One Year of CHESS

1
One Year of CHESS

• Tom Henzinger

2
The CHESS Faculty

• Board of Directors
• Tom Henzinger, tah_at_eecs.berkeley.edu
• Edward A. Lee, eal_at_eecs.berkeley.edu
• Alberto Sangiovanni-Vincentelli,
  alberto_at_eecs.berkeley.edu
• Shankar Sastry, sastry_at_eecs.berkeley.edu
• Other Key Faculty
• Alex Aiken, aiken_at_cs.berkeley.edu
• Dave Auslander, dma_at_me.berkeley.edu
• Ras Bodik, bodik_at_cs.berkeley.edu
• Ruzena Bajcsy, ruzena_at_cs.berkeley.edu
• Karl Hedrick, khedrick_at_me.berkeley.edu
• Kurt Keutzer, keutzer_at_eecs.berkeley.edu
• George Necula, necula_at_cs.berkeley.edu
• Masayoshi Tomizuka, tomizuka_at_me.berkeley.edu
• Pravin Varaiya, varaiya_at_eecs.berkeley.edu
• Partner Universities
• Vanderbilt University

EE
Embedded Systems
CS
ME
3
The CHESS Mission

• The goal of the Center is to provide an
  environment for graduate research on the design
  issues necessary for supporting next-generation
  embedded software systems.
• The research focus is on developing model-based
  and tool-supported design methodologies for
  real-time fault-tolerant software on
  heterogeneous distributed platforms.
• The Center maintains a close interaction between
  academic research and industrial experience.
• A main objective is to facilitate the creation
  and transfer of modern, "new economy" software
  technology methods and tools to "old economy"
  market sectors in which embedded software plays
  an increasingly central role, such as aerospace,
  automotive, and consumer electronics.

4
Embedded Software Systems

• Computational
• but not first-and-foremost a computer
• Integral with physical processes
• sensors, actuators
• Reactive
• at the speed of the environment
• Heterogeneous
• hardware/software, mixed architectures
• Networked
• shared, adaptive

Source Edward A. Lee
5
The Safety Issue
6
The Cost Issue
7
The Embedded Software Challenge

• Problem indicators
• System integration cost too high (40-50)
• Cost of change too high
• Design productivity crisis

Process

• Root cause of problems is the emerging new role
  of embedded software
• Exploding integration role
• New functionalities
• Expected source of flexibility in systems

Embedded Software
mC
D M A
ROM
FPGA
DSP
PROBLEM Lack of Design Technology aligned with
the new role.
Process
Source Janos Sztipanovits
8
Todays Techniques Wont Do

• Largely Missing
• Abstraction (especially Time)
• Compositionality
• Inheritance Polymorphism
• Portability Reusability
• Verifiability Reliability
• Robustness

Source Edward A. Lee
9
Todays Techniques Wont Do

• Largely Missing
• Abstraction (especially Time)
• Compositionality
• Inheritance Polymorphism
• Portability Reusability
• Verifiability Reliability
• Robustness

Source Edward A. Lee
10
Complexity Management in Engineering
Calculate
Model
Mathematics
Predict
Abstract
Bridge Aircraft Software ?!
System
Test
11
Two Opposing Forces
Requirements
Verification
Model
Environment
Implementation
Resources
12
The CHESS Approach

• Model-based Design (the view from above)
• principled frameworks for design
• merging specification, modeling, and design
• manipulable (mathematical) models
• enabling analysis and verification
• enabling effective synthesis of implementations
• Platform-based Design (the view from below)
• exposing key resource limitations
• hiding inessential implementation details
• Tools
• concrete realizations of design methods

13
Two Opposing Forces
Source Alberto Sangiovanni-Vincentelli
14
Two Opposing Forces
Requirements
Verification
Component
Component
Implementation
Resources
15
Two Opposing Forces
Requirements
(timing, fault tolerance, etc.)
Verification
no change
Component
Component
Deep Compositionality
Implementation
no change
Resources
16
Foundational Theory Research

• The science of computation has systematically
  abstracted away the physical world. The science
  of physical systems has systematically ignored
  computational limitations.
  Embedded software systems, however, engage the
  physical world in a computational manner.
• It is time to construct a Hybrid Systems Science
  that is simultaneously computational and
  physical.
  
• Time, concurrency, robustness, continuums, and
  resource management must be remarried to
  computation.

17
A Hybrid Systems Science
Environment Processes (physical models)
Reactivity
Embedded System Design
Schedulability
Software Processes (models of computation)
18
and Applied Systems Research

• Models and Tools
• Model-based design (platforms, interfaces,
  metamodels, virtual machines, abstract syntax and
  semantics, etc.)
• Tool-supported design (simulation, verification,
  code generation, interoperability, etc.)
• Applications
• Flight control systems
• Automotive electronics
• National experimental embedded software platform

19
Some Major CHESS Projects

• Foundations of Hybrid and Embedded Software
  Systems
• NSF ITR (Information Technology Research)
• Hybrid System Design Tools for Software-Enabled
  Control
• DARPA SEC (Software-Enabled Control)
• Process-based Software Components for Embedded
  Systems
• DARPA MoBIES (Model-Based Integration of Embedded
  Software)
• Automating the Development and Analysis of
  Embedded Systems
• DARPA PCES (Program Composition for Embedded
  Software)
• Platform-Based Design of Heterogeneous Systems
• MARCO GSRC (Gigascale System Research Center)
• Rich Interfaces for Component-Based Design
• ONR
• Towards Predictability and Portability in
  Embedded Software
• NSF Hybrid and Embedded Systems Program
• Etc.

20
and A New Engineering Education

• Mathematics
• used to be calculus, differential equations,
  linear algebra
• now also logic, topology, universal algebra,
  combinatorics
• Fundamental Limits
• used to be thermal noise, the speed of light
• now also computability, complexity
• Building Blocks
• used to be capacitors, resistors, transistors,
  gates, op amps
• now also processors, DSP cores, software
  components

Source Edward A. Lee
21
Education The Starting Point
Berkeley has a required sophomore course EECS 20
that addresses mathematical modeling of signals
and systems from a computational perspective.
The web page at the right illustrates a broad
view of feedback, where the behavior is a fixed
point solution to a set of equations. This view
covers both traditional continuous feedback and
discrete-event systems.
Source Edward A. Lee
22
Education Beyond EECS 20

• New Courses
• Engineering Science
• Upper-division course on embedded systems
• Several graduate courses
• Curriculum Council
• Mold the new curriculum
• Spread the new curriculum
• Summer Institute
• How to facilitate change
• Recruit participants from community colleges

23
Program Preview

• Applications
• Flight control
• Automotive systems
• Models
• Heterogeneous models of computation
• Hybrid systems theory
• Software
• Compilation
• Validation
• Systems
• Embedded processors
• Networked embedded systems
• Code generation

24
Poster Session at Noon

• CAL A Language for Describing Dataflow Actors
• Joern Janneck
• CHIC A Tool for Checking Interface Compatibility
• Arindam Chakrabarti
• GIOTTO Scheduling and Implementation
• Ben Horowitz
• The METROPOLIS Design Environment
• Guang Yang
• METROPOLIS in the Design of Wireless Sensor
  Networks
• Rong Chen
• TinyGALS A Programming Model for Event-Driven
  Embedded Systems
• Elaine Cheong
• A Programmable Microkernel for Real-Time Systems
• Marco Sanvido
• Etc.