Catálogo de publicaciones - libros

The ability of an IP backbone network to deliver robust and dependable communications relies on quickly restoring service after failures. Service-level agreements (SLAs) between a network service provider and customers typically include overall availability and performance objectives. To achieve the desired SLA, we have developed a methodology for the combined analysis of performance and reliability (performability) of networks across multiple layers by modeling the probabilistic failure state space in detail and analyzing different restoration alternatives. This methodology has been used to analyze large commercial IP-over-Optical layer networks. In this paper we extend our methodology to evaluate restoration approaches for an IP-based satellite backbone network. Because of the environment in which they operate (long delay links, frequent impairments), satellite networks pose an interesting challenge to typical restoration strategies. We describe the potential multi-layer restoration alternatives and compare their performability. Interestingly, while it is commonly thought that SONET ring restoration at the lower layer improves overall reliability, we find that it may not always improve performability in this environment.

Broadband satellite-based IP networks have been considered as the technology to enable a strong and promising next-generation market. In satellite communication systems, the channel performance might be severely degraded due to the dynamic weather conditions such as the precipitation, and thereby lead to network congestion. Therefore, congestion control is critical in such networks to satisfy QoS based Service Level Agreement (SLA). Moreover, the inherent large bandwidth-delay product of satellite channels impedes the deployment of existing numerous congestion control schemes. In this paper, we propose a modified Random Early Detection (RED) based congestion avoidance/control mechanism that incorporates a fuzzy logical controller to tune the queue thresholds of RED according to the dynamic weather conditions. We will show using analysis and simulations that the newly developed congestion control method is effective and efficient for broadband satellite-based IP networks.

Paths with loops, even transient ones, pose significant stability problems in networks. As a result, much effort has been devoted over the past thirty years to designing distributed algorithms capable of avoiding loops. We present a new algorithm, DIV (DIV), that guarantees that no loops, transient or steady-state, can ever form. DIV’s novelty is in that it is restricted to shortest paths, can easily handle arbitrary sequences of changes and updates, and provably outperforms earlier approaches in several key metrics. In addition, when used with distance-vector style path computation algorithms, DIV also prevents ; hence further improving convergence. The paper introduces DIV and its key properties. Simulation quantifying its performance gains are also presented.

This paper presents an algorithm for finding a robust routing plan in core networks that takes into consideration network topology, available capacity, traffic demand, and quality of service (QoS) requirements. The algorithm addresses the difficult problem in routing and traffic engineering of optimal path selection. Our approach is inspired by the concept of “between-ness” from graph theory, from which we introduce quantitative metrics for link and path criticality. Paths are ranked according to path criticality and the algorithm tries to avoid placing flows on the most critical paths, maximizes throughput over the short term in the presence of QoS constraints, and attempts to increase the bandwidth of the critical paths for future use. The proposed approach shows promise relative to previous proposals in simulations on benchmark and experimental networks.

Intra-domain traffic engineering is essential for the operation of an Internet Service Provider. Demand-oblivious routing [2] promises excellent performance guarantee with changing and uncertain traffic demands. However, it is difficult to implement it. We investigate an efficient and deployable implementation of oblivious routing. We study its performance by both numerical experiments and simulation. The performance study shows that the multipath implementation achieves a close approximation to oblivious routing [2], especially when approximate knowledge of traffic is available. The study also shows its robustness under varying traffic demands, link failures and an adversary attack. Its performance is excellent even with a 100% error in traffic estimation.

In this paper we develop a mathematical model to capture BGP table fluctuations. This provides the necessary foundations to study short- and long-term routing table growth. We reason that this growth is operationally critical for network administrators who need to gauge the amount of memory to install in routers as well as being a potential deciding factor in determining when the Internet community will run out of IPv4 address space.

We demonstrate that a simple model using a simple arrival process with heavy tailed service times is sufficient to reproduce BGP dynamics including the “spiky” characteristics of the original trace data. We derive our model using a classification technique that separates newly added or removed prefixes, short-term spikes and long-term stable prefixes. We develop a model of non-stable prefixes and show it has similar properties in their magnitude and duration to those observed in recorded BGP traces.

In todays Internet, a trend towards distributed content delivery systems can be observed. These systems still have to offer the same functionality as centralized architectures, e.g. content localization. While several advantages like load distribution and cost reduction could be gained by using a decentralized distribution platform, they come with a tradeoff in resource consumption. In this paper, we analyze and evaluate a decentralized Content Distribution Application (CDA) with respect to the time needed to locate specific content. We introduce a node model based on queueing theory and provide methods to compute important characteristics like the mean search time. In this model, we do not only consider the external offered load, but also the internal traffic created as a consequence of decentralization.

Optical Burst Switching (OBS) is a promising switching technique that has received much attention in recent years. A fair scheduling algorithm can ensure fair bandwidth allocation among different users and isolate ill-behaved users from normal users in a network. However, fair scheduling in OBS networks is difficult to implement due to lack of inexpensive and large buffers in optical domain. In this paper, we propose a new scheme to provide fairness in OBS scheduling by considering the time-slotted version of OBS, the Slotted OBS or SOBS. We give a fair scheduling algorithm with (1) time complexity called the . Our simulation results demonstrate that by applying the algorithm to the control packets (CP) at the end of each time slot, fair scheduling can be achieved. The scheme significantly improves the fairness among different flows compared to other scheduling algorithms in OBS such as LAUC-VF and Pipeline with VST. We also give an approach to supporting Quality of Service (QoS) in OBS based on the proposed scheme.

Optical networks with Wavelength Division Multiplexing (WDM), especially Optical Packet Switching (OPS) networks, have attracted much attention in recent years. However, to make OPS practical, it is crucial to reduce the packet loss ratio at the switching nodes. In this paper, we study a new type of optical switching scheme for OPS which combines optical switching with electronic buffering. In the new scheme, the arrived packets that do not cause contentions are switched to the output fibers directly; other packets are switched to receivers and converted to electronic signals and will be stored in the buffer until being sent out by transmitters. Since full range wavelength converters are difficult to implement with available technology, we focus on the more practical fixed wavelength conversion, with which one wavelength can be converted to another fixed wavelength. We will show with both analytical models and simulations that the packet loss ratio can be greatly reduced with few additional receivers and transmitters. Thus, we believe that the proposed hybrid switching scheme can greatly improve the practicability of OPS networks.

Wavelength-division-multiplexing (WDM), Slotted, Loss ratio, Optical Packet Switching, Buffering, Fixed wavelength conversion.

In this paper we investigate the maximum message throughput of the ActiveMQ server in different application scenarios. We use this throughput as a performance criterion. It depends heavily on the installed filters and the message replication grade. In previous work, we have presented measurement results and an analytical model for simple filters. This work extends these studies towards more complex configuration options. It provides measurement results and analytical performance models for complex AND-, OR-, and IN-filters. The results are useful to understand the performance of JMS servers and help to dimension large distributed JMS-based systems.