Have you ever been curious about what “quality of service” (QoS) really means, with regard to Internet services? Browsing through this collection of high-quality conference papers, and reading some of them, will provide the novice reader with good insight into this area. Meanwhile, the expert networking professional will gain valuable knowledge about proposed architectures, protocols and methods, simulation results, and comparative analysis for the next generation network (NGN).
The first 22 research papers were presented at the Fifth International Workshop on Quality of Future Internet Services (QofIS 2004), which took place in Barcelona, Spain in September 2004. The first session was dedicated to QoS for two specific Internet applications: peer-to-peer networks (P2P) and voice over IP (VoIP). The first P2P paper presents an analytical model of file distribution, showing that the service capacity grows exponentially with the number of chunks a file consists of. The second P2P paper proposes an application-level multicast protocol for file distribution that allows for more efficient bandwidth usage. Three papers deal with VoIP—assessing the perceptual quality of voice transmission, analyzing the impact of using forward error correction (FEC) in VoIP flows over the Internet, and proposing a multiprotocol label switching (MPLS)-based NGN architecture that can guarantee end-to-end QoS.
The papers from the second edition session discuss local area and ad hoc wireless networks. The first two present analytical models for researching the 802.11 backoff and distributed coordination function (DCF) delays that are more accurate than previous ones. The third paper proposes a new protocol that improves the QoS in ad hoc networks, while the last one in this section analyzes some medium access control (MAC) policies for such networks.
The four papers from the third session analyze various service differentiation and congestion control issues in a transmission control protocol/Internet protocol (TCP/IP) environment. These include an interesting priority algorithm, a novel management technique that can provide soft service guarantees to aggregate traffic, a proposed intermediate marker in a three-color QoS scheme, and an additiveincrease/multiplicative decrease (AIMD)-based TCP load-balancing architecture.
In the fourth session, dedicated to traffic engineering (TE) and routing, five research papers were presented. The first studies a model that uses a hybrid interior gateway protocol plus MPLS (IGP+MPLS) routing approach to achieve an efficient routing of packet flows. The second one proposes a genetic algorithm to increase the weight of loaded links and the cache hit ratios. The third one describes a prediction-based mechanism for lightpath assignment in optical networks, which reduces the required overhead. The next one proposes a novel algorithm for the deployment of overlay servers on the Internet, in order to assure QoS. The last paper from this session compares various proposed extensions of MPLS of open shortest path first (OSPF) routing protocols, with respect to network resilience.
The fifth session, “Enforcing Mobility,” contains four papers, covering the integration of handoff with a differentiated services (DiffServ) mechanism, a novel wireless channel modeling using cross-layer mappings, an efficiency analysis of AA2/MPLS for different traffic types, and an extensive analysis of QoS in Bluetooth-based sensor networks.
Colocated with QofIS 2004, the first Workshop on Quality of Service Routing (WQoSR 2004) was motivated by the challenges that must be addressed by QoS routing, to enable the support of advanced Internet services. From 28 submitted papers, eight were selected and organized into two sessions.
Four papers were presented in the first session, “Algorithms and Scalability Issues.” The first paper presents a novel distributed algorithm that provides QoS by only enabling free-of-cycles routes, which are known to ensure network stability. The second paper presents a method to facilitate the scale-up of multihomed systems. The third paper analyzes service differentiation in an optical packet switched network, and the last proposes a novel approach to inter-domain QoS routing.
The “Novel Ideas and Protocol Enhancements” session collects innovative papers in this research area. These papers present a policy-based connectionless routing architecture and algorithms, an algorithm that solves the combined routing and bandwidth allocation problem for the case in which flows may be split along several paths, a new metric for QoS in MPLS traffic engineering, and a framework for QoS routing for wireless sensor networks, based on learning-based meta-strategies.
The fourth International Workshop on Internet Charging and QoS Technology (ICQT 2004) was also colocated with QofIS 2004. Eight papers were arranged into three sessions.
The “Auctions and Game Theory” session contains three research papers dedicated to auction-based models and strategies—for a hierarchical two-level bandwidth allocation problem, and for another bandwidth allocation problem in a stochastic environment. A mechanism to provide an end-to-end guarantee of services is also discussed. The last paper in this session is dedicated to a game-theoretic reexamination of the TCP Vegas congestion control mechanism.
The “Charging in Mobile Networks” session proposes a novel roaming settlement model of sharing and bartering excess capacities, and analyzes Web content prefetching in universal mobile telecommunications system (UMTS) third-generation cellular networks.
The last two papers are grouped in the “QoS Provisioning and Monitoring” session. Both papers analyze, from different points of view, novel QoS architectures and concepts between service providers.
To conclude, these papers provide a state-of-the-art discussion, from various eclectic approaches, of this rich research area. I highly recommend the book.
