Ferenc Vajda

From Cluster Monitoring to Grid Monitoring Based on GRM

GRM was originally designed and implemented as part of the P-GRADE graphical parallel program development environment running on supercomputers and clusters. In the framework of the biggest European Grid project, the DataGrid we investigated the possibility of transforming GRM to a grid application monitoring infrastructure. This paper presents the architectural redesign of GRM to become a standalone grid monitoring tool.

A data-driven algorithm and systolic architecture for image morphology

The paper presents a systolic processor for binary and gray scale morphological operations. The principle of the proposed implementation scheme is based on an input data-driven algorithm: the envelope scan method (ESM) in cooperation with a new formalization of the window-operation provides promising, cost-effective logic gate implementation introduced as the envelope scan processor (ESP). It is showed that the proposed architecture can also be easily applied to adaptive operations, since the SE-adaptation is directly supported by the algorithm. Analysis shows that the ESP offers remarkable results in terms of the utilized logic components in comparison to other recent processor designs.<<ETX>>

An Efficient Implementation Technique of Adaptive Morphological Operations

In the paper the implementation aspects of mathematical morphology based on a new formalization of the window operation is presented. The proposed technique called the Reverse Window Operation (RWO) can also be applied to adaptive morphological operations. In order to reduce the control complexity an algorithm is proposed, which uses a region-based parameter modification for the structuring element adaptation. Finally, an application of a data-dependent processing technique using the adaptive filter scheme is demonstrated.

Efficient morphology architecture

Following a brief introduction of mathematical morphological operations a novel architecture called boundary scan processor is introduced and its application for binary and gray-scale morphology is demonstrated. The hardware complexity of the processor is analyzed and compared with other recently published architectures. Typical examples in space (time) variant and adaptive (data dependent) morphology are also given.

Concurrent systems, programming primitives and languages: a comparative study

Abstract Key issues of communication by message passing are discussed. Two languages such as Occam and Ada which provide different approaches for process interaction are compared. Language notations for message passing (process vs. task) and for specifying process interaction (channels and input/output processes vs. remote procedure call) are surveyed. Specification, naming requirement, synchronization, nondeterminism are contrasted. Defining three levels of description such as virtual, logical and physical level, three implementations such as the Multibus Interprocessor Protocol (MIP), the Personal SuperComputer (PSC) and a Transputer Net (TN) are discussed and compared.

Implementation issues of image texture analysis and segmentation

Abstract After a brief description of the most popular image texture analysis algorithms the paper deals with their calculation complexity. Considering their direct implementation based on DSPs, superscalar RISc processors, ASICs, algorithmically dedicated standard parts and FPGAs, the features of these implementation categories are compared.

Session 4: Java and Jini [breaker page]

his session deals with problems related to Java and/or Jini. Nowadays both of them refer to areas of hot topics and they are used as buzzwords. Therefore everybody who is very far from these research areas knows that Java is not a coffee or an island near a famous vulcano and Jini is not a jinni who just left a bottle to follow the orders of his master. According to the creator of the language Java is meant to provide a network-centric computer platform that had among other characteristics the ability to allow programmers to write applications once and run them on virtually any hardware or software architecture. A little more detailed it includes an object-oriented programming language, a platform independent bytecode and a virtual machine specification. Java is a robust, secure environment, which provides a new platform and a new approach to distributed computing. Built on the top of the Java software infrastructure, Jini technology enables different services and devices to work together in a community. It is organized without extensive planning, installation or human intervention, as Jini technology-based services are accessible either directly or through surrogates written in Java language. The formal Jini specifications describes among other things the discovery and join protocols, lookup services, distributed leasing concepts and event programming and transaction models. It provides an infrastructure and a programming model to federate services into a single, dynamic distributed system. This session has five papers dealing with quite different aspects of distributed Java and Jini-based systems. The paper by D. Cotroneo, C. Di Flora and S. Russo deals with a Jini-based framework called PRINCEPS meaning a Plugable Reliable Infrastructure for Network Components and Enhanced Properties of Service. The framework provides a mechanism for selecting services according to functional and non-functional requirements. CORBA-based, socket-based and centralized services can be considered. The second paper of the session by A. Bouchi, B. Toursel and R. Olejnik titled “An Observation Mechanism of Distributed Objects in Java” discusses an execution environment for distributed applications called ADAJ (Adaptive Distributed Application in Java). It presents an observation mechanism by counting the number of method invocations for global objects. It concentrates to irregular applications developed in distributed Java. The subject of the third paper by Z. Laszlo is mobile agents in a distributed database system. It presents an infrastructure for multimedia database searches based on CORBA and mobile agent technology. The system was implemented in Java and can be run in the Oracle database. The fourth paper by J-H Kim and R. S. Ramakrishna is investigating the problem of load distribution in Jini. It proposes a load distribution mechanism called JINT and uses CORBA event service for reporting load information. Some experimental results are also provided for the application of the JINT approach. The last paper of the session is titled “Assignment Schemes of Multiple Services in Jini” by V. Georgiev and V. Getov. It introduces and compares different schemes for assignment of multiple services in Jini and also presents four levels of system information gathering. The performance modeling and evaluation is based on queuing systems and contributes to developing a methodology for building Jini-based large applications. The papers of the session present interesting novel proposals, solutions and experiments for distributed systems based on Java implementations and applications of Jini as middleware. T