Explain the third layer exchange technology in detail
Abstract Analysis of the basic characteristics of the third floor exchange technology, and in-depth analysis of the classification, working principle and implementation methods of the third layer exchange technology, comparing the forwarding characteristics of two types of third-level exchange technology packets. This view is questioned on the current LAN Internet "Routing, Switching" or "Layer 2 Exchange Based on Destination MAC Address". Key words third floors exchange technology Traditional router three-layer switch packets to packet exchange exchange exchange with the current network business traffic growth, and business flow mode changes to more business streams across the subnet boundary, The traffic flow through the router has also increased, and the traditional router low speed, complex network bottlenecks are convex. The appearance of the third layer exchange technology is well solved the problem of low-transgeneration rate, high delay, and other network bottlenecks caused by traffic flow across network segments in the LAN. The application field of the third layer exchange device is also from the initial backbone, and the aggregation layer has always penetrated into the edge of the access layer. The application of the third-level switching equipment is increasing in the network interconnection. However, there is a big difference in the understanding of the third layer exchange technology and its equipment. Some professional magazines and scientific abstracts are easily misleading for the incorrect introduction of the third floor exchange technology principle. The third layer exchange is a new generation of local area networks and exchange technologies, and its products are in architectural, and they are implemented with different Layer Ethernet switches and traditional routers. This paper analyzes the basic characteristics, classification and realization principles of the third layer exchange technology, and it is hoped that it is beneficial to people to further understand and use third-layer exchange equipment. 1 Third layer exchange technology Basic Features The third layer exchange technology is also known as IP switching technology. It combines the advantages of both the second-layer switch and the third layer router into an organic whole. It is a mechanism to enhance the second-layer exchange function using information using information in the third layer of protocol, which is a new generation of local area network routing and exchange. technology. The third layer exchange technology has the ability to obtain transmission performance in the past 10 times in the current system 1/10. Since the third-layer switch can replace most of the router to perform the most features of the traditional router, it should have the basic feature of the route. We know that the core functions of routes mainly include data packet forwarding and routing processing. Data message forwarding is the most basic feature of the router and the third layer switch, used to transfer data packets between subnets; routing sub-features include creation and maintenance of routing tables, completing this feature to enable routing protocol such as RIP or OSPF To discover and establish a network topology view to form a routing table. Once the routing process is completed, the routing process is sent to the destination is the task of the packet forwarding sub-function. The work of packet forwarding sub-function includes checking the IP packet head, IP packet fragmentation and reorganization, modifying the survival time (TTL) parameter, recalculates IP header checks, MAC address resolution, IP package data link package And the error and control processing (ICMP) of the IP package, etc. The third floor exchange also includes a series of special service functions, such as the format conversion of packets, information flow priority, user authentication, and packet filtering and other security services, IP address management, LAN protocols, and broad area network protocols. When the third layer switch is used for forwarding the subnet or VLAN during the local area network or the VLAN does not perform routing processing, only the third layer service flow forwarding, in which case the device may not need routing. Since the traditional router is a software-driven device, all packet exchange, routing and special service functions, including processing multiple underlying technologies and a variety of third layer agreements are almost all by software, and can be upgraded by software. Function, therefore has good scalability and flexibility. But it also has the disadvantages of complex, high prices, relatively low throughput and relatively high throughput changes. The third floor exchange technology has largely compensated for the shortcomings of traditional routers. The following methods are usually used when designing the third-level exchange of products: • Cutting the number of protocols, often only only for IP; · Only exchange and routing functions, limit special services; • Use a dedicated integrated circuit (ASIC) construct more features Instead of running these functions in software above the RSIC processor. The third floor exchange product uses structured, modular design methods, and the architecture has a good level.
Software modules and hardware modules are clear, with coordination, information can be saved, fully distributed, or caches throughout the device. For example, the location of the third layer of IP packets is determined in the position of the frame, and the address bits can be extracted by hardware, and the routing calculation or address lookup is completed by hardware; on the other hand, routing table construction and maintenance can continue The software is completed by the software in the RSIC chip. In summary, the implementation of the third layer exchange technology and product is attributed to the rapid development of modern chip technology, especially ASIC technology. 2 Third layer exchange technology classification, principles and implementations are currently mainly two types of third floors exchange technology: the first category is a message to the message exchange, each message must go through the third layer processing (ie at least the route Processing), and data flow forwarding is based on the third layer address; the second class is a stream exchange, which does not process all packets in the third layer, and only the first packet in the stream, complete the routing processing, and based on The third layer address forwards the message, subsequent packets in the stream use one or more shortcut technologies, and the design of such technologies is to facilitate line speed routes. Understanding the key to the third layer of exchange first needs to distinguish the different forwarding methods of these two types of packets. A prominent feature of the message to the message processing method is that it can adapt to the topology of the route. By running standard protocols and maintains the routing table, the packets to the packet switching device can dynamically reroute packets, bypassing network fault points and congestion points without waiting for the high-level protocol test packet loss. The stream exchange method does not have these features, because subsequent packets take shortcut without the need for a third layer, so that it cannot identify changes in the routing table for the standard protocol. Therefore, the flow exchange method may require additional protocol to obtain topological changes or congestion information in order to reach the correct place in the exchange system. 2.1 Message to Packet Switching Technology Principle and Implementation Method Packets to Packet Exchange Following such a data stream process: packets enter the first layer of the OSI reference model in the system, that is, the physical interface, and then accept the destination Mac in the second layer. Check that if the second layer can be exchanged, two-layer exchange is performed, otherwise it will enter the third layer, that is, the network layer. At the third layer, the packet should be used to determine, address parsing and certain special services. After the processing is completed, the message has been updated to determine the appropriate output port, the packet is transmitted to the physical media through the first layer. Traditional router is a typical device that meets the third layer of messages to packet exchange technology, which is completely software-based working mechanisms that have been overcome by modern hardware-based third-floors. At present, the third-layer switching equipment provided by each manufacturer has almost the same hardware structure on the architecture. The central silicon exchange array connects the CPU module through the CPU interface bus. Connect the I / O interface module through the I / O interface bus, which is a concentration point of the traffic gathering and exchange of the device, which provides a parallel switching path of each of the access ports of the device. The data stream across the I / O interface module is forwarded through a silicon exchange array. Each I / O interface module contains one or more forwarding engines, and the ASIC is completed all packets *, including routing, packet category, third layer forwarding and business flow decision, this will forward The method of distributing the ASIC of each I / O port is a key part of the third layer switching device to be linear. The CPU module mainly completes the background of the device, such as various route protocols related to routing processing, create and maintain routing tables, system configuration, etc., and import routing table information into each I / O interface module distributed forwarding engine. ASIC. Thus, the distributed forwarding engine ASIC of each interface module directly proceeds to the forwarding strategy of the packet according to the routing table, and there is no need to pass all the packets like the traditional router. 2.2 Flow Exchange Technology Principle and Implementation Method In stream exchange, the first message is analyzed to determine if it identifies a "stream" or a set of messages with the same source address or destination address. The flow exchange saves the processing time to check each packet. Subsequent packets in the same stream are exchanged to the destination address based on the second layer.
The flow exchange requires two techniques. The first tip is to identify which feature of the first message identifies a stream, which allows the remaining packets to take shortcut, ie the second layer path. The second tip is that once the path through the network is established, it is long enough to use the advantages of shortcuts. How to detect streams, identify packets belonging to specific streams and the establishment of changes in circulation roads through the network. There are currently a variety of stream exchange technologies, such as the 3CM's fast IP, submitted by Cisco to the IETF Multi-Protocol Mark Switch (MPLS), the ATM Forum Multi-Protocol (MPOA), and IPSilon's IP Exchange. We can divide them into two main types: end system drive flow exchange and network center flow exchange. Limited to the space, now only briefly introduce the fast IP work of 3COM. 3Com's fast IP belongs to the end system drive flow exchange technology, its working principle is based on NHRP standard (draft). The source host sends a quick IP connection request, which is routed through the network like a data packet. If the destination host runs fast IP, it sends a NHRP answering message containing its MAC address to the source host If the source host and destination host exists, there is a two-layer exchange path. When the NHRP response message arrives at the source host, the target terminal host MAC address and port mapping table will be generated, and the source host can be based on the purpose. The terminal host MAC address switches data packets directly through the switch, no longer passing the router; if there is no switch between the two ends, there is no exchange path without the NHRP response, the packet is routing as before. The fast IP software is mainly running on the source, de intending host's network interface card (NIC) driver. It is coordinated with NHRP exchange with the host's IP protocol stack and NIC driver. In summary, the fast IP is trying to improve the forwarding performance of the route on the switched network, but it does not have potential flexibility, and no security is provided through packet filtering, and the NHRP protocol software is installed on the host participating in the fast IP exchange. In fact, the maintenance workload of the equipment is added. Multi-stream exchange techniques were initially developed under the premise of routing. The message to the message exchange product has proven to be the case. The flow exchange method is more complex and difficult to understand compared to the message exchanged product exchanged products. In dynamic network environments, successfully identify, establish, manage, and undo a lot of flows, which measures needed to be a problem to be studied. At present, the third-level switching device currently applied to the LAN is mostly based on packets to packet switching technology. Flow exchange is more likely to find its location in a wide area network. 3 On the 3rd floor switch "Routing, Switching" or "Device MAC Address" This view is questioned. Many articles have been introduced to the working principle of the three-layer switch, three floors in switches. The packet forwarding mechanism is basically consistent, that is, "Routing, switch" or "a second floor exchange based on the purpose of the mac address". These texts are checked? The principle of the third layer exchange is the principle of the third layer exchange, assuming two hosts A, B using IP are communicated. Send a site A When starting to send, compare your IP address with the IP address of the B station, determine if the B station is in the same subnet. If the destination station B is in the same subnet, the second floor is performed. Forward. If the two sites are not in the same subnet, if the send station A is to communicate with the destination station B, the send station A is to issue an ARP request (address analysis) package to "Default Gateway", and "Default Gateway" IP The address is actually a three-layer switching module of a three-layer switch. When the transmission station A broadcasts an ARP request to the IP address of the "Default Gateway", if the three-layer switching module is known in the previous communication process already knows the MAC address of the B station Then reply B's MAC address to the send station A.
