Yang Zhao and Edward A. Lee and Jie Liu

EECS Department, University of California, Berkeley

Technical Report No. UCB/EECS-2006-82

May 28, 2006

http://www2.eecs.berkeley.edu/Pubs/TechRpts/2006/EECS-2006-82.pdf

Discrete-event (DE) models are formal system specifications that have analyzable deterministic behaviors in terms of event values and time stamps. However, since time is only a modeling property, they are primarily used in performance modeling and simulation. In this paper, we extend discrete-event models with the capability of mapping certain events to physical time and propose them as a programming model, called PTIDES. We seek analysis tools and execution strategies that can preserve the deterministic behaviors specified in DE models without paying the penalty of totally ordered executions. This is particularly intriguing in time synchronized distributed systems since there is a consistent global notion of time and intrinsic parallelism among the nodes. Based on causality analysis of DE systems, we define relevant dependency and relevant orders to enable outof- order executions without hurting determinism and without requiring backtracking. For a given network characteristic, we can check statically whether deploying the model in the network can preserve the real-time properties in the specification.

---

BibTeX citation:

@techreport{Zhao:EECS-2006-82,
    Author= {Zhao, Yang and Lee, Edward A. and Liu, Jie},
    Title= {Programming Temporally Integrated Distributed Embedded Systems},
    Year= {2006},
    Month= {May},
    Url= {http://www2.eecs.berkeley.edu/Pubs/TechRpts/2006/EECS-2006-82.html},
    Number= {UCB/EECS-2006-82},
    Note= {See the <a href="http://ptolemy.eecs.berkeley.edu/publications/papers/07/RTAS/">published version</a>},
    Abstract= {Discrete-event (DE) models are formal system specifications
that have analyzable deterministic behaviors in terms
of event values and time stamps. However, since time is only
a modeling property, they are primarily used in performance
modeling and simulation. In this paper, we extend discrete-event
models with the capability of mapping certain events to physical
time and propose them as a programming model, called PTIDES.
We seek analysis tools and execution strategies that can preserve
the deterministic behaviors specified in DE models without paying
the penalty of totally ordered executions. This is particularly
intriguing in time synchronized distributed systems since there
is a consistent global notion of time and intrinsic parallelism
among the nodes. Based on causality analysis of DE systems,
we define relevant dependency and relevant orders to enable outof-
order executions without hurting determinism and without
requiring backtracking. For a given network characteristic, we
can check statically whether deploying the model in the network
can preserve the real-time properties in the specification.},
}

EndNote citation:

%0 Report
%A Zhao, Yang 
%A Lee, Edward A. 
%A Liu, Jie 
%T Programming Temporally Integrated Distributed Embedded Systems
%I EECS Department, University of California, Berkeley
%D 2006
%8 May 28
%@ UCB/EECS-2006-82
%U http://www2.eecs.berkeley.edu/Pubs/TechRpts/2006/EECS-2006-82.html
%F Zhao:EECS-2006-82