Routing and Switching

Big talk communication see now, this chapter I think the best, I understand the best. The first few chapters are relatively simple, the middle chapters a lot of unfamiliar concepts, 1:30 will not understand. I do not know whether I am old, always remember things ... See Chapter 12 after the front to see and forget, the overall understanding of the communication I can not say, in other words that do not understand ... I really don't know how to read this kind of book.

RC for IP network: hub (hub), Ethernet switch, router

Hub is the core device for shared Ethernet . Shared Ethernet each access terminal share a bus, who want to use to grab the line, the hub works is broadcast, a packet needs to be transmitted to all ports, resulting in a waste of resources and delay time, to introduce a switched Ethernet, the core device is an Ethernet switch , Multiple sets of communication can be made simultaneously. The switch holds the MAC address table for each terminal and can transmit data directly without broadcasting to all ports. Ethernet switches are divided into two-layer switches and three-layer switches, and the two-layer switch works on the data-link layer (Ethernet layer), and the three-layer switch can work on the data link layer and the IP layer. Layer Two switches are not responsible for any routing functions, each terminal connected to it in the same IP segment, three layer switch has routing capabilities, depending on the settings, the terminal connected with it may be in the same IP address segment, or may not be.

The router works in the IP network layer, is based on the IP packet header or MPLS tag routing forwarding, the performance indicator is the number of packets per second (PPS, the maximum number of packages per second), its work is stored and forwarded, the internal can be divided into the control plane part and the data channel part. The core of the control plane is the routing table and routing protocol, and the route forwarding packet is based on the routing table. The routing table has the destination network segment and the subnet mask, the outbound port of the router to the destination address, the next hop router entry IP address and other information, the router can select the appropriate port for the packet according to the routing table, if not found, then discard the packet. Now the router is forwarded according to the forwarding of data, and the transfer is generated by the routing table, its table items and routing table items have a direct correspondence, but the format is different, more suitable for fast lookup.

route Table fetching method: 1. Manually specify a route , that is, a static route, that explicitly knows the exact direction of an IP address segment, or that there is a uniform default egress provided by all packets that do not have a clearly identified direction in the routing table. 2. Dynamic routing, using dynamic routing protocol to obtain routing information. The routing table acquisition method designed by IP experts is a hybrid approach, which is obtained through a manual set-up part and a routing protocol. The routing protocol is a standardized protocol designed to meet the needs of the router to get the routing table. Through a series of routing protocols, all routers of IP network can get the whole network quickly and accurately by information to guide the direction of IP packets. The routing protocol is only responsible for obtaining the routing table, and the whereabouts of the IP packet into the router is the function of the router's routing query and data forwarding function.

routing protocols can be divided into internal routing protocols and external routing protocols, and BGP (Border Gateway Protocol) is the only external routing protocol.

1.RIP2 and RIPng: Distance vector protocol, which sends so-called "distance vectors" information to neighboring routers every 30 seconds, and the routing table is stored only to the destination

The next hop address of the best path to the site, the maximum number of hops allowed is 15 hops, which is considered unreachable.

2.OSPF: Open Shortest Path First, typical internal gateway protocol, famous algorithm, Dijkstra algorithm to find the shortest path (data structure of the classic algorithm, so far remember the pain of writing code), run the steps:

When the ① router initializes or changes the network structure, it generates a link-state broadcast packet that contains information about all the connected links on the router (the port information), and the router begins to work by first identifying the network segment of its own port and depositing the information into the LSA packet.

② all routers Exchange link state data by refreshing the method. A refresh is a router that transmits its LSA packets to all of its neighboring OSPF routers, and the neighboring router updates its database based on the link state information it receives, and forwards the link state information to its neighboring routers until the network is stable.

③ after a vigorous LSA transfer process, each router calculates the shortest path to all network segments based on the SPF algorithm and writes a list of routing entries. The routing table contains the cost of the router to each reachable destination and the next router to be forwarded to that destination.

This chapter is about how routers work, and the examples used in the book are easy to understand.

Routing and Switching