Objectives
As pointed out in the PITAC report, software is the new physical
infrastructure of the information age. It is fundamental to economic
success, scientific and technical research and national security.
Our current ability to construct the large and complex software
systems demanded for continued economic and military success is
inadequate.
The workshop will discuss the following challenges in networked
systems that will require further major advances in software and
systems technology:
- System Integration and Dynamic Adaptation.
A new challenge in networked systems is that stable application
performance needs to be maintained actively during run-time due
to the dynamically changing communication and computing platforms.
Consequently, the run-time architecture must include active control
for adapting the system/software to changing conditions. Global
system characteristics need to be achieved by increased run-time
use of reflection (systems that utilize their own models), advanced
interface modeling, self-adaptation, and self-optimization
- Effects of Dynamic Structure. The structure
of networked systems is complex and highly dynamic. Because systems
are formed by ad hoc networks of nodes and connections, the system
lacks fine-grain determinism for end-to-end behaviors that span
subsystem and network boundaries. In addition, there are end-to-end
system qualities such as timeliness and security attached to these
behaviors that can only be evaluated in this dynamically integrated
context. Simulation and analysis research directions have begun
to address this, but efforts are in their infancy.
- Effects of Faults. Faults and disruptions in
the underlying communication and computing infrastructure are
the normal events. Since well understood techniques for fault
tolerant computing, such as n-modular redundancy, are not applicable
in the dynamically changing networked architecture, new technology
is required for building safe and reliable applications on the
dynamic, distributed platforms.
- Design for Reliability. Although there are
a variety of metrics and established practices for characterizing
the expected failure behavior of a system after it is fielded
and there are established practices for specifying the desired
reliability of a system, the evaluation of system or software
reliability prior to fielding is nonetheless something of a black
art. The process for certifying that a system meets high reliability
goals under the range of conditions expected in actual use currently
involves exhaustive analysis of a system, including its development
history. In the end, the accepted ways of characterizing expected
system reliability do not give most system engineers the confidence
they would like to have in concluding that a system will have
particular reliability characteristics.
- Effects of Scale. Another risk that overlays
all proposed solutions is scale. Scale also addresses both run-time
and design time concerns. Typically, demonstrations are the convincing
drivers to technology adoption. Demonstrations of new technologies
however are usually small-scale, focused efforts. It is an open
problem how to scale up a demonstration that addresses the number
of nodes and connections, and the number of software developers,
analysts, and integrators to provide enough proof to justify technology
transition
In this workshop we will bring together experienced researchers
who have been involved in tool development and networked system
development, with leaders in industry who are stakeholders in the
progress. The overall aim is to exchange ideas for continued research
in this area and to reduce the gap between foundations and software
practice.
Topics
Workshop topics include but are not limited to:
- Model-based software development of network-centric systems;
- Foundations of future design and programming abstractions;
- Active fault management in network-centric systems;
- Intelligent, robust middleware;
- Model based development of certifiable systems;
- Reliability metrics;
- Architectures for networked systems;
- Probabilistic modeling of networked systems
- Pro-active methods for preventing predictable faults
- Self-Adaptive architectures
- Active control of QoS
- Scaling
Focus Questions
Some focus questions for the workshop are:
- How to model or compute from the models software reliability?
- How to organize the development of large-scale, highly dynamic
systems such that fault management is properly addressed?
- How can QoS be built into ad-hoc networks?
- What techniques are available which compellingly reduce the
aggregate certification effort?
Organization issues & deadlines
The objective of the 2005 edition of the Monterey Workshops is
to elaborate a snapshot of the state of the art concerning the construction
of safe networked systems. We aim to achieve this by structuring
the program as follow:
- Introductory talks will be given to provide a broad vision in
the area,
- Invited talks will focus on specific scientific results from
participants.
- Structured discussion on research approaches in the context
of a set of "Challenge Problems". The Challenge Problems
will address specific issues from current literature and practice.
So far, two problems are identified:
feel free to adapt these problems and to integrate hypotheses
that seems suitable.
- Open discussion summarizing Workshop findings: promising approaches,
open problems, excpected new challenges and technology roadmaps.
Invitees are requested to provide a 1 to 3 page abstract focusing
on a research approach, or challenge problem. Based on the submitted
abstracts, the Program Committee will select invited talks in both
areas. These contributions should be provided via electronic
mail by September 4th 2005.Selection
of invited talks will be made by September 11th, 2005.
Those participants who plan to contribute to the program with new
Challenge Problems, we will request to make the problem description
available through the web site for others before the meeting. Participants
contributing to the program with research approaches are invited
to prepare a few slides to outline how you could address some issues
deduced from the Challenge Problems.
Based on the discussions, participants will be invited to submit
extended material for publication in a special issue. |
