Bring you into the world of switched Ethernet

A long time ago, the concept of switched Ethernet has been put forward. Today, we will explain the basic knowledge about switched Ethernet, so that you can have a deeper understanding of it, switched Ethernet is initially based on a coaxial cable. in, Xeroxinter and DEC proposed DIX.

◆ In, the standard was proposed.

◆ CSMA/CD communication process, transmission-listener-interference-random wait-transmission.

◆ Traditional switched Ethernet uses bridges to separate hosts and routers to connect network segments.

◆ Switched Ethernet, usually the host is not connected. When communication is required, the peer host is connected through the switch device and then disconnected. Exchange devices include exchange hubs and switches.

◆ The physical logic of switched Ethernet is star. Splits the conflict domain and limits the network conflict to the minimum range.

◆ There are nine groups of RMON, including port statistics, history, alarms, and events.

◆ Data traffic is differentiated by organization and administration, by host type, by physical distribution, and by application type.

◆ 80/20 rule, 80% local, 20% other network segments. 20/80 rules, on the contrary.

◆ A single-port 64-byte packet forwarding of the switch is 1488810pps, And the 64-byte packet forwarding of the router is smaller than 100pps.

◆ Layer-3 switching technology implements hardware route forwarding. The forwarding route table is also established by the software through the routing protocol.

◆ Layer-3 switching and routing are used to determine the path based on the Logical Address, run layer-3 checksum, use TTL, process and handle information, analyze packets, and update SNMP management with MIB.

◆ Three-layer switching advantages: hardware packet forwarding, low latency, and low cost.

◆ Layer-4 switching is based on data streams to achieve one route and multiple exchanges. Consider the port number and protocol fields.

◆ LAN design principles: Investigate physical links, analyze data stream features, adopt hierarchical models, and consider Redundancy

◆ LAN Management System functions: configuration, monitoring, and fault isolation.

◆ Network performance, bandwidth, and latency must be guaranteed. It depends on an important factor, the cable type and layout.

◆ Increase the bandwidth, increase the overall bandwidth, and reduce the number of users on a shared media.

◆ The standard for fast switched Ethernet (100 M) is 802.3u.

◆ Dedicated data packets are not required for self-negotiation using physical chips. Send 16 Bi packets. The entire packet is sent at 16 Ms intervals.

◆ The speed cannot be completed through self-negotiation, but the working method may cause problems. One segment is forced to be 10 m full duplex, and the other end is automatically negotiated to be 10 M half duplex.

◆ Self-negotiation priority: 100BASE-TX full duplex, 100BASE-T4, 100BASE-TX, 10BASE-T full duplex, 10BASE-T

◆ Self-negotiation of gigabit switched Ethernet has been implemented, but the self-negotiation of switched Ethernet on the optical fiber fails.

◆ The vswitch belongs to the MDIX device and the PC is an MDI device. Physical chip implementation.

◆ Half duplex adopts backpressure technology for traffic control.

◆ Full-duplex throttling follows the 802.3x standard and uses 64-Byte "PAUSE" MAC frames. This frame uses the multicast address 01-80-c2-00-00-01.

◆ PAUSE is used between the terminal and the switch, and cannot solve the congestion of stable state. End-to-End traffic control and more complex operations than simple stop-to-start operations.

◆ Port aggregation is only applicable to the MAC Mechanism of the 802.3 protocol family.

◆ Flow control Command flow-control

◆ Configure the port aggregation link-aggregation port_num1 to port_num2

◆ 3526 supports three switched Ethernet groups and one optical group, each with eight optical groups. E0/1, e0/9, e0/17, G1/1

◆ Vlan Division: Based on ports, mac, protocol, and Subnet

◆ 802.1Q standard for virtual bridging LAN (VLAN.

◆ 802.1Q defines the vlan architecture (MAC frame format), services provided, protocols and algorithms involved in implementation.

◆ The dot1q tag header contains two-byte tag protocol identifier (TPID) and two-byte tag control information (TCL ). The fixed value of TPID is 0x8100. TCL includes priority, CFI, and vlan id.

◆ All frames with the dot1q label header are tagged frames.

◆ GARP general property registration protocol, which is applied to GVRP and GMRP

◆ There are five GARP messages: join in, leave, empty, join empty, and leave all.

◆ GVRP is a dynamic VLAN Registration Protocol and is enabled as gvrp.

◆ GVRP is divided into three categories: normal allows dynamic vlan creation, registration, and logout. Fixed allows you to manually create and register VLANs to prevent vlan cancellation and other interfaces from registering VLANs known to this interface. Forbidden cancels all VLANs except vlan1, and prohibits the creation and registration of other VLANs on interfaces.

◆ PVLAN configuration, isolate-user-vlan enable. After establishing a ing relationship, you must first remove the original ing relationship to the interface.

◆ Trunk only allows default vlan without tags, and hybird allows multiple VLANs to pass through without tags.

◆ The switching Ethernet frame is fixed with a fixed length. The layer-3 Switch uses a method that matches the longest address mask of the router-precise address matching for packet processing.

◆ Stream-based exchange, the first packet goes through three layers of processing, and the others are forwarded twice. For packet switching, each packet must undergo a layer-3 check.

◆ 802.1D Spanning Tree Protocol: transmits a special configuration information BPDU between bridges. Function: select the root bridge, calculate the shortest path, select the specified bridge, select a port, and select the port contained in the build tree.

◆ BPDU includes the root bridge ID, minimum path overhead, specified bridge ID, and specified port ID.

◆ The bridge ID is expressed by the combination of the bridge priority and mac address.

◆ BPDU uses the fixed mac address 01-80-c2-00-00-00 as the destination address. The SAP value is 0x42.

◆ The root bridge is the one with the smallest bridge ID.

◆ BPDU priority Comparison Principle: The BPDU priority is the highest priority when the BPDU priority is the lowest priority.

◆ Topology change messages include topology change notification messages, topology change response messages, and topology change messages.

◆ The scheduled sending cycle of STP is hello time, the message life cycle is configured as message age, and the maximum life cycle is max age.

◆ Method to avoid temporary loop: Set the intermediate state. The blocked State enters the intermediate state after a forward delay, and the intermediate state passes through a forward delay to enter the forwarding state.

◆ Several statuses of the stp port: disabled does not send or receive any packets. Blocking does not receive and send data, but does not send bpdu, and does not perform address learning. Listening does not receive or send data. It accepts and sends bpdu without address learning. Learning does not receive or forward data. It accepts and sends bpdu to start address Learning. Forwarding accepts and forwards data, accepts and sends bpdu, and learns the address.

◆ The aging time value of the mac address table is greater than the time required for re-calculation of the spanning tree. A longer value is used for a disk for 15 minutes. After re-calculation of the spanning tree, a shorter buffer timeout value is used.

◆ Quick Tree Generation Improvement: 1. if the old root port is blocked and the specified port of the New Root port connection CIDR block is in the forwarding state, the new root port can be forwarded without delay. 2. The specified non-edge interface waiting for forwarding sends a handshake packet to the downstream. If the downstream responds with approval, the interface will not be delayed for forwarding. The handshake must be in the point-to-point link and will be passed down until the network edge. 3. edge interfaces are forwarded without delay.