 |
Distributed
Strategies
in
Active Network Applications
|
This
project is aimed to analyze properties of Active Networks programming languages
and Execution Environments.
Establishing
provable properties requires a mathematical perspective on the software
architecture.
None
of the classical computing models (such as Turing machines, register machines,
lambda calculus) is concerned with interaction, as we would normally understand
the term.
In
the classical theory, rather little attention is paid to the way in which
two automata may interact, in the sense that an action by one entails a
complementary action by another.
It
appears that if we are interested in interactive behavior, then a non-deterministic
automaton cannot correctly be equated behaviorally with a deterministic
one.
Aim
of the project is the study of the interaction forms occurring among the
active network applications.
The
project is a collaboration among:
-
ICAR
- CNR, Palermo - Italia
-
CRATOS
- Universita' Cattolica del Sacro Cuore, Piacenza - Italia
-
ICSI
- International Computer Science Institute, Berkeley - USA
See
also: Distributed
Strategies in Active Networks Applications.
Introduction
Emerging
paradigms for distributed computing move their domain from the network
terminal nodes towards the whole network (terminal and internal nodes).
In
such a way the network can be seen as a whole powerful computing engine.
The classical "store-and-forward" model for the data delivering is replaced
by an innovative "store-compute-and-forward" model, thus originating the
"Active Networks" model.
Outline
of the Project
Our project
proposes the study of the problems which are related to interactions among
active network applications.
-
How active
network applications may exchange information?
We
will try to individuate all the possible interaction forms.
Currently
we adopt a differentiation between interference and communication.
-
Which
are the tools we need to accomplish the interaction?
We
should focus on the basic functions which will allow the existing AN environments
to implement the proposed interaction forms.
-
We will
extend current AN environments to make them capable of supporting the different
interaction forms.
-
We will
study the characteristics and capabilities of the interaction forms using
formal methods.
Currently
adopted tools:
-
The ABONE
testbed.
-
The PLAN
language and execution environment.
-
The PI-Calculus
formal description languages.
Papers
and Presentations
-
L. Delgrossi,
G. Di Fatta, D. Ferrari , G. Lo Re, "Interference and Communications among
Active Network Applications", Lectures Notes in Computer Science 1653,
Active Networks IWAN 99, S. Covaci (Ed.) Springer,
pp.
97-108. (pdf
86kB) (ps
596kB)
-
G. Di
Fatta, "The Active Networking Challenge", ICSI, 30 April, 1999. (ppt
285 kB).
-
G. Di
Fatta, "Interference and Communications among Active Network Applications",
IWAN'99 The First International Working Conference on Active Networks.Berlin,
1 July, 1999. (ppt.gz
8 kB)
last modified (09/27/99)