Download >>> https://tinurli.com/28181p
This book constitutes the revised selected papers of the 7th IFIP TC6 Conference, TSIF 2014, held in Berlin, Germany in June 2014. The problems and concepts that were covered and discussed at the conference ranged from network protocols to improving system reliability. The topics are grouped into five sections: Communication Infrastructure Protocols, Infrastructure Management Systems, Network Security & Safety, System Reliability Engineering & Design Tools and Techniques. The aim of this conference was to identify new ideas for better communication systems with increased safety. This scientific event offered a forum for academics and industry professionals to meet together, exchange their knowledge on different aspects of communication networks and systems engineering with delegates coming from all over the world including countries such as Brazil, China or USA. The book is organized in five sections: Communication Infrastructure Protocols, Infrastructure Management Systems, Network Security & Safety, System Reliability Engineering & Design Tools and Techniques. In the Communication Infrastructure Protocols section, reliability is a very important issue when designing protocols. In this section three papers were presented about different protocols with the aim of increasing the network availability. In [1], a new congestion control algorithm was proposed based on Explicit Congestion Notification (ECN). ECN-based end-to-end congestion control algorithms have been proposed to reduce network overload and improve network performance. However, few of them exploit ECN marking for end-to-end congestion control. Scukanec et al. proposed a new congestion control algorithm that takes advantage of ECN marking from routers. In [2], a novel traffic shaping mechanism is proposed, which provides a flexible and guaranteed bandwidth to a group of users. This mechanism, known as PIE (Proportional Integral Evolutionary) traffic shaping, considers the relationship among packet delivery delay, group throughput and group delay. The proposed mechanism can shape HTTP(S) request packets into different time periods according to the offered network bandwidth ratio between TCP flows. With this technique it is possible to guarantee that each TCP flow has an equal share of bandwidth during periods of congestion. Finally, in [3], the authors propose an algorithm for controlling router buffer sizes in order to reduce buffer overflows. This problem is relevant when using protocols such as TCP where packets arrive in bursts and the network has to keep a buffer for each connection. Although such buffers provide important advantages in terms of reduced delay and jitter, they cause the router to overload when packets come in bursts. In [4] is proposed a framework that unifies structured and unstructured information in a hierarchical way with speed and accuracy. A hierarchical model with flexible levels of granularity is proposed and tested on different data sources. It considers two kinds of information: structured (e.g., logs, registries or tables) and unstructured (e.g. e-mails, chat conversations). The authors propose a model able to automatically assimilate both structured and unstructured information. Thus, also the number of bytes written on disk is smaller than with other models. In [5] it is presented another framework for finding informative patterns in data. For example data in computer systems and networks and in medicine can be used to understand and deal with health problems. Depending on the kind of problem it may be possible to find patterns in different kinds of data, for instance by looking at several medical records containing information about different patients. For instance, it has been shown that heart attack patients can show specific signs that could not be found by looking at their medical records separately. cfa1e77820
Kommentare