basic queueing theory
play

Basic Queueing Theory CS 450 : Operating Systems Michael Saelee - PowerPoint PPT Presentation

Basic Queueing Theory CS 450 : Operating Systems Michael Saelee <lee@iit.edu> Agenda - Queueing theory? Huh? - Probability refresher / Crash course - Queueing theory & Kendalls notation - Mean value analysis of basic queues


  1. Basic Queueing Theory CS 450 : Operating Systems Michael Saelee <lee@iit.edu>

  2. Agenda - Queueing theory? Huh? - Probability refresher / Crash course - Queueing theory & Kendall’s notation - Mean value analysis of basic queues

  3. § Queueing Theory? Huh?

  4. Remember, we started our discussion of scheduling at a high level — “policy” - mostly described heuristics -based (i.e., hand-wavy) approaches

  5. to obtain empirical data, we can examine a “live” operating system’s scheduler

  6. to exercise rigor , should leverage some branch of mathematics well-suited to analyzing scheduling systems … queueing theory!

  7. queueing theory models wait queues using probability theory - e.g., arrival/service rate distributions - supports mathematical analysis & rigor

  8. wide application: - checkout lines - telecom switch - traffic light system - network quality of service

  9. we’ll barely scratch the surface — queueing theory is an area ripe for research — but you’ll see some basic applications - will also help explain underpinnings of simulators used for scheduling!

  10. § Probability refresher / Crash course

  11. Probability theory = quantitative analysis of random phenomena - assign a weighted probability to every event in a sample space ( Ω ) - use these probability distributions to better understand the behavior of the phenomena

  12. Core abstraction: random variable - a R.V . is a function that maps the sample space onto numeric values (e.g., X: Ω → ℝ ) - discrete R.V .s map to a countable set - continuous R.V .s map events onto an uncountable set 
 (e.g., real-valued)

  13. E.g., double coin toss (discrete event space) Ω = { TT, TH, HT, HH }  0 , if ω = TT   1 , if ω = TH  X ( ω ) = 2 , if ω = HT   3 , if ω = HH 

  14. Typically interested in a variety of statistics of random variables (and corresponding events): - probability of event n : P ( X = n ) (or p ( n ) ) - expected value (mean): E ( X ) - variance: σ 2 ( X ) ; standard deviation: σ ( X ) - coefficient of variance: C X = σ ( X )/ E ( X ) 
 (unitless measure)

  15. <latexit sha1_base64="3hdbwTSVxBON+FiA5gMGQhY0lE=">ACD3icbZDLSgMxFIYz9VbrbdSlm2BRKgxlptbqplB047KCvUA7lEyaUMzmSHJCGXoG7jxVdy4UMStW3e+jWk7C209kPDx/+eQnN+LGJXKtr+NzMrq2vpGdjO3tb2zu2fuHzRlGAtMGjhkoWh7SBJGOWkoqhpR4KgwGOk5Y1upn7rgQhJQ36vxhFxAzTg1KcYKS31zN6oV3lZ9WoK+uLxBOnElSmViQVx2rZJ1bZevCqvTMvF20ZwWXwUkhD9Kq98yvbj/EcUC4wgxJ2XHsSLkJEopiRia5bixJhPAIDUhHI0cBkW4y2cCT7TSh34o9OEKztTfEwkKpBwHnu4MkBrKRW8q/ud1YuVfuQnlUawIx/OH/JhBFcJpOLBPBcGKjTUgLKj+K8RDpENROsKcDsFZXHkZmqWio/munK9dp3FkwRE4BgXgEtQA7egDhoAg0fwDF7Bm/FkvBjvxse8NWOkM4fgTxmfP/s6mMs=</latexit> <latexit sha1_base64="3hdbwTSVxBON+FiA5gMGQhY0lE=">ACD3icbZDLSgMxFIYz9VbrbdSlm2BRKgxlptbqplB047KCvUA7lEyaUMzmSHJCGXoG7jxVdy4UMStW3e+jWk7C209kPDx/+eQnN+LGJXKtr+NzMrq2vpGdjO3tb2zu2fuHzRlGAtMGjhkoWh7SBJGOWkoqhpR4KgwGOk5Y1upn7rgQhJQ36vxhFxAzTg1KcYKS31zN6oV3lZ9WoK+uLxBOnElSmViQVx2rZJ1bZevCqvTMvF20ZwWXwUkhD9Kq98yvbj/EcUC4wgxJ2XHsSLkJEopiRia5bixJhPAIDUhHI0cBkW4y2cCT7TSh34o9OEKztTfEwkKpBwHnu4MkBrKRW8q/ud1YuVfuQnlUawIx/OH/JhBFcJpOLBPBcGKjTUgLKj+K8RDpENROsKcDsFZXHkZmqWio/munK9dp3FkwRE4BgXgEtQA7egDhoAg0fwDF7Bm/FkvBjvxse8NWOkM4fgTxmfP/s6mMs=</latexit> <latexit sha1_base64="3hdbwTSVxBON+FiA5gMGQhY0lE=">ACD3icbZDLSgMxFIYz9VbrbdSlm2BRKgxlptbqplB047KCvUA7lEyaUMzmSHJCGXoG7jxVdy4UMStW3e+jWk7C209kPDx/+eQnN+LGJXKtr+NzMrq2vpGdjO3tb2zu2fuHzRlGAtMGjhkoWh7SBJGOWkoqhpR4KgwGOk5Y1upn7rgQhJQ36vxhFxAzTg1KcYKS31zN6oV3lZ9WoK+uLxBOnElSmViQVx2rZJ1bZevCqvTMvF20ZwWXwUkhD9Kq98yvbj/EcUC4wgxJ2XHsSLkJEopiRia5bixJhPAIDUhHI0cBkW4y2cCT7TSh34o9OEKztTfEwkKpBwHnu4MkBrKRW8q/ud1YuVfuQnlUawIx/OH/JhBFcJpOLBPBcGKjTUgLKj+K8RDpENROsKcDsFZXHkZmqWio/munK9dp3FkwRE4BgXgEtQA7egDhoAg0fwDF7Bm/FkvBjvxse8NWOkM4fgTxmfP/s6mMs=</latexit> <latexit sha1_base64="ck8pdC+ekZH4nUmSP+ZG7r8lEyk=">AB2XicbZDNSgMxFIXv1L86Vq1rN8EiuCozbnQpuHFZwbZCO5RM5k4bmskMyR2hDH0BF25EfC93vo3pz0JbDwQ+zknIvSculLQUBN9ebWd3b/+gfugfNfzjk9Nmo2fz0gjsilzl5jnmFpXU2CVJCp8LgzyLFfbj6f0i7+gsTLXTzQrMr4WMtUCk7O6oyaraAdLMW2IVxDC9YaNb+GS7KDUJxa0dhEFBUcUNSaFw7g9LiwUXUz7GgUPNM7RtRxzi6dk7A0N+5oYkv394uKZ9bOstjdzDhN7Ga2MP/LBiWlt1EldVESarH6KC0Vo5wtdmaJNChIzRxwYaSblYkJN1yQa8Z3HYSbG29D7odOn4MoA7ncAFXEMIN3MEDdKALAhJ4hXdv4r15H6uat6tDP4I+/zBzjGijg=</latexit> <latexit sha1_base64="17UuOYqWZInhVmp/m7fMJfwWJtk=">ACBHicbZDLSgMxGIX/qbdaq45u3QSLUmEoM1Wrm4LgxmUFe4G2lEyaUMzmSHJCGXoG7jxVdy4UMRncOfbmF4W2nog4eOchOQ/fsyZ0q7bWXW1jc2t7LbuZ387t6+fZBvqCiRhNZJxCPZ8rGinAla10xz2olxaHPadMf3U7z5iOVikXiQY9j2g3xQLCAEayN1bNPa8VWVZxV46LZOoHEJPUmaWXiIFH1nLJz7lw4l06lZxfckjsTWgVvAQVYqNazvzr9iCQhFZpwrFTbc2PdTbHUjHA6yXUSRWNMRnhA2wYFDqnqprN5JujEOH0URNIsodHM/X0jxaFS49A3J0Osh2o5m5r/Ze1EB9fdlIk40VSQ+UNBwpGO0LQc1GeSEs3HBjCRzPwVkSE2pWhTYc6U4C2PvAqNcskzfO9CFo7gGIrgwRXcwB3UoA4EnuAF3uDderZerY95XRlr0dsh/JH1+QNrsZdR</latexit> <latexit sha1_base64="17UuOYqWZInhVmp/m7fMJfwWJtk=">ACBHicbZDLSgMxGIX/qbdaq45u3QSLUmEoM1Wrm4LgxmUFe4G2lEyaUMzmSHJCGXoG7jxVdy4UMRncOfbmF4W2nog4eOchOQ/fsyZ0q7bWXW1jc2t7LbuZ387t6+fZBvqCiRhNZJxCPZ8rGinAla10xz2olxaHPadMf3U7z5iOVikXiQY9j2g3xQLCAEayN1bNPa8VWVZxV46LZOoHEJPUmaWXiIFH1nLJz7lw4l06lZxfckjsTWgVvAQVYqNazvzr9iCQhFZpwrFTbc2PdTbHUjHA6yXUSRWNMRnhA2wYFDqnqprN5JujEOH0URNIsodHM/X0jxaFS49A3J0Osh2o5m5r/Ze1EB9fdlIk40VSQ+UNBwpGO0LQc1GeSEs3HBjCRzPwVkSE2pWhTYc6U4C2PvAqNcskzfO9CFo7gGIrgwRXcwB3UoA4EnuAF3uDderZerY95XRlr0dsh/JH1+QNrsZdR</latexit> <latexit sha1_base64="ZzSe56npIEy+O0YUG4qA/G13BGU=">ACD3icbZBNS8MwGMdTX+d8q3r0EhzKhDLaqdPLYOjF4wT3AlsZaZuYWlaklQYpd/Ai1/FiwdFvHr15rcx23rQzQcSfvz/z0Py/L2IUals+9tYWl5ZXVvPbeQ3t7Z3ds29/aYMY4FJA4csFG0PScIoJw1FSPtSBAUeIy0vNHNxG89ECFpyO/VOCJugAac+hQjpaWeVIvtqv8tBoV9dX1BcKJkyaV1IK86lhl68w6ty6sSs8s2CV7WnARnAwKIKt6z/zq9kMcB4QrzJCUHceOlJsgoShmJM13Y0kihEdoQDoaOQqIdJPpPik81kof+qHQhys4VX9PJCiQchx4ujNAaijnvYn4n9eJlX/lJpRHsSIczx7yYwZVCfhwD4VBCs21oCwoPqvEA+RDkXpCPM6BGd+5UVolkuO5ju7ULvO4siBQ3AEisABl6AGbkEdNAGj+AZvI348l4Md6Nj1nrkpHNHIA/ZXz+APn6mMc=</latexit> <latexit sha1_base64="3hdbwTSVxBON+FiA5gMGQhY0lE=">ACD3icbZDLSgMxFIYz9VbrbdSlm2BRKgxlptbqplB047KCvUA7lEyaUMzmSHJCGXoG7jxVdy4UMStW3e+jWk7C209kPDx/+eQnN+LGJXKtr+NzMrq2vpGdjO3tb2zu2fuHzRlGAtMGjhkoWh7SBJGOWkoqhpR4KgwGOk5Y1upn7rgQhJQ36vxhFxAzTg1KcYKS31zN6oV3lZ9WoK+uLxBOnElSmViQVx2rZJ1bZevCqvTMvF20ZwWXwUkhD9Kq98yvbj/EcUC4wgxJ2XHsSLkJEopiRia5bixJhPAIDUhHI0cBkW4y2cCT7TSh34o9OEKztTfEwkKpBwHnu4MkBrKRW8q/ud1YuVfuQnlUawIx/OH/JhBFcJpOLBPBcGKjTUgLKj+K8RDpENROsKcDsFZXHkZmqWio/munK9dp3FkwRE4BgXgEtQA7egDhoAg0fwDF7Bm/FkvBjvxse8NWOkM4fgTxmfP/s6mMs=</latexit> <latexit sha1_base64="3hdbwTSVxBON+FiA5gMGQhY0lE=">ACD3icbZDLSgMxFIYz9VbrbdSlm2BRKgxlptbqplB047KCvUA7lEyaUMzmSHJCGXoG7jxVdy4UMStW3e+jWk7C209kPDx/+eQnN+LGJXKtr+NzMrq2vpGdjO3tb2zu2fuHzRlGAtMGjhkoWh7SBJGOWkoqhpR4KgwGOk5Y1upn7rgQhJQ36vxhFxAzTg1KcYKS31zN6oV3lZ9WoK+uLxBOnElSmViQVx2rZJ1bZevCqvTMvF20ZwWXwUkhD9Kq98yvbj/EcUC4wgxJ2XHsSLkJEopiRia5bixJhPAIDUhHI0cBkW4y2cCT7TSh34o9OEKztTfEwkKpBwHnu4MkBrKRW8q/ud1YuVfuQnlUawIx/OH/JhBFcJpOLBPBcGKjTUgLKj+K8RDpENROsKcDsFZXHkZmqWio/munK9dp3FkwRE4BgXgEtQA7egDhoAg0fwDF7Bm/FkvBjvxse8NWOkM4fgTxmfP/s6mMs=</latexit> <latexit sha1_base64="3hdbwTSVxBON+FiA5gMGQhY0lE=">ACD3icbZDLSgMxFIYz9VbrbdSlm2BRKgxlptbqplB047KCvUA7lEyaUMzmSHJCGXoG7jxVdy4UMStW3e+jWk7C209kPDx/+eQnN+LGJXKtr+NzMrq2vpGdjO3tb2zu2fuHzRlGAtMGjhkoWh7SBJGOWkoqhpR4KgwGOk5Y1upn7rgQhJQ36vxhFxAzTg1KcYKS31zN6oV3lZ9WoK+uLxBOnElSmViQVx2rZJ1bZevCqvTMvF20ZwWXwUkhD9Kq98yvbj/EcUC4wgxJ2XHsSLkJEopiRia5bixJhPAIDUhHI0cBkW4y2cCT7TSh34o9OEKztTfEwkKpBwHnu4MkBrKRW8q/ud1YuVfuQnlUawIx/OH/JhBFcJpOLBPBcGKjTUgLKj+K8RDpENROsKcDsFZXHkZmqWio/munK9dp3FkwRE4BgXgEtQA7egDhoAg0fwDF7Bm/FkvBjvxse8NWOkM4fgTxmfP/s6mMs=</latexit> <latexit sha1_base64="3hdbwTSVxBON+FiA5gMGQhY0lE=">ACD3icbZDLSgMxFIYz9VbrbdSlm2BRKgxlptbqplB047KCvUA7lEyaUMzmSHJCGXoG7jxVdy4UMStW3e+jWk7C209kPDx/+eQnN+LGJXKtr+NzMrq2vpGdjO3tb2zu2fuHzRlGAtMGjhkoWh7SBJGOWkoqhpR4KgwGOk5Y1upn7rgQhJQ36vxhFxAzTg1KcYKS31zN6oV3lZ9WoK+uLxBOnElSmViQVx2rZJ1bZevCqvTMvF20ZwWXwUkhD9Kq98yvbj/EcUC4wgxJ2XHsSLkJEopiRia5bixJhPAIDUhHI0cBkW4y2cCT7TSh34o9OEKztTfEwkKpBwHnu4MkBrKRW8q/ud1YuVfuQnlUawIx/OH/JhBFcJpOLBPBcGKjTUgLKj+K8RDpENROsKcDsFZXHkZmqWio/munK9dp3FkwRE4BgXgEtQA7egDhoAg0fwDF7Bm/FkvBjvxse8NWOkM4fgTxmfP/s6mMs=</latexit> <latexit sha1_base64="3hdbwTSVxBON+FiA5gMGQhY0lE=">ACD3icbZDLSgMxFIYz9VbrbdSlm2BRKgxlptbqplB047KCvUA7lEyaUMzmSHJCGXoG7jxVdy4UMStW3e+jWk7C209kPDx/+eQnN+LGJXKtr+NzMrq2vpGdjO3tb2zu2fuHzRlGAtMGjhkoWh7SBJGOWkoqhpR4KgwGOk5Y1upn7rgQhJQ36vxhFxAzTg1KcYKS31zN6oV3lZ9WoK+uLxBOnElSmViQVx2rZJ1bZevCqvTMvF20ZwWXwUkhD9Kq98yvbj/EcUC4wgxJ2XHsSLkJEopiRia5bixJhPAIDUhHI0cBkW4y2cCT7TSh34o9OEKztTfEwkKpBwHnu4MkBrKRW8q/ud1YuVfuQnlUawIx/OH/JhBFcJpOLBPBcGKjTUgLKj+K8RDpENROsKcDsFZXHkZmqWio/munK9dp3FkwRE4BgXgEtQA7egDhoAg0fwDF7Bm/FkvBjvxse8NWOkM4fgTxmfP/s6mMs=</latexit> <latexit sha1_base64="3hdbwTSVxBON+FiA5gMGQhY0lE=">ACD3icbZDLSgMxFIYz9VbrbdSlm2BRKgxlptbqplB047KCvUA7lEyaUMzmSHJCGXoG7jxVdy4UMStW3e+jWk7C209kPDx/+eQnN+LGJXKtr+NzMrq2vpGdjO3tb2zu2fuHzRlGAtMGjhkoWh7SBJGOWkoqhpR4KgwGOk5Y1upn7rgQhJQ36vxhFxAzTg1KcYKS31zN6oV3lZ9WoK+uLxBOnElSmViQVx2rZJ1bZevCqvTMvF20ZwWXwUkhD9Kq98yvbj/EcUC4wgxJ2XHsSLkJEopiRia5bixJhPAIDUhHI0cBkW4y2cCT7TSh34o9OEKztTfEwkKpBwHnu4MkBrKRW8q/ud1YuVfuQnlUawIx/OH/JhBFcJpOLBPBcGKjTUgLKj+K8RDpENROsKcDsFZXHkZmqWio/munK9dp3FkwRE4BgXgEtQA7egDhoAg0fwDF7Bm/FkvBjvxse8NWOkM4fgTxmfP/s6mMs=</latexit> e.g., 6-sided dice roll P ( X = n ) = p ( n ) = 1 6 , n = 1 , 2 , 3 , 4 , 5 , 6 — probability mass function X Note: P ( X = n ) = 1 n X E ( X ) = n · p ( n ) = 3 . 5 n

Recommend


More recommend