WAVE and Web technologies in the Integration of Adaptive 
Networks and Applications
NPAC at Syracuse University and the University of Surrey 
England - 
Principal Investigators Geoffrey Fox and Peter Sapaty

Innovative Claims (A)

1)We will deploy on both the general Internet and an application specific IntraNet, a 
novel distributed mobile adaptive system WAVE which has already been developed and 
demonstrated in prototype form to exhibit key features of  the intelligent adaptive 
networks called for in the BAA.  We will integrate WAVE with Web technology which 
we expect to be an important COTS technology framework of growing use in DoD 
applications. 
2)We will demonstrate and measure the fault tolerance and performance characteristics of 
WAVE in a relatively simple but large scale application - Web Search - showing that the 
self organization characteristics of WAVE leads to greater fault tolerance than 
conventional technologies in a heterogeneous adaptive network. WAVE itself will be 
used to simulate network faults and bandwidth adaptivity and heterogeneity.
3)We will use WAVE as the substrate of an existing Web - based Command and Control 
Application which is based on a novel Java collaborative environment TANGOSIM 
which is built on top of a distributed event driven simulator. This will exhibit combined 
adaptivity at both application and network level to satisfy real time requirements of the 
application in given network scenario by adjusting simultaneously video and image 
compression levels as well as network management.
4)We will develop a novel  MULTINET hierarchical resource management system which 
will be integrated at both application and network level. This combines ideas of multigrid 
iterative solvers with distributed deterministic annealing (generalized neural networks) 
optimization algorithms and has been successfully applied to parallel computing data 
decomposition and other classification and optimization problems.



Technical Description (G)
1) This proposal extends and applies a novel model and technology called WAVE for 
distributed processing in large open computer network. WAVE is based on a program 
mobility where a special recursive code navigates distributed environment and provides 
highly parallel, distributed and dynamic control. The technology may operate without any 
central memory or control and can support robust distributed algorithms efficiently 
working in a rapidly changing environment.  It may also may provide high recoverability 
from the system failures in both software and hardware.  The WAVE language, a core of 
the technology, describes parallel and asynchronous navigation in distributed systems 
rather than traditional data processing, and allows one to receive solutions of complex 
distributed computing and network control problems in a very dense mode, up to two 
orders of magnitude shorter than in C or Java. The WAVE technology is currently 
implemented in C as a network of cooperating WAVE language interpreters operating via 
the Internet, and is successfully used in commercial and defense projects. Possible 
applications include(but are not limited to): integration of distributed heterogeneous 
databases with distributed and parallel inference, distributed interactive simulation of 
large systems, like battlefields, with high mobility of interacting entities and dynamics of 
terrain, intelligent management and control of advanced transport and telecommunication 
networks, programming collective behavior of autonomous platforms (robots), intelligent 
command and control systems of different natures, air and space defense systems based 
on highly organized distributed computer networks driven by mobile programs, as well as 
a distributed dynamic 3D virtual reality. This technology is well suited to meet the 
requirements of the BAA and in this proposal it will be demonstrated in two large scale 
applications. Firstly we will integrate WAVE with Web technology providing both a high 
performance plug-in ( C language) version and a portable Java version. This will be 
demonstrated first in a large Web search application where we will use WAVE itself to 
simulate faulty links and demonstrate the adaptive scaling characteristics of  WAVE. We 
will take an existing Web search system used by NPAC and implement the gatherer 
component with WAVE technology. This relatively simple demonstration of WAVE 
technology will be followed by the large scale system demonstration described in the 
following section.