Explore chapters and articles related to this topic
Quality of Service in Switch/Routers
Published in James Aweya, Designing Switch/Routers, 2023
Differentiated services (DiffServ) is a more recent QoS architecture defined by the IETF [RFC2474] [RFC2475] [RFC3260]. DiffServ specifies a simple and scalable coarse-grained, class-based mechanism for traffic forwarding in IP networks. It defines a mechanism for classifying and marking packets as belonging to a specific CoS. It includes the concept of marking IP packets with priority values that allow routers to classify the marked packets into traffic classes, each associated with certain forwarding behaviors in the network. Using DiffServ-based classification, each packet is placed into a limited number of traffic classes (or queues), compared to the fine-grained, flow-based architecture (Integrated Services (IntServ) architecture [RFC1633]) where network traffic is differentiated based on the requirements of an individual flow. DiffServ allows each router on the network to differentiate traffic based on the limited range of packet markings and traffic classes. Each traffic class can be managed differently, allowing routers to provide preferential treatment for higher-priority traffic on the network.
5G Edge-Based Video Surveillance in Smart Cities
Published in Zoran S. Bojkovic, Dragorad A. Milovanovic, Tulsi Pawan Fowdur, 5G Multimedia Communication, 2020
DiffServ layer 3 (L3) can be used for both traffic marking and classification. DiffServ-supported networks mark their traffic so the router knows which service to apply to the packet. By setting a field called the differentiated services code point (DSCP), the marking is done in the IP header, and the intelligence of a DiffServ network is set up in the routers. Each DSCP value represents a QoS class, also known as a behavior aggregate. A particular DSCP value is mapped to a particular routing behavior. This behavior is known as a per-hop behavior (PHB) and is implemented in the router using different queuing disciplines. Network elements that are DiffServ-compliant must conform and implement the specifications of the PHB. The classes usually are CS0 (000000 or 0) – CS7 (111000 or 56); higher classes provide increasingly better service treatment. For instance, video surveillance live stream standard value for DSCP is 32 decimal (or 100000 binary). However, users can select different value for live surveillance, DSCP 40 as suggested in Ref. [10]. Less strict DSCP classification is used if we deal with the transmission of recorded video signal that should be replayed – because recorded video typically uses TCP protocol.
Ad Hoc Networks
Published in Syed Ijlal Ali Shah, Mohammad Ilyas, Hussein T. Mouftah, Pervasive Communications Handbook, 2017
Jamal N. Al-Karaki, Ibrahim Al-Oqily
One of the QoS routing solutions for single-channel but multi-hop AP-based WMNs is proposed in [5]. The paper in fact proposes a DiffServ architecture for WMNs which also includes QoS provisioning at both network and MAC layers. DiffServ architecture defines three types of classes: (i) expedited forwarding (e.g., real-time data), (ii) assured forwarding (e.g., reliable delivery), and (iii) best effort (normal delivery). The routers differentiate the packets based on these classes and the packets within the same class are aggregated and forwarded to the upper AP as shown in Figure 4.3. Note that this mechanism is different than providing per-flow guarantees. The QoS routing protocol has two functions in this architecture. It should determine the route as well as the wireline gateway where the data will be forwarded. The paper proposes an on-demand QoS routing protocol which has four components: load classifier, path selector, call admission control (CAC) routine, and route repair routine. Load classifier monitors the traffic load and based on the load the path selector can switch to another path. The path selector mainly determines the gateway and a path to that gateway. The path is found through GPSR [14] protocol. Once it is found, CAC procedure is run to check whether the path satisfies the bandwidth requirements. If not, path selector should find another path with the given requirements. Route repair routine is used whenever a QoS path is broken. An interesting extension to this paper would be to consider multi-hop channels when designing the DiffServ and the corresponding QoS routing protocol.
A QoS guarantee approach for multimedia software system
Published in International Journal of Computers and Applications, 2020
For QoS guarantee, traditional rule-based methods [10] require explicit margins of QoS parameters; so, the scalability and efficiency cannot be satisfied. In order to address these issues, end-to-end approaches for QoS guarantee were proposed. Active approaches require the injection of probe packets into the network. The pioneering active approach [11] traceroutes between 37 participating sites are collected and analyzed to characterize the end-to-end performance issues. The authors of [12,13] propose to detect path outage among hosts using ping and localizes the observed path outage with traceroute. PlanetSeer [14] relies on active probes to diagnose the root cause of Internet path failures that are detected by passive monitoring the end users of a CDN service deployed on PlanetLab. Commercial systems such as Keynote [15] and Gomez [16] are also available to detect issues from the end users’ perspective by active probing. All these work employ active probing while ‘Argus’ purely depends on passive monitoring. Although the DiffServ architecture [17] supports end-to-end QoS guarantee, only a limited number of static QoS classes are provided. Obviously, it cannot meet the active fine-grained QoS demands of diverse services. Moreover, end-to-end QoS guarantee by DiffServ may require underlying hardware to support QoS guarantee, which will introduce an overhead to those core devices and lead to low efficiency. In addition, fine-grained QoS quantitative guarantee is difficult to be realized in this architecture because these fine-grained service anomalies are hard to be traced among ISPs due to commercial security problem. Besides, current research [10,18] on QoS diagnosis or QoS guarantee focused mainly on one specific criteria, such as packet loss rate, packet delay, and delay jitter, on one application. In fact, other factors, i.e. contexts in this study from different resources, should be considered in a comprehensive way because of their joint effects on a QoS metric.