Topics include the fundamental laws of system dynamics, bottleneck analysis, probability, and queueing networks.
The Wisconsin Center for Academically Talented Youth (WCATY), an affiliate of the UW School of Education, runs several online and summer programs for talented students from around the state and region. In 2009 and 2010, I taught an introductory college-level CS for WCATY's accelerated summer learning program. The course packed seven hours of instruction into each day (two hours of lectures and five hours of Java programming practice) and covered all of the topics on the AP Computer Science syllabus in only three weeks. We are planning to offer this course again in 2012.
In 2011, the CS department partnered with WCATY to offer a new Computer Science Research Internship program. Six Madison high schoolers partnered with CS department faculty to learn about computer science and conduct research projects. I served as the research coordinator for the program and taught a two week introductory "boot camp" for students without programming experience.