I am attending these workshops from the perspective of someone familiar with UNIX as a way of being introduced to Windows NT administration and software development. I have been doing UNIX Systems Administration and software development for about 15 years and have not yet done much with Windows.

As for the workshop for researchers, I think I will get exposure to issues such as Windows NT software development and considerations of upwards scalabilty of NT with respect to various scenarios. This will be valuable involvement I might have in scientific applications implementation, large database systems and distributed operating system research and application implementation and consultation. It will also directly relate to a research project I will be involved in, where various distributed systems approaches will be compared (DCOM, DCE and CORBA). A copy of the abstract for a research grant proposal that Bina Ramamurthy, a faculty member in CS, has submitted to NSF is attached. I will be involved in this project at several levels. This project will also benefit me in that I am actively involved in a campus-wide group (Distributed Computing Working Group) which is a vehicle for evaluation and recommendation as my University considers various distributed computing approaches.

Please feel free to contact me with any questions.

Harry DeLano
Computer Science
SUNY Buffalo
delano@cs.buffalo.edu

Abstract: Interoperability and Security Features for Open Distributed Computing

The main goal of this research is to study the features needed to build a a highly interoperable and secure architecture for open distributed computing. Open distributed computing refers to collaborative computing among many large, independent distributed computing environments. Such systems are commonly used in a wide range of application areas including mobile computing, simulation, design automation, enterprise-wide integrated computing, just to mention a few prominent ones. Distributed computing infrastructures offer virtually unlimited connectivity and accessibility to information. Connectivity is exploited to provide collaboration among the distributed components of the system. Such collaboration relies heavily on the extent of interoperability among the various components of the distributed system. Moreover, security is extremely critical for establishing trust among the interoperable environments and for warding off the vulnerability brought about by distributed control of resources. The two highly inter-dependent issues, namely, interoperability and security are the focus of this proposal.

A new interoperability protocol will be defined. It will specify how to field a request from a client, validate the request, select the appropriate server using the directory services, and translate and delegate the request to the selected server. The approach that will be used in realizing the interoperabilty protocol is to extend the facilities and services of currently available middleware packages. A prototype ODC architecture will be designed and developed to test the protocol. Three different environments (i) DCE on Solaris, (ii) CORBA on Solaris and (iii) DCOM on Windows NT have been selected to build the prototype. Specific issues to be addressed in this research include: (i) mechanisms for realizing interoperability, and (ii) mechanisms for handling insecurity that arises due to interoperation, and (iii) reducing the security checking overheads. Extensive performance evaluation under various request patterns, optimal client-server configurations, load balancing and scalability are some issues that will be addressed in the future extensions of the proposed research.

This research has other broader goals with educational connotations. These goals are: (i) involving undergraduates in scientific research, (ii) incorporation of topics of current interest in undergraduate curriculum using special projects, and (iii) building a model distributed systems laboratory.