Implement Multi-layer exchange (BCMSN notes)

1. Understand basic components and basic principles of key components required by layer-3 Switching
2. Compare two-layer switching and multi-layer switching, and explain the flow chart of data packets in each switching mode.
3. Understand the structure of multi-layer exchange data tables
4. Understand centralized forwarding and distributed forwarding, command-based switching, and topology-based switching.

Layer-3 switching: it combines the layer-2 switching function with the layer-3 routing function, so that the throughput of layer-3 packet forwarding can reach several million packets per second. It has the routing processing function and packet forwarding function. Packet forwarding is implemented by hardware. The performance of Layer 3 and Layer 4 is similar to that of Layer 2 forwarding.
The forwarding process of packets passing through layer-2 Packet Exchange: receive data packets on the port --- process the input port ---- check and filter the ACL of the entry --- Perform a second-layer search ---- find the corresponding exit or exit VLAN----VLAN ACL filter ---- frequent QOS output limits ---- detects the physical link of the output interface ---- queues the physical link ---- forwards the frame.

Layer-3 forwarding process: receive data packets on the port --- process the input port ---- check and filter the ACL of the entry --- Perform Layer 2 lookup ---- router forwarding module ---- check the Input Access Control List ---- check the IP address capable of forwarding ---- search IP Forwarding route table ---- output Access Controller security check (ACL) --- find the corresponding exit or exit VLAN----VLAN ACL filtering-Limits of frequent QOS output-physical links of the detection output interface-queuing the physical link-forward the frame out.

Layer-3 forwarding in the input and output of the IP packet changes: the destination MAC address to the next hop address, the original MAC address to the MAC address of the router, TTL = TTL-1, validation and re-calculation.

When a vswitch performs a search, for example, it searches for the target MAC address based on binary matching. The vswitch regards the target MAC as a search keyword and runs the HA algorithm, then we get a pointer pointing to the content address memory (memory address memory) in the switch, where it is the matched value.

The vswitch also has a triple content address table, which can match 0, 1, or none. That is, not all matching addresses. Centralized forwarding: Only one Forwarding Table (such as the 6500 series and 4000 Series) is available. All data is forwarded through the central forwarding engine, which contains a forwarding table.

Distributed forwarding: You can make local forwarding query and forwarding decisions on each module and port (for example, 3550 and 6500 with distributed forwarding cards ).
Netflow-Based Switching (Stream-Based Switching) means that one route is exchanged multiple times. It works together with a route processor and a dedicated integrated circuit through the first layer-3 forwarding of the switch (the router retrieves the forwarding information table, obtains the output interface, and modifies the data packet, after completion, rewrite and forward information to the hardware forwarding table, and then forward the data packets ). When the second packet needs to be forwarded in layer-3 mode, it will directly look for the hardware forwarding table without using the routing processor.
The traditional three-tier packet forwarding mode of routing is called processor switching (only thousands of packets can be forwarded in one second ), stream-based data exchange is based on hardware (Millions of packets can be forwarded in one second ).
Topology-Based Switching Mode (CISCO exclusive, Cef): Creates a forwarding table and a neighbor table in hardware (dedicated Integrated Circuit) before data packet forwarding, all data packet forwarding uses hardware instead of routing processor.
Configure multi-layer switching:
1. Understand the features and working principles of CEF-Based Multi-layer Switching
2. Configure and verify the CEF
3. diagnose CEF faults
Cef (ciscoexpressforward) includes two plane function modules: control plane and data plane. The control plane (with the vro as the basic structure) uses the responsible software to create a forwarding information table and a neighbor table, and stores the information on the layer-3 forwarding module and interface. The data plane is responsible for hardware-forwarded ip Unicast. When an ip packet to be routed arrives at the interface, the data plane of the layer-3 forwarding module is responsible for searching the forwarding information table, finding the next hop mac and rewriting the data frame.
Two tables are used for CEF search: the forwarding table is calculated by the route table of the router, and the neighbor table is found through arp. Layer-3 data packets are first searched through the triple forwarding information table to find the best match, get the neighbor information, get the neighbor pointer, rewrite the data frame with the layer-2 information, and then forward.
Working Process of a CEF-Based Multi-layer switch: The interface receives data packets (the destination MAC address is the MAC address of the switch, and the switch knows that data packets need to be forwarded by Layer 3) ---- the switch uses the destination IP address to search for the IP information table to obtain the pointer to the neighbor table ---- find the MAC address and VLAN Number of the next hop in the neighbor table ---- rewrite the packet with the next hop MAC and the switch MAC ---- forward data from the output port.

Arp suppression: when a new destination address is requested by the customer's pc, the link path corresponding to the mac does not exist in the forwarding information table of the layer-3 Switch. Before the switch obtains the correct link path, all data sent to this ip address will be discarded.

Multi-layer switching load balancing: when there are multiple reachable paths to a destination address, you can create multiple neighbor paths in the neighbor table, up to 6. The allocation rule can be based on the source and destination IP addresses. Source, destination address pair.
On switches that support cef, the cef function is enabled by default. The noiproute-cachecef command can disable the cef function.

Showinte *** ceinte *** ce-num | beginL3 package count
Showinte *** ceinte *** ce-num | includeswitched
Showipcef view the running status of the cef
Showadjacency view switch neighbors
Debugipcef diagnoses cef Information
Debugipcefipc

Routing Between VLANs:
1. master various interface types used by multi-layer switches for VLAN Routing
2. Explain the working principle and configuration of routes between VLANs
3. Understand the working principle and configuration of the end router.
4. Check the configurations of routes between VLANs.

Common vswitch ports:
1. Switch Port (L2 physical port)
(1) Access Port: generally used to connect to the terminal, which belongs to a VLAN
(2) trunk port: It belongs to multiple upper VLANs and can transmit traffic from multiple VLANs.
2. Virtual Interface (logical interface)
3. Routing port (some physical ports on the layer-3 Switch are a routable port, which does not belong to any VLAN. On the layer-3 switch, you can set the interface to a layer-3 routing port by running noswitchport)

Route Configuration between VLANs:
Iprouting activates vswitch Routing
Routerip_routing_protocol <option> enable Routing Protocol
Inte *** creates or enters a virtual port, representing an upper VLAN
Single-arm routing: 802.1Q has a nativevlan (default VLAN), which does not contain tagged. Use a sub-interface to mark the nativevlan, in this way, there is a default gateway in nativevlan (that is, the sub-interface above .)

Related Articles]

• How to use a layer-3 Switch to build an enterprise VLAN
  

• Introduction to layer-3 Switching
  

• Analysis of layer-3 switching technology and product selection