China
Africa
Explore chapters and articles related to this topic
The Era of High-Performance Networks
Published in James Aweya, Designing Switch/Routers, 2023
Routers are the main network devices that provide the network-wide intelligence for moving information in internetworks, from enterprise networks to service provider networks, and the Internet as a whole. IP routing is a general term used to represent the collection of methods and protocols used to determine the paths across multiple internetworks that a packet can take in order to get from its source to its destination. Each packet is routed hop-by-hop through a series of routers, and across multiple networks from its source to the destination. Each hop represents a routing device (or router).
High-Performance Switch-Routers
Published in James Aweya, Designing Switch/Routers, 2023
Redundant, hot-swappable components provide non-stop service delivery as explained by the following examples:Route Processor Module: The route processor (or control engine) is responsible for running the control plane functions in the system which comprises the tasks performed by the routing protocols (e.g., RIP, OSPF, IS-IS, BGP) and the management protocols (e.g., SNMP, ICMP, IGMP), and the tools used for system configuration and troubleshooting. The IP routing protocols are used for advertising network topologies, exchanging routing information, and computing routes to network destinations (intra-network and inter-network routes). A switch/router may be configured with dual CPU/route processor modules supporting sub-second detection and failover (Figure 1.5).Switch Fabric Element Redundancy: The switch fabric is responsible for transferring packets between the line cards and between the line cards and the route processor. A switch/router may be configured with one or more redundant switch fabric modules supporting millisecond or lower failover performance (Figure 1.6).Hitless Route Processor Failover: Stateful failover ensures that the forwarding engines on the line modules (in a distributed forwarding architecture) are not impacted by a route processor failover (Figure 1.5). This capability enables non-stop packet forwarding in the event of a route processor module failover.Redundant Power Supplies: A switch/router may be designed to support N+M power supply module redundancy for AC and DC power configurations.Redundant Chassis Cooling Fans: Traditional computing devices and network devices use cooling fans to accomplish a number of functions: draw cooler air into the chassis from the outside, expel warmer air from within the chassis to the outside, and blow cooler air across internal processors and heat sinks, and power supplies. This is done to maintain a cooler ambient temperature and improve system performance. A switch/router may support N+M cooling fan modules to allow continuous operation irrespective of system traffic loading and outside temperature conditions.Additional design features include intake and exhaust temperature sensors, and cooling fan spin detection to aid in rapid identification of abnormal or failed operating conditions to help minimize MTTR.
A Survey on Packet Switching Networks
Published in IETE Journal of Research, 2022
The traffic demand has been increased because of delay-sensitive applications and increasing bandwidth, and the circuit-based networks can no longer fulfil traffic demands. Hence, Multiprotocol Label Switching (MPLS) has been introduced to optimize the network resources and manage the traffic [19]. MPLS is a technique used to forward the packets based on labels. The lookup process in the switching table in the MPLS network is less time-consuming and complex than the routing lookup process in IP routing. MPLS network provides many advantages and features like Quality of Service (QoS) and traffic engineering [20]. In an MPLS network, the packets forwarded are based on labels, and these labels are injected by the ingress edge router and removed by the egress edge router. Therefore, the source and destination workstations do not know about the labels. This technique is also called label switching.
An efficient and DoS-resilient name lookup for NDN interest forwarding
Published in Connection Science, 2021
Dacheng He, Dafang Zhang, Yanbiao Li, Wei Liang, Meng-Yen Hsieh
Name lookup is the core operation of NDN forwarding, just like IP lookup in TCP/IP networks. During the processing, the packets are separated into interest packets and data packets. Consumers request the contents by interest packets and obtain data packets back with target contents. In NDN, one interest packet can pull back only one data packet. Moreover, the data packet is forwarded following the reverse path from where the interest packet came from. The most time-consuming operation in NDN forwarding is to determine how to forward Interest packets. The IP lookup only needs to match one route table called Forwarding Information Base (FIB) in IP routing. However, due to its complexity, name lookup needs to match 3 NDN routing tables: Content Store (CS), Pending Interesting Table (PIT), and FIB. The data transmission in the NDN Router is depicted in Figure 2, the flowchart of name lookup processes within three tables. As can be observed, NDN's name lookup is much more complicated than IP address lookup in the TCP/IP network.
Internet of Things: A Comprehensive Review of Enabling Technologies, Architecture, and Challenges
Published in IETE Technical Review, 2018
Bhagya Nathali Silva, Murad Khan, Kijun Han
6LoWPAN is another protocol developed by IETF, which is operating in the network layer of the infrastructure. Since 6LoWPAN was built taking IPv6 as the base, it facilitates interoperability with other IP networks, as well as with other wireless devices on IEEE 802.15.4. 6LoWPAN allows each constrained device to be accessed uniquely within the network, making the administration tasks easier. Moreover, it is responsible for fragmenting and reordering of IPv6 packets, compressing protocol stack headers, enabling stateless addressing, providing a basis for “mesh-under” routing and assuring consistency with the upper layers [62]. In IP routing over 6LoWPAN, additional header information is not a mandatory field, so that it reduces unnecessary packet overhead while saving more space for data to be transferred [63]. Moreover, 6LoWPAN has a mesh address header to support routing of packets in a mesh network, but leaves the details of routing to the link layer [64]. 6LoWPAN header is identified by the type field represented in the first two bits of the header. There are four types of headers defined for 6LoWPAN communications, i.e. (1) If the packet is not for 6LoWPAN processing, the header is set to No 6LoWPAN (00); (2) If the header is set as Dispatch (01), it indicates that the packet is ready for IPv6 header compression; (3) The Mesh-Addressing (10) header-type forward IEEE 802.15.4 frames to the link layer as required, to create multi-hop networks; and (4) Fragmentation (11) header is used if the packet size exceeds IEEE 802.15.4 frame size [65].