We discuss an implementation of QoS routing extensions to the open shortest path first (OSPF) routing protocol and evaluate its performance over a wide range of operating conditions. Our evaluations are aimed at assessing the cost and feasibility of QoS routing in IP networks. The results provide insight into the respective weights of the two major components of QoS routing costs, processing cost and protocol overhead and establish strong empirical evidence that the cost of QoS routing is well within the limits of modern technology and can be justified by the performance improvements.

Considers the support of hard real-time connections in ATM networks. In an ATM network, a set of hard real-time connections can be admitted only if the worst-case end-to-end delays of cells belonging to individual connections are less than their deadlines. There are several approaches to managing the network resources in order to meet the delay requirements of connections. This paper focuses on the use of traffic regulation to achieve this objective. Leaky buckets provide simple and user-programmable means of traffic regulation. An efficient optimal algorithm for selecting the burst parameters of leaky buckets to meet connections' deadlines is designed and analyzed. The algorithm is optimal in the sense that it always selects a set of burst parameters whose mean value is minimal and by which the delay requirements of hard real-time connections can be met. The exponential size of the search space makes this problem a challenging one. The algorithm is efficient through systematically pruning the search space. There is an observed dramatic improvement in the system performance in terms of the connection admission probability when traffic is regulated using this algorithm.

Recent studies provide evidence that Quality of Service (QoS) routing can provide increased network utilization compared to routing that is not sensitive to QoS requirements of traffic. However, there are still strong concerns about the increased cost of QoS routing, both in terms of more complex and frequent computations and increased routing protocol overhead. The main goals of this paper are to study these two cost components, and propose solutions that achieve good routing performance with reduced processing cost. First, we identify the parameters that determine the protocol traffic overhead, namely (a) policy for triggering updates, (b) sensitivity of this policy, and (c) clamp down timers that limit the rate of updates. Using simulation, we study the relative significance of these factors and investigate the relationship between routing performance and the amount of update traffic. In addition, we explore a range of design options to reduce the processing cost of QoS routing algorithms, and study their effect on routing performance. Based on the conclusions of these studies, we develop extensions to the basic QoS routing, that can achieve good routing performance with limited update generation rates. The paper also addresses the impact on the results of a number of secondary factors such as topology, high level admission control, and characteristics of network traffic.

In recent years, a number of link scheduling algorithms have been proposed that greatly improve upon traditional FIFO scheduling in being able to assure rate and delay bounds for individual sessions. However, they cannot be easily deployed in a backbone environment with thousands of sessions, as their complexity increases with the number of sessions. In this paper, we propose and analyze an approach that uses a simple buffer management scheme to provide rate guarantees to individual flows (or to a set of flows) multiplexed into a common FIFO queue. We establish the buffer allocation requirements to achieve these rate guarantees and study the trade-off between the achievable link utilization and the buffer size required with the proposed scheme. The aspect of fair access to excess bandwidth is also addressed, and its mapping onto a buffer allocation rule is investigated. Numerical examples are provided that illustrate the performance of the proposed schemes. Finally, a scalable architecture for QoS provisioning is presented that integrates the proposed buffer management scheme with WFQ scheduling that uses a small number of queues.

An important challenge for the future growth of the Internet is to design routers that can forward the exponentially increasing volume of traffic, and at the same time provide the service differentiation needed by new applications. In this paper, we describe the architecture, implementation, and initial experiences with a system designed to meet this challenge. This system, which we call a QoS capable switch-router (QSR), combines the salient features of switching and routing technologies to provide high throughput and support the different classes of service being defined by the IETF. It consists of a core (ATM) switch fabric connecting intelligent adapters, each capable of both routing and switching pockets. A control engine is responsible for routing, RSVP signalling, and resource management. We have built a prototype network of 3 systems connected to several UNIX hosts, and have conducted preliminary performance measurements on this network.

A CAC algorithm must efficiently determine if a new connection can be admitted by verifying that its QoS requirements can be met without violating those of previously admitted connections. In hard real-time systems, the QoS requirements are specified in terms of end-to-end cell deadlines and no cell loss due to buffer overflow. A CAC algorithm must account for interdependencies among connections caused by statistical multiplexing of cells in ATM networks. Arbitrarity of network topology may lead to cyclic dependencies among various connections. We present an efficient CAC algorithm that addresses the above issues. The algorithm uses a traffic descriptor called the maximum traffic rate function to effectively compute bounds on end-to-end delays of connections and buffer requirements within the network. Our work differs from most previous work in that it does not require traffic restoration inside the network.

We study a new network architecture based on standard FDDI networks. This network, called FDDI-based reconfigurable network (FBRN), is constructed using multiple FDDI token rings and has the ability to reconfigure itself in the event of extensive damage to the network. Thus, an FBRN has the potential to provide high available bandwidth even in the presence of numerous faults. Realization of this potential depends crucially on a reconfiguration algorithm that guides the reconfiguration process. We design and analyze a reconfiguration algorithm for FBRNs. Our algorithm is optimal in the sense that it always produces a configuration that results in the maximum available bandwidth for a given fault pattern. This algorithm has a polynomial time complexity. We also show that the available bandwidth of an FBRN is dramatically improved with our reconfiguration algorithm.

We study a new network architecture called FBRN (FDDI-based reconfigurable network). An FBRN consists of multiple FDDI token rings and has the ability to reconfigure itself in the event of network faults. Thus, an FBRN has the potential to provide high available bandwidth even in the presence of numerous faults. Realization of this potential depends crucially on the choice of reconfiguration algorithm. We design and analyze a distributed reconfiguration algorithm for FBRNs. Our algorithm is optimal in the sense that it always produces a configuration that has the maximum available bandwidth possible for the given fault pattern. This algorithm has polynomial time complexity and hence can be efficiently implemented. We evaluate the algorithm's performance in terms of the available bandwidth in the network as a function of the number of faults.

We address the issue of guaranteeing the end-to-end deadlines of hard real-time connections in an ATM network. In an ATM network, a set of hard real-time connections can be admitted only if the end-to-end delays of cells belonging to individual connections are not more than their deadlines. We systematically decompose an ATM network into constant delay and variable delay servers to facilitate the delay analysis. Effective traffic description is the key part of such a process. We propose a comprehensive traffic description function that provides adequate information about the worst case traffic behavior of connections anywhere in the network. We also study some simple approximations of this function that perform reasonably well in practice. We analyze and compare the performance of ATM networks with FCFS and WRR link scheduling policies under different loading conditions.

This paper considers the support of hard real-time connections in ATM networks. In an ATM network, a set of hard real-time connections can be admitted only if the worst case end-to-end delays of cells belonging to individual connections are less than their deadlines. Although there are several approaches to manage the network resources in order to meet the delay requirements of connections, we focus on the use of traffic regulation to achieve this objective. Leaky buckets provide simple and user-programmable means of traffic regulation. We design and analyse an efficient optimal algorithm for selecting the burst parameters of leaky buckets to meet connections' deadlines. Our algorithm is optimal in the sense that it always selects burst parameters to meet the delay requirements of hard real-time connections whenever some such assignment exists. The exponential size of the search space makes this problem a challenging one. Our algorithm is efficient and we observe a dramatic improvement in the system performance in terms of the connection admission probability when traffic is regulated using our algorithm.

We address the issue of guaranteeing application-to-application deadlines of messages in distributed real-time systems. Most of the previous studies have focused on either host subsystem or network subsystem. Our study considers the entire integrated system. We develop two methods to test whether the message deadlines are met. The first method called the independent method, obtains conservative estimates of message delays by independently analyzing each subsystem. This method is easy to use and efficient, but may sometimes give pessimistic results. The second method called the integrated method, computes the delays more accurately by modeling interactions among the subsystems in greater detail. Performance of the two methods is evaluated an terms of the probability of guaranteeing message sets for given host utilization.

FDDI-Based Reconfigurable Networks have an architecture that is suitable for delivering messages that have hard real-time constraints as well as certain fault-tolerance requirements. This architecture uses multiple FDDI networks to connect hosts and provides for automatic reconfiguration to maintain high network bandwidth in spite of faults. An important open problem is how resources in such networks should be managed in order to guarantee that the fault-tolerant real-time requirements of messages are met. This paper presents an efficient and practical solution to this problem. Our solution consists of off-line and on-line components. On-line management deals with run-time manipulation of messages and network resources. A message grouping approach simplifies on-line management. Off-line management deals with message grouping, bandwidth allocation and schedulability verification. Three approaches are investigated: spatial redundancy, temporal redundancy and an integrated approach. It is shown that the integrated approach has the best performance. Our solution is compatible with the FDDI and SAFENET standards.

Networks used in mission-critical applications must be highly fault-tolerant. The authors propose an FDDI-based reconfigurable network (FBRN) that can survive multiple faults. An FBRN consists of multiple FDDI trunk rings and has the ability to reconfigure itself in the face of extensive damage. They consider three classes of FBRNs depending on their degree of reconfigurability: N-FBRN (nonreconfigurable), P-FBRN (partially reconfigurable) and F-FBRN (fully reconfigurable). They analyze the performance of FBRNs in terms of their available bandwidth in the presence of faults. In mission-critical systems, it is not enough to ensure that the network is connected in spite of faults; the traffic carrying capacity of the network must be sufficient to guarantee the required quality of service. They present a probabilistic analysis of the available bandwidth of FBRNs given the number of faults. They find that a fully reconfigurable FBRN can provide a high available bandwidth even in the presence of a large number of faults. A partially reconfigurable FBRN is found to be an excellent compromise between high reliability and ease of implementation.>

We study the performance of FDDI networks in terms of their guarantee probability, i.e., the probability that a set of synchronous messages are guaranteed to meet their deadlines. Traditional techniques such as queuing analysis cannot be directly used to derive the guarantee probability. To counter this problem, we develop a new geometric model of schedulability. Based on this model, we obtain a numerical method to compute the exact values of the guarantee probability. A closed-form approximation for the guarantee probability is also derived, and is shown to be relatively accurate and computationally efficient. The network performance is then systematically examined in terms of the guarantee probability. We find that there is a high probability that a randomly chosen message set can be guaranteed even when the real-time traffic is increased beyond the worst case achievable utilization bound. Hence, FDDI networks are applicable for real-time applications in a wide range of loading conditions.>