Werner Zorn

Developing network management applications in an application-oriented way using mobile agent

Networks are becoming larger and more and more complex. Current management paradigms are unsuitable for them. As a new distributed computing paradigm, a mobile agent has many special features. It offers a very promising solution for many problems in the current network management architecture. Because a mobile agent is an autonomous and intelligent program unit, developing management application using mobile agent follows an application-oriented way. In this way, the developer only needs to focus on the functionality of applications. In this paper, we first investigate the paradigm of network management by a mobile agent (MbMA), its fundamental architecture and advantages. After that, we give a general taxonomy of mobile agent and examples in the context of network management.

End-to-end transactions in three-tier systems

Many modern enterprise applications follow the three-tier architecture. Typical three-tier systems have some nice properties: clients are thin, application states are stored in a database, and the middle servers are stateless. This makes three-tier systems well scalable and manageable. However, it is challenging to implement end-to-end reliable and safe transactions for such applications without violating their nice properties. We present a pragmatic end-to-end transaction protocol which ensures that every issued request will be processed exactly-once by the server(s). The idea is to use client log-based recovery and application server replication. The application server remains stateless; all replicas play the same role and can serve client requests. Therefore, the system is well scalable and the failover is fast.

Parallel Knowledge Processing in Open Networks

The objective of this paper is the description of WAVE model and language for highly parallel processing in open computer networks, as well as an efficient distributed implementation of the language. The WAVE language interpreter may be a basic building block for parallel knowledge processing systems working without any centralised resources in an integral logic & data flow mode. Application of the model for intelligent management of distributed computer networks, based on integration of the WAVE interpreters with traditional network management tools, is discussed.

A Role-Based Access Control Model and Implementation for Data-Centric Enterprise Applications

Access control is concerned with limiting the activity of legitimate users in an application. Role-based access control (RBAC) uses role to indirectly describe the access rights. This indirectly mapping is very flexible. However, current RBAC models are not suitable to describe fine-grained access control for data-centric enterprise applications. In this paper, we present a pragmatic role-based access control model for data-centric application. The access control is fine-grained and flexible. An object-oriented implementation is also presented. Users are identified by digital certificates. The proposed model is designed for three-tier enterprise application.

Simulation of Traffic Engineering in IP-Based Networks

A practical function of traffic engineering in IP-based networks is the mapping of traffic onto the network infrastructure to achieve specific performance objectives in respect of traffic-oriented and/or resource-oriented. The simulation of such traffic engineering function is necessary needed while planning and optimizing the network resources. This paper describes an new model for analytical simulation of the traffic mapping in IP-based networks in respect of both real-time and best-effort traffic. We mainly conceptualize the simulation as an behavioral modeling of network dynamics. We then mathematical formulate this behavioral modeling as two sub-optimization tasks corresponding to both real-time and the best-effort traffic delivering. As an example of our algorithm implementations, an genetic algorithm for real-time traffic mapping is presented. The article also gives some numerical results from practical examples.

Resource Planning in Converged Networks

Network resource planning aims to determine cost-effective resources needed to carry current volumes of traffic. Determining where and when to add bandwidth or expand routing or switching capacity are the most fundamental resource planning decisions. This paper describes this problem in converged networks and an model-based approach as an efficient solution for it. Computational tests of the approach with different backbones ware carried

Generic Management Architecture (GEMA) — ein neuer Netzwerkmanagementansatz

Heutige Netzwerkmanagementsysteme lassen sich in die zwei Klassen zentralisiertes und verteiltes Management unterteilen. Selbst Systeme, die der zuletztgenannten Gruppe anzurechnen sind, basieren vorwiegend auf einer statischen Verteilung funktionaler Elemente, die im Client-Server-Konzept interagieren. In Erganzung zu diesen existierenden Ansatzen wird hier eine flexiblere Architektur vorgestellt, die neben den bekannten Vorteilen der Verteilung von Anwendungen zusatzlich die Komplexitatsreduzierung der Managementapplikationen durch Einfuhrung der objektorientierten Paradigmen der Vererbung und Redundanz beinhaltet.

Mehrwertdienste in Rechnernetzen — Technik und Anwendungen

Mehrwertdienste (VAS) sind alle Netzdienste, die dem zugrundeliegenden Transportnetzwerk wesentliche Leistungsmerkmale hinzufugen. Hierzu zahlen einmal die klassischen Basis-Dienste wie Electronic Mail, News, Directory Service, zum andern darauf aufsetzende Dienste fur den Austausch von Dokumenten, Faksimilen, Graphik, den Dialogzugriff auf Datenbanken u.v.a.m. Unter der Bezeichnung CSTA (Computer Supported Telefonic Application) wachsen derzeit die rechner- und telefon-orientierten Mehrwertdienste zusammen. Der Beitrag behandelt die Techniken, inkl. Administration, stellt Nutzenbetrachtungen an, und versucht, einen Uberblick uber die Marktentwicklung zu geben.

Zugang zu internationalen Netzen

Beim Blick auf die internationale Netzlandschaft stellt man fest, das derzeit eine erhebliche Anzahl von Rechnernetzen im Wissenschaftsbereich existieren (ARPANET, CSNET, BITNET u.a.), die sowohl in standigem Wachstum, als auch in einer zunehmenden Verflechtung untereinander begriffen sind. Aus deutscher Sicht stellt sich die Frage, wie neben dem Aufbau einer nationalen Infrastruktur Anschlusse an diese Netze realisiert werden konnen.

Systemoptimierung mit Hilfe von synthetischen Belastungsprofilen

Bei dem zu optimierenden System handelt es sich um das Rechenzentrum der Fakultat fur Informatik an der Universitat Karlsruhe. Hierbei sollen folgende, in Wechselwirkung zueinander stehende Teilsysteme unterschieden werden: Hardware Betriebssystem Benutzerauftrage Den Kern der Hardware bildet eine BURROUGHS B6700 Doppelprozessoranlage mit 208 k Kernspeicher, 40 M Bytes Uberlagerungsspeicher und 360 M Bytes Hintergrundspeicher. Daran angeschlossen sind uber einen speziellen Datenkommunikationsprozessor ca. 60 Dialogstationen, in der Mehrzahl alphanumerische und graphische Bildschirmgerate. Besonderheiten der Hardware-Architektur sind Kellerstruktur mit Display-Registern, eintrittsinvariante Code in der Syntax der polnischen Notation, virtueller Speicher mit Deskriptortechnik.