Architecture-Independent Programming for Wireless Sensor Networks

New automated, application-independent methodology for designing
and deploying sensor networks

Following this book's clear explanations, examples, and
illustrations, domain experts can design and deploy nontrivial
networked sensing applications without much knowledge of the
low-level networking aspects of deployment. This new approach is
based on the Abstract Task Graph (ATaG), a data-driven programming
model and an innovative methodology forarchitecture-independent
programming and automatic software synthesis for sensor networks.
ATaG programs are concise, understandable, and network-independent
descriptions of global application functionality that can be
automatically compiled onto any target deployment.

The book begins with an overview chapter that addresses the
important issues of programming methodologies and compares various
programming models for sensor networks. Next, the authors set forth
everything you need for designing and deploying sensor networks
using ATaG, including:

• Detailed description of the ATaG model's features

• System-level support for architecture-independent
  programming

• Examination of the graphical programming and software synthesis
  environment for ATaG

• Case study illustrating the process of end-to-end application
  development and software synthesis using ATaG

Throughout the book, the authors provide code excerpts and
figures to help clarify key concepts and explain each step.

For programmers, the graphical formalism of the ATaG program,
coupled with the fact it uses an existing language (Java), means
that no special training is needed to start developing and
deploying applications in ATaG. Everything you need to know is
clearly set forth in this book.

Auteur

Amol B. Bakshi, PhD, is a Research Assistant Professor in the
Ming Hsieh Department of Electrical Engineering at the University
of Southern California (USC), Los Angeles. He also manages the
Integrated Asset Management project at the USC-Chevron Center of
Excellence for Research and Academic Training on Interactive Smart
Oilfield Technologies. Dr. Bakshi's PhD research was on programming
models for networked sensor systems and led to the creation of the
ATaG programming model and software synthesis toolkit. His current
interests include semantic Web technologies for information
integration, smart oilfield technologies, model integrated
computing, and sensor networks.

Viktor K. Prasanna, PhD, is Charles Lee Powell Chair in
Engineering and Professor of Electrical Engineering and Professor
of Computer Science at the University of Southern California (USC),
Los Angeles. He is also an associate member of the Center for
Applied Mathematical Sciences (CAMS) at USC, and a member of the
USC-Chevron Center of Excellence for Research and Academic Training
on Interactive Smart Oilfield Technologies. He has served on the
editorial boards of the Journal of Parallel and Distributed
Computing, Proceedings of the IEEE, IEEE Transactions on VLSI
Systems, and IEEE Transactions on Parallel and Distributed Systems.
He was editor-in-chief of the IEEE Transactions on Computers and
was the founding chair of the IEEE Computer Society Technical
Committee on Parallel Processing. He is a Fellow of the IEEE and
the ACM.

Résumé
New automated, application-independent methodology for designing and deploying sensor networks Following this book's clear explanations, examples, and illustrations, domain experts can design and deploy nontrivial networked sensing applications without much knowledge of the low-level networking aspects of deployment. This new approach is based on the Abstract Task Graph (ATaG), a data-driven programming model and an innovative methodology forarchitecture-independent programming and automatic software synthesis for sensor networks. ATaG programs are concise, understandable, and network-independent descriptions of global application functionality that can be automatically compiled onto any target deployment.

The book begins with an overview chapter that addresses the important issues of programming methodologies and compares various programming models for sensor networks. Next, the authors set forth everything you need for designing and deploying sensor networks using ATaG, including:

• Detailed description of the ATaG model's features

• System-level support for architecture-independent programming

• Examination of the graphical programming and software synthesis environment for ATaG

• Case study illustrating the process of end-to-end application development and software synthesis using ATaG

Throughout the book, the authors provide code excerpts and figures to help clarify key concepts and explain each step.

For programmers, the graphical formalism of the ATaG program, coupled with the fact it uses an existing language (Java), means that no special training is needed to start developing and deploying applications in ATaG. Everything you need to know is clearly set forth in this book.

Contenu
Preface. Acknowledgments.

1. Introduction.

1.1 Sensor networks and traditional distributed systems.

1.2 Programming of distributed sensor networks.

1.2.1 Layers of programming abstraction.

1.2.1.1 Service-oriented specification.

1.2.1.2 Macroprogramming.

1.2.1.3 Node-centric programming.

1.2.2 Lessons from parallel and distributed computing.

1.3 Macroprogramming: What and Why?

1.4 Contributions and Outline.

2. The Abstract Task Graph.

2.1 Target applications and architectures.

2.2 Key Concepts.

2.2.1 Data Driven Computing.

2.2.1.1 Program flow mechanisms.

2.2.1.2 Why data driven?

2.2.2 Mixed Imperative-Declarative Specification.

2.3 Syntax.

2.3.1 The Structure of an ATaG Program.

2.3.2 More on Task Annotations.

2.3.3 Illustrative examples.

2.4 Semantics.

2.4.1 Terminology.

2.4.2 Firing rules.

2.4.3 Task graph execution.

2.4.4 get() and put().

2.5 Programming idioms.

2.5.1 Object tracking.

2.5.2 Interaction within local neighborhoods.

2.5.3 In-network aggregation.

2.5.4 Hierarchical data fusion.

2.5.5 Event-triggered behavior instantiation.

2.6 Future work.

2.6.1 State-based dynamic behaviors.

2.6.2 Resource management in the runtime system.

2.6.3 Utility based negotiation for task scheduling and resource allocation.

2.6.4 Analyzing feasibility of compilation.

3. DART:The Data Driven ATaG Runtime.

3.1 Design objectives.

3.1.1 Support for ATaG semantics.

3.1.2 Platform independence.

3.1.3 Component-based design.

3.1.4 Ease of software synthesis.

3.2 Overview.

3.3 Components and functionalities.

3.3.1 Task, Data, and Channel Declarations.

3.3.2 UserTask.

3.3.2.1 Service.

3.3.2.2 Interactions.

3.3.2.3 Implementation.

3.3.3 DataPool.

3.3.3.1 Service.

3.3.3.2 Interactions.

3.3.3.3 Implementation.

3.3.4 AtagManager.

3.3.4.1 Service.

3.3.4.2 Interactions.

3.3.4.3 Implementation.

3.3.5 NetworkStack.

3.3.5.1 Service.

3.3.5.2 Interactions.

3.3.5.3 Implementation.

3.3.6 NetworkArchitecture.

3.3.6.1 Service.

3.3.6.2 Interactions.

3.3.6.3 Implementation.

3.3.7 Dispatcher.

3.3.7.1 Service.

3.3.7.2 Interactions.

3.3.7.3 Implementation.

3.4 Control flow.

3.4.1 Startup.

3.4.2 get() and put().

3.4.3 Illustrative example.

3.5 Future work.

3.5.1 Lazy compilation of channel annotations.

3.5.2 Automatic priority assignment for task scheduling.

4. Programming and Software Synthesis.

4.1 Terminology.

4.2 Meta-modeling for the ATaG domain.

4.2.1 Objectives.

4.2.2 Application model.

4.2.3 Network model.

4.3 The programming interface.

4.4 Compilation and software synthesis.

4.4.1 Translating task anno…