Switch Development (iii)--deep analysis of the principle and design of three-layer network switch

Introduction

Traditional routers have isolated networks, isolated broadcasts, routing and forwarding, and firewall operations in the network, and as the network continues to evolve, the load of routers is growing rapidly. One of the important reasons is that the VLAN (virtual local area network) technology is widely used in networks because of security and management convenience. VLAN technology can logically isolate different network segments, ports, or even hosts, and the communication between the different VLANs is routed through the router to complete the forwarding. Because of the large data traffic in the local area network, the large amount of information exchange between the VLANs is forwarded through the router, and the router becomes the bottleneck of the network with the increasing of data traffic. In order to solve this bottleneck of local area network, many enterprises in the internal, school and community building LAN used three-layer switch. The three layer switching technology introduces the switching technology to the network layer, and the application of layer three switches from the backbone layer and the aggregation layer of the initial Network center to the access layer of the network edge.

first, the third layer of exchange technology

1, three-layer switching concept

The third layer of switching technology, also known as IP switching technology or high-speed routing technology, is relative to the traditional concept of exchange proposed. As we all know, the traditional exchange technology is in the OSI Network Standard Model of the second layer-the data link layer to operate, and the third layer of exchange technology is in the network model of the third layer to achieve high-speed packet forwarding. Simply put, the third layer of exchange technology is: the second layer of switching technology + third-tier forwarding technology, which is a third layer of information in the Protocol to strengthen the second layer of exchange function mechanism. A device with a third-layer switching capability is a second-tier switch with a third-tier routing capability, but it is an organic combination of the two, not simply stacking the hardware and software of a router device on a LAN switch.

2, three-layer exchange principle

From the hardware implementation, at present, the interface module of the second layer switch is through the high-speed Backplane/bus exchange data. In the third layer of switches, router-related third-tier routing hardware modules are also plugged into the high-speed backplane/bus, which allows the routing module to exchange data quickly between other modules that need to be routed, thus breaking the traditional interface speed limit for external routers (10mbit/s---100mbit/s )。 On the software side, the third layer switches the traditional software-based routers back to the definition

(1) Data packet forwarding: such as IP/IPX packet forwarding, these regular processes through the hardware high-speed implementation;

(2) The third layer of routing software: such as the update of routing information, routing table maintenance, routing calculation, routing determination and other functions, with optimized, efficient software implementation.

Assuming that there are two sites that use the IP protocol, the process of communicating through a third layer of switches is: If send site A at the beginning of the sending, known as the destination station B IP address, but do not know that it sends the required MAC address on the LAN, you need to use address resolution (ARP) to determine the MAC address of B. A compares its IP address with the IP address of B, and uses the subnet mask configured in its software to extract the network address to determine whether B is within the same subnet as itself. If B and a are in the same subnet, a broadcasts an ARP request, B returns its MAC address, a gets the MAC address of B, caches the address, and forwards the data with this MAC address packet, and the second-level switch module looks for the MAC address table to determine the packet to the destination port. If two sites are not within the same subnet, a must issue an ARP (address resolution) packet to the default gateway, and the IP address of the "default gateway" is already set in the system software, which actually corresponds to the third layer switch module of the third layer. When a to the "Default gateway" IP address broadcast an ARP request, if the third switch module in the previous communication process has obtained B's MAC address, then to send station a reply to B's MAC address; otherwise the third Layer switch module broadcasts an ARP request to the destination station based on the routing information, and B gets the ARP request to The third layer switch module replies to its MAC address, and the third switch module saves this address and replies to the Send station a. Later, when the packet between A and B is forwarded, the MAC address of the final destination site will be sealed, the data forwarding process is all handed to the second layer of exchange processing, the information can be exchanged at high speed [1].

3. Characteristics of the third layer Exchange

The salient features are as follows:

(1). Organic hardware combination makes data exchange faster

(2). Optimized routing software makes routing process more efficient;

(3). In addition to the necessary routing decision process, most of the data forwarding process is processed by the second layer;

(4). Multiple subnets interconnect only the logical connection with the third Layer switch module, not as the traditional external routers need to increase the port, to protect the user's investment.

The goal of the third layer exchange is that there is no need to go through the router to forward packets as long as there is a more direct second-tier path between the source and destination addresses. The third layer of exchange uses a third-tier routing protocol to determine the delivery path, which can be used only once or stored for later use. The packet is then quickly routed through a virtual circuit bypassing the router.

The emergence of the third layer switching technology solves the network segment in the LAN, the network segment must rely on the router to manage the situation, to solve the traditional router low speed, the complexity caused by the bottleneck problem. Of course, the three-layer switching technology is not a simple overlay of network switches and routers, but a combination of the two, forming an integrated, complete solution.

Second, VLSW4150 series switch

VLSW4150 series switches are a high-performance, multi-tiered solution for your business. VLSW 4150 also applies to ISPs and service providers, especially for some large operators, and will be able to enhance their competitiveness in the IP market.

1. Overall design

The VLSW4150 switch has 24 RJ45 10/100base-tx adaptive ports with 2 optional fiber 10/100basetx Ethernet ports, 100BASEFX Fast Ethernet port, or Gigabit Ethernet (SX,LX,ZX) And also provides a serial port and a 100M Ethernet port to configure the switch.
The architecture of the VLSW4150 switch supports speeds up to 11Gbps, multilayer switching speeds of up to 6.6Mpps, support for 8,192 MAC addresses, and speed limits in 1M increments for better control of network traffic and network security Support for frame filtering of Taggedvlan and mac-based as well as RIP, OSPF, and BGP routing protocols.

The VLSW4150 switch provides a stack technology that can manage multiple switches with a single logical IP address, and can be mirrored in a single port, providing a Web-based network management system and a CLI mode to debug the switch. The VLSW4150 supports SNMP protocol, Rmon, and Telnet features for easy administration.

2. Hardware structure

The hardware structure of the VLSW4150 three-layer switch is divided into two parts, the processor module and the switching module, which are connected through the PCI interface, and together with the corresponding peripheral circuitry to form a complete three-layer switch system, see Figure 1.

Figure 1 Hardware structure

(1) Processor module

As shown in 2, the processor unit uses a MOTOROLAPOWERQUICCIICPU, with some external storage devices and some peripheral circuitry to form the processor portion of the three-layer switch. The processor module mainly runs the embedded operating system, configures the maintenance of the system and routing table, rather than the component of the data forwarding path. CPLD saves some CPU initialization configuration to ensure that the CPU starts normally after power on, Flash chip is used to store all the software and related configuration required by the three-layer switch, SDRAM will be loaded into the Flash program after the system is started, and the system can be guaranteed to run normally. On the one hand, the processor module provides a Fast Ethernet interface and an asynchronous port for configuration and commissioning of the switch, and on the other hand, through the PCI interface and the switching module, the switching module is controlled by the PCI interface and the data is transferred [2].

Figure 2 Hardware composition of the processor module

(2) switching module

3, the switching module uses the Broadcom Company's BCM5645 as the ASIC chip, through the PCI interface and the Processor module communication complete data transmission, through the memory interface provided by 5645, can provide a 64M external SDRAM to the switching module, This improves the throughput and switching speed of the switch. 5645 Connect 24 hundred Gigabit Ethernet and 2 Gigabit Ethernet via the Mii interface and the Gmii Interface (3) respectively.

Figure 3 Hardware composition of the switching module

3. Software structure

The software system of the VLSW4150 three-layer switch adopts the modular and distributed design method, which is based on the real-time multitasking operating system. The structure of software system is hierarchical structure, one layer is built on the basis of another layer, each layer uses the service provided by its next layer, and provides a higher level of service for the above layer, the advantage is that it can shield the bottom operation from the upper software, improve the portability of the upper software and improve the maintainability of the software.

As shown in 4, the software is broadly divided into three levels:

(1) Drive layer

The driver connects the upper software and the hardware system, and translates the routing update, management and configuration commands of the upper software into the format that can be recognized by the hardware system, so as to update its internal data structure such as routing table, Address table and so on, to control and manage the hardware Exchange system. At the same time, the device driver transmits the routing update messages received by the underlying hardware, the control management frames and the various information received to the upper layer software processing;

(2) protocol stack

Implementation of TCP/IP, 802.1D and 802.1Q protocols, for the upper-level applications to provide a good interface;

(3) Application layer

Mainly includes the routing module and the network management module, the routing module realizes the protocol such as RIP and OSPF, namely realizes the third layer route the main function, the network management module realizes the SNMP and the Rmon and so on the network management module, causes the three layer switch to have the partial network management function, guaranteed three layer switch to be better normal operation.

Figure 4 Software Architecture

III. application of the third layer switch

The main purpose of the third layer switch is to replace the traditional router as the core of the network, so there is no need for wide-area connectivity, but also the need for routers, can be replaced by a third layer of switches. In the enterprise network and Campus network, the third layer switch is usually used in the core layer of the network, with the gigabit port or the gigabit port on the third layer switch to connect different subnets or VLANs. The third layer switch solves the situation that the LAN VLAN must rely on the router to manage, and solves the network bottleneck caused by the traditional router's low speed and complexity. Using three-layer switch to divide VLAN in LAN can satisfy many kinds of flexible logic combination, prevent broadcast storm, and set different access rights between different VLANs, so as to increase the overall security of network and greatly improve the work efficiency of network administrator. and the third layer switch can rationally configure the information resources, reduce the network configuration cost, make the connection between the switches become flexible.

Switch Development (iii)--deep analysis of the principle and design of three-layer network switch