Evangelos Vayias

Intelligent Agents for ATM Network Control and Resource Management: Experiences and Results from an Implementation on a Network Testbed

This paper presents the architecture and implementation of a distributed multi-agent system designed to provide flexible control and efficient resource management of a communications network. The system architecture enables different service providers and the network provider to use different control algorithms on the same logically partitioned physical infrastructure. All the mechanisms have been implemented and successfully tested on real experimental telecommunications networks. The system components are first described and then results from experiments and performance issues are discussed. The problem of interfacing a multi-agent system to network devices is then discussed. Interfacing agents with the network requires some form of programmable interface, since it is necessary, for example, to override conventional ATM signaling. At the moment programmable network device interfaces are not always available, particularly in commercially available devices and, thus, generic operations using management protocols have to be used. A generic architecture for embedding custom network control functionality into ATM networks is described.

A framework for designing ATM management systems by way of abstract information models and distributed object architectures

ATM network management has not yet reached the level of versatility and comprehensiveness exhibited by other aspects of this networking technology. Since the currently dominant general standards for broadband network management, primarily targeted towards large carrier networks, are too complex and intricate for smaller environments, e.g. LANs or corporate networks, ATM equipment manufacturers, in their effort to provide systems that exploit the power of ATM and yet remain practical and simple, usually resort to devising ad hoc proprietary extensions of simpler management frameworks, originally developed for other networking technologies. Such incompatible extensions, however, remain useless in the common case where the network is heterogeneous. The notion of abstract information modeling may be employed for improving on this situation. Building on this concept, the paper proposes a framework for developing ATM management systems intended for heterogeneous small- to medium-size networks. The general guidelines are illustrated through discussing a specific compliant management application intended for the remote monitoring of ATM network platforms and developed in the framework of a European research project. The application features a WWW interface and, as such, provides an example of the compatibility of information abstraction with the Web-based management techniques and of the benefits arising from the combination of the two notions. Lastly, and as the concept of abstracting information is inherent to recent frameworks for designing and programming distributed object systems, the paper explores this relation by discussing an alternative design of the monitoring application as a distributed object system.

CAC and traffic shaping for performance control in ATM: the two-class paradigm

Abstract Retaining only a few service classes in the ATM layer of broadband networks has the distinct advantage of simplifying the traffic control functions, which seems to be a key prerequisite for the successful introduction of ATM. In the paradigm under consideration, only two service classes are assumed, which share the storage and bandwidth resources at the network nodes according to an absolute-time-priority scheme, being able to support two general types of applications: real-time and non-real-time ones. The low-priority class is thus of the available-bit-rate (ABR) type, with some service quality guarantee obtained through allocating a minimum amount of bandwidth to it. Proper connection admission control (CAC) for both classes and traffic shaping of the low-priority traffic are employed in order to maintain contracted quality-of-service levels for both classes. Peak rates for the real-time traffic and effective rates for the non-real-time one form the basis of the control functions. A burst-level traffic model based on the M/D/1 queue is proposed for sources without strict delay constraints and the derived effective rate is validated by simulation and results from experiments on an ATM testbed. Following these results, a coherent CAC and traffic shaping scheme is presented. This scheme makes use of the proposed burst-level model to achieve an adaptive shaping of the non-real-time sources. For all links intended to support both classes, real-time sources are allocated, their demanded peak rate and the bandwidth left available by these sources is exploited by the ABR type sources (non-real-time), through adapting the parameters of the shaper. Simulation results validate the concept of this novel ABR-like service. The set of simulation and experimental results contained in the paper constitute the heart of its contribution.

Packet-level traffic control for the internet: a framework and its experimental validation

This paper presents a theoretical framework, which can be the basis for traffic control and resource management mechanisms, that provide QoS with quantitative guarantees in the Internet. The starting point is a fairly general traffic shaping algorithm for effective rate enforcement which is based on a packet spacing law. Multiplexing several independent shaped streams results in a tight upper bound to the queue-length distribution at the multiplexer. The algorithm effectively works with any kind of input traffic (non-stationary, non-Markovian, heterogeneous, long-range dependent (LRD), etc.). Based on the definition of shaped streams, we devise mechanisms for aggregating, splitting and policing such streams without affecting the queuing performance at subsequent network nodes. Also, a calculus for end-to-end QoS in this framework is presented. The theoretical statements are supported by simulation and experimental results on a network testbed. Since the enforced effective rate features the convenient additivity property Σ i f i = C , it can be handily used for many traffic control and accounting functions, like policing, admission control, bandwidth allocation and charging.

Management architecture in tactical communication networks

The objective of this paper is not only to monitor, configure and control the immediate situation, but also to continually collect and assess information-so inter-and intra-service communications must be available throughout the chain of command. Vital decision-based on the current status and location of every available resource-must be centrally coordinated. Military communications should be quickly reconfigured with optimum survivability potential regardless of the particular configuration. Not only the initial speed of response, but also the rate at which fresh intelligence reports can be assimilated is crucial. To apply this intelligence in the most effective manner requires a unique mix of communications technology, systems integration and software know-how. Without fail-safe communications, todays command and control centers would be unable to monitor operations, assess developments, allocate resources, establish plans or execute actions. To meet these communication needs, systems engineers analyze the requisite traffic load and establish variable connections between a complex set of components ranging from sensors to voice and data transmission links, data processing centers and distribution networks.

An Approach to Network Control and Resource Management Based on Intelligent Agents

Network Control is currently carried out mainly by means of signalling protocols. Although these protocols are robust and facilitate standardisation, they present several drawbacks such as the inability of the Service Providers to adapt network resource control to the particular needs of their services, or the difficulty in deploying advanced traffic control schemes in order to achieve increased utilisation. An open distributed software architecture for Network Control based on concepts such as Intelligent Agents can provide a solution to this problem. This paper highlights the major issues of this problem and presents the architecture and implementation of a system that enables open and flexible network control and resource management using intelligent agents. The description of the architecture is complemented by experimental results as well as a generic implementation framework for deploying open programmable control capabilities in existing networks.

The design of an object-oriented simulation tool for evaluating ATM network resource control schemes

Changing the algorithmic components of ATM equipment such as ATM switches, requires full access to the switch control software, which is not available in most commercial products. Hence, most researchers still resort to simulation in order to evaluate new traffic control algorithms. This paper presents the architecture of an ATM simulator that operates at the call level, exposes the power of such a tool in evaluating different control architectures for ATM networks, and presents the status of a prototype implementation. ATM simulators based on the proposed architecture include a Web-based component (Java applet) which acts as user interface. This component enables tele-training and dissemination of important simulation results to remote sites.