Memory Controllers for Real-Time Embedded Systems

Predictable and Composable Real-Time Systems

  • Benny Akesson
  • Kees Goossens
Part of the Embedded Systems book series (EMSY, volume 2)

Table of contents

  1. Front Matter
    Pages i-xxi
  2. Benny Akesson, Kees Goossens
    Pages 1-27
  3. Benny Akesson, Kees Goossens
    Pages 29-43
  4. Benny Akesson, Kees Goossens
    Pages 45-61
  5. Benny Akesson, Kees Goossens
    Pages 63-103
  6. Benny Akesson, Kees Goossens
    Pages 105-142
  7. Benny Akesson, Kees Goossens
    Pages 143-170
  8. Benny Akesson, Kees Goossens
    Pages 171-186
  9. Benny Akesson, Kees Goossens
    Pages 187-194
  10. Benny Akesson, Kees Goossens
    Pages 195-201
  11. Back Matter
    Pages 203-221

About this book

Introduction

 

Verification of real-time requirements in systems-on-chip becomes more complex as more applications are integrated. Predictable and composable systems can manage the increasing complexity using formal verification and simulation.  This book explains the concepts of predictability and composability and shows how to apply them to the design and analysis of a memory controller, which is a key component in any real-time system.

This book is generally intended for readers interested in Systems-on-Chips with
real-time applications.   It is especially well-suited for readers looking to use SDRAM memories in systems with hard or firm real-time requirements. There is a strong focus on real-time concepts, such as predictability and composability, as well as a brief discussion about memory controller architectures for high-performance computing.

Readers will learn step-by-step how to go from an unpredictable SDRAM memory,
offering highly variable bandwidth and latency, to a predictable and composable
shared memory, providing guaranteed bandwidth and latency to isolated applications.
This journey covers concepts for making memories and arbiters behave in a
predictable and composable manner, as well as architecture descriptions of hardware
blocks that implement the concepts.

  • Provides an overview of trends in embedded system design that make design of real-time SoCs difficult, error-prone, and expensive;
  • Introduces the concept of predictability, which is required for formal verification of real-time systems;
  • Introduces the concept of composability, which is a divide and conquer technique that enables performance verification per application, instead of monolithic verification for all applications together;
  • Describes a novel approach to composability, which applies to any predictable shared resource, thus widely extending the scope of composable platforms. This is the first approach that can efficiently support SDRAM, which is an essential system component;
  • Provides an overview of the SDRAM architecture at a level that is relevant for system designers, not memory designers, and explains why SDRAM architectures are difficult to use in real-time systems;
  • Describes concepts, architectures, implementation and worst-case performance analysis of predictable SDRAM accesses, as well as predictable and composable memory arbitration, which can be applied to all memory types

Keywords

Embedded Systems Embedded Systems Integrated Circuits and Systems Integrated Circuits and Systems Memory Controllers Memory Controllers Predictable and Composable Systems Predictable and Composable Systems Real-time systems Real-time systems SDRAM SDRAM SOC SOC Systems-on-Chip Systems-on-Chip Verification of Embedded Systems Verification of Embedded Systems

Authors and affiliations

  • Benny Akesson
    • 1
  • Kees Goossens
    • 2
  1. 1., Faculty of Electrical EngineeringEindhoven University of TechnologyEindhovenNetherlands
  2. 2., Faculty of Electrical EngineeringEindhoven University of TechnologyEindhovenNetherlands

Bibliographic information

  • DOI https://doi.org/10.1007/978-1-4419-8207-0
  • Copyright Information Springer Science+Business Media, LLC 2012
  • Publisher Name Springer, New York, NY
  • eBook Packages Engineering
  • Print ISBN 978-1-4419-8206-3
  • Online ISBN 978-1-4419-8207-0
  • Series Print ISSN 2193-0155