1 / 12

Hierarchical Real-Time Systems for Imprecise Computation Model

The 5th EuroSys Doctoral Workshop ( EuroDW 2011). Hierarchical Real-Time Systems for Imprecise Computation Model. Guy Martin Tchamgoue Department of Informatics, Gyeongsang National University, South Korea. Agenda. Problem An Application Proposed Solution Preliminary Results

onaona
Download Presentation

Hierarchical Real-Time Systems for Imprecise Computation Model

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. The 5th EuroSys Doctoral Workshop (EuroDW 2011) Hierarchical Real-Time Systems for Imprecise Computation Model • Guy Martin Tchamgoue • Department of Informatics, • Gyeongsang National University, South Korea

  2. April 10, 2011 Agenda • Problem • An Application • Proposed Solution • Preliminary Results • Ongoing Work

  3. April 10, 2011 Problem • Statement • To build large complex real-time systems from simpler ones • Flexibility and Reliability • To handle both hard and soft real-time systems • To enhance the resource utilization on system overload • To facilitate the development and deployment process

  4. Virtual Machine or Partition 1 • Virtual Machine or Partition 2 • Virtual Machine or Partition N • Virtual Machine or Partition N-1 April 10, 2011 An Application Audio/Video Applications Multimedia Streaming Real-Time Image Tracking Other Applications Operating System Operating System Operating System Operating System Operating System

  5. April 10, 2011 Compositional Scheduling Framework • Hierarchical resource sharing among components under different scheduling policies • Higher-level schedulers: • Consider each component as a scheduling unit • Do not need to know about the internal complexity of its components • Lower-level schedulers: • Focus on local scheduling for a given resource supply C3 C4 C1 C1 C2 C2

  6. April 10, 2011 Preliminary Results • Publications • Tchamgoue et al., “Compositional Scheduling Framework for Imprecise Computation Model”, Work-in-Progress Session, RTAS’10, April 2010 • Tchamgoue et al., “Hierarchical Real-Time Scheduling Framework for Imprecise Computations”, 8th IEEE/IFIP EUC, December 2010 • System Model • Imprecise Computation Tasks: T = {i (pi , mi , oi, fi)} • pi : period • mi: execution time of the mandatory part • oi: execution time of the optional part • fi: reward function for optional execution

  7. April 10, 2011 • The Imprecise Resource Model: (, , ) • : resource period • : resource allocation for mandatory part • : resource allocation for optional part • (, , ) periodically supplies at least  time units and at most  +  time units. • Example: • The resource supply of (, , )

  8. April 10, 2011 • The Interface Model • Specifying the real-time requirement of a component • Interface Model I(P, M, O, F) • P : period • M : mandatory part requirement • O : optional part requirement • F : reward function of optional execution I0 (5, 3, 2, F0(t)) MF-EDF 5 (10, 3, 2, F1(t)) 6 (15, 4, 2, F2(t)) R1=1(10, 3, 2) R2=2(15, 4, 2) Component C0 I1(10, 3, 2, F1(t)) I2(15, 4, 2, F2(t)) MF-RM MF-EDF 1(40, 5, 3, f1(t)) 2(30, 4, 2, f2(t)) 3(50, 4, 2, f3(t)) 4 (25, 3, 3, f4(t)) Component C1 Component C2

  9. April 10, 2011 • Other Results • Schedulability Analysis • To guarantee the minimum requirements of a component • Utilization Bound • To determine the largest possible utilization bound that makes a component schedulable • Scheduling Algorithm with Guarantee of Reward • To provide a minimum reward guarantee to an upper layer that provides resources to lower layers

  10. April 10, 2011 Ongoing Work • Component Interface Generation • How to optimally compute the resource and interface parameters • Component Reward function • To derive the reward function for the component’s interface • Optimal scheduling of optional parts • To schedule optional parts in order to maximize the total reward of a component

  11. April 10, 2011 Thank you for your attention. EuroDW 2011

  12. April 10, 2011 Imprecise Computation* (IC) Model • Proposed for flexible scheduling of hard deadline tasks • An IC task consists of: • A mandatory subtask for an acceptable result, and • An optional subtask for • refining the mandatory task result ,and • improving the quality of the task according to the amount of execution time. period : Mandatory part : Optional part time * J. W.-S. Liu, K.-J. Lin, W.-K. Shih, A. C.-S. Yu, C. Chung, J. Yao, and W. Zhao, “Algorithms for scheduling imprecise computations,”IEEE Computer, vol. 24, no. 5, pp. 58–68, 1991.

More Related