NASA Scientific and Technical Aerospace Reports

Load balancing, fault-tolerance, and scalability play an important role in distributed computing. They are of interest for
many distributed systems problems as well as for many problems in other areas. Collaborative applications, for example, need
to provide some level of fault tolerance to allow for failures of participants’ machines. In networking, as another example,
techniques are needed to balance the load offered to network links and to tolerate machine failures. In this dissertation, the
author presents three projects with different goals to which all or some of these issues are relevant, and presents different ways
to address them. The main focus will be on the World Wide Web, but the techniques presented here are not restricted to it.
The Web can be viewed as a large, unreliable, distributed platform composed of different machines with different speeds that
may slow down, speed up, and crash-fail at any time, and networks with different capacities that may get congested and
partitioned. All of these factors make the Web a very challenging environment. The projects the author presents range from
resource allocation, to parallel computing, to infrastructure support for collaborative applications. The dissertation is organized
as follows: first the author presents WebSeAl and discusses how it addresses scalability, load balancing, and fault masking in
Chapter 2. In Chapter 3, he will describe Charlotte and the techniques it uses to deal with machines of different speeds and
crash-failed machines. In Chapter 4, he presents an infrastructure Web-based application which contains various aspects
relating to scalability and fault masking.
DTIC
Allocations; Client Server Systems; Distributed Processing; Grid Computing (Computer Networks); Internets; Parallel
Processing (Computers); Resource Allocation; World Wide Web
20060001830 New York Univ., New York, NY USA
Polymorphic Type Inference and Abstract Data Types
Laufer, Konstantin; Jul. 1, 1992; 147 pp.; In English
Contract(s)/Grant(s): N00014-90-J1110; N00014-91-5-1472
Report No.(s): AD-A440329; No Copyright; Avail.: Defense Technical Information Center (DTIC)
Many statically typed programming languages provide an abstract data type construct, such as the package in Ada, the
cluster in CLU, and the module in Modula2. However, in most of these languages, instances of abstract data types are not
first-class values, and they cannot be assigned to a variable, passed as a function parameter, or returned as a function result.
The higher-order functional language ML has a strong and static type system with parametric polymorphism. In addition, ML
provides type reconstruction and does not require type declarations for identifiers. Although the ML module system supports
abstract data types, their instances cannot be used as first-class values for type-theoretic reasons. In this thesis, the author
describes a family of extensions of ML. While retaining ML’s static type discipline, type reconstruction, and most of its
syntax, he adds significant expressive power to the language by incorporating first-class abstract types as an extension of ML’s
free algebraic data types. In particular, he is now able to express multiple implementations of a given abstract type,
heterogeneous aggregates of different implementations of the same abstract type, and dynamic dispatching of operations with
respect to the implementation type. Following Mitchell and Plotkin, he formalizes abstract types in terms of existentially
quantified types, and proves that this type system is semantically sound with respect to a standard denotational semantics. He
then presents an extension of Haskell, a non-strict functional language that uses type classes to capture systematic overloading.
This language results from incorporating existentially quantified types into Haskell and gives first-class abstract types with
type classes as their interfaces. One can now express heterogeneous structures over type classes. The language is statically
typed and offers comparable flexibility to object-oriented languages.
DTIC
Computer Programming; High Level Languages; Inference; Optimization; Polymorphism; Programming Languages;
Semantics
20060001836 Aptima, Inc., Woburn, MA USA
From Cognitive Task Analysis to Simulation: Developing a Synthetic Team Task for AWACS Weapons Directors
Hess, Stephen M.; MacMillan, Jean; Serfaty, Daniel; Elliott, Linda; Jan. 1, 2005; 8 pp.; In English; Original contains color
illustrations
Contract(s)/Grant(s): F41624-98-C-6010
Report No.(s): AD-A440345; No Copyright; Avail.: CASI: A02, Hardcopy
To effectively study team variables as they impact performance in a particular domain, it is possible to develop medium
fidelity simulations that abstract some details of the performance environment while maintaining others. This paper reports the
results of a successful effort to create a synthetic task environment that captures key elements of a team task based on
Cognitive Task Analysis of the important features of the task from a teamwork and cognition viewpoint. The authors studied
the performance of AWACS Weapons Director (WD) teams and, based on the CTA data collected and insights from
133