University of Saskatchewan - Computer Science
University of Saskatchewan
Higher Education
Dwight
Makaroff
Saskatchewan, Canada
I have been teaching at universities and colleges for 15 of the last 20 years, and doing research in networks and distributed systems.
Specialties: undergrad education, grad student supervision, conference program committee work, simulation, measurement, operating systems, sensor networks, distributed application performance debugging.
Associate Professor
Graduate and undergraduate teaching. Research in computer systems and computer networks. Administrative committee work.
Professor
Teaching, Research, and Administration.
Teaching Operating Systems, Programming Principles and Practices, Computer Systems Organization, Computer Networks, Computer Performance Evaluation, Advanced Parallel and Distributed Systems.
Course coordinator for CMPT 400/405 Honours for many years.
Chair of Academic Programs Committee (College of Arts/Science, Division of Science), University council, Academic Programs Committee (College of Graduate Studies and Research).
Department Curriculum Committee, Department Undergrad Committee, Department Research Seminar Coordinator at times in the past years.
I have been the Graduate Committee chairperson for 2013-2014 and 2014-2015.
In 2015-2016, I was on sabbatical at Google (Mountain View, CA) and DATA61/CSIRO (Sydney, Australia).
Member of University Review Committee (2018-2021).
Visiting Researcher
Community PVR (Distributed Media Caching using local Incentives)
Visiting Research Scientist
Content Management for Wearable Computers (sensors and distribution)
Visiting Researcher
Intrusion Detection in Wireless Mesh Networks
Visiting Researcher
Project for Sabbatical - Disk Tail Latency Performance Measurement
B. Comm.
Computational Science
BComm in Computational Science
M. Sc.
Computer Science
MSc in Computational Science
Associate Professor
Graduate and undergraduate teaching. Research in computer systems and computer networks. Administrative committee work.
Professor
Teaching, Research, and Administration.
Teaching Operating Systems, Programming Principles and Practices, Computer Systems Organization, Computer Networks, Computer Performance Evaluation, Advanced Parallel and Distributed Systems.
Course coordinator for CMPT 400/405 Honours for many years.
Chair of Academic Programs Committee (College of Arts/Science, Division of Science), University council, Academic Programs Committee (College of Graduate Studies and Research).
Department Curriculum Committee, Department Undergrad Committee, Department Research Seminar Coordinator at times in the past years.
I have been the Graduate Committee chairperson for 2013-2014 and 2014-2015.
In 2015-2016, I was on sabbatical at Google (Mountain View, CA) and DATA61/CSIRO (Sydney, Australia).
Member of University Review Committee (2018-2021).
Ph. D.
Computer Science
PhD in Computer Science
ICPE 2011
ICPE 2011
IBM CASCON 2014
Hadoop task schedulers like Fair Share and Capacity have been specially designed to share hardware resources among multiple organizations. The Capacity Scheduler provides a complex set of parameters to give fine control over resource allocation of a shared MapReduce cluster. Administrators and users often run into performance problems because they do not know the meaning of different task scheduler parameters and the impact they can have on the running MapReduce workloads across different organizations. The interaction between parameter settings is particularly problematic. In this paper, we implemented a Capacity Scheduler simulator component, integrated it into an existing simulator and then validated the scheduler performance in the simulator with small test cases, consisting of standard benchmark sort programs with different resource requirements. The next step was to study the impact of Capacity Scheduler parameters on different MapReduce workload submission patterns with a more complex set of benchmark programs. Among other results, we found maxCapacity and minUserLimitPCT to be influential parameters as was indicated in our previous work and that using separate queues for short and long jobs provides the best performance in terms of response ratio, execution time and makespan compared to submitting both short and long jobs in the same queue.