Hoffmann, Brandon2015-10-132015-10-132015-05https://hdl.handle.net/11299/174781University of Minnesota M.S.E.C.E. thesis. May 2015. Major: Electrical Engineering. Advisor: David Lilja. 1 computer file (PDF); viii, 58 pages.Recently the amount of data generated and stored on computers has seen outrageous growth and the trend will only continue. With the 24 hour global business structure being the way it is now, backup windows are shrinking and/or data is expected to be available at all times. Because of this, having effective and efficient data protection has become increasingly important. It is therefore necessary to move past the outdated static backup configurations and adopt intelligent dynamic backup systems. With that in mind we introduce the Affinity dynamic backup scheduling algorithm. Using a dynamic backup simulator we examine this algorithm as well as others and examine the performance trade-offs between these algorithms. Using this algorithm we have seen incremental improvements in the three primary metrics of Storage Throughput Utilization, Storage Distribution and Backup Time Consistency. With the insight gained from our simulation we discuss the benefits and trade-offs of dynamic scheduling algorithms and also dive into ideas and changes important to the future of backup systems.enbackupdynamicschedulingThesis or Dissertation