Computer network Knowledge arrangement--Foundation

Computer network Knowledge arrangement--Foundation

computer network Top-down method notes

End systems are connected to each other through communication links and packet switches .
The sending-side system segments the data, each with the first byte, and the resulting packets become grouped .

The two most famous types in a packet switch are:
Routers
Link Layer Switches

-End systems via the Internet service provider (provider) access network. Just like in China Telecom Unicom and so on. Access Network

The physical link of the router on which the end system is connected to its edge . The edge router is the first router on the end system to any other remote end system path.
Network access can be broadly divided into three categories: residential access, corporate access, wireless access.
1. Residential Access

Refers to connecting a home-end system to an edge router. One form is connected to a residential ISP with a dial-up modem via a normal analog telephone line.
There are two common types of broadband residential access: Digital Subscriber line DSL and hybrid fiber copper shaft cable (hybrid fiber-coaxial cable HFC).

DSL

By limiting the distance between the user and the ISP modem, the DSL can transmit and receive data at a much higher rate, its data transfer rate is usually asymmetric in two directions, and the ISP is routed to the home at a higher rate than the home-to-ISP router. Because home users are more likely to be consumers of information.
The actual download and upload rate available to the user is a function of the distance between the home modem and the ISP modem, the twisted pair specification, the electromagnetic interference strength, and other factors that need to be considered.

HFC
Like a DSL, HFC requires a special modem, called a cable modem. The HFC network is divided into two channels, namely downlink channel and Uplink channel. Downlink channels are often allocated with greater bandwidth.
It is an important feature of shared media, if several users simultaneously download different MP3 on the downlink channel, the actual rate of each user accepting MP3 will be significantly lower than the downlink rate, on the other hand, if there are very few active users in the Web surfing, Each of these users can accept Web pages at the full rate of the downstream rate, because users rarely request pages at exactly the same time.

DSL advocates sharply that DSL establishes a point-to-point connection between the home and the ISP, so all DSLs are private rather than shared. Proponents of HFC say that a reasonable range of HFC networks provides greater bandwidth than DSL.

2. Company Access

In corporate and university campuses, local area networks (LANs) are typically used to connect end users with edge routers. Ethernet technology is the most popular access technology in the current company network. Connect the end system batches using a copper-or twisted-pair wire. Similar to HFC, Ethernet uses shared media. Therefore, the end system shares the transfer rate of the LAN. Shared Ethernet technology has migrated towards switched Ethernet technology. Switched Ethernet uses a star topology where the host is connected directly to a "switch" to allow all hosts to send and receive at the same time as the full rate of the LAN.

3. Wireless access
Two major wireless access methods. In a wireless local area network (Wireless LAN) , wireless users transmit/receive services between the base stations located within a dozens of-meter radius.
In the wide area Wireless access network (Wide-area wierless access Networks) , the packet is sent by the same infinite pole facility used for cellular phones, and the base station is managed by the telecommunications provider, providing wireless access to tens of thousands of radius users.
Many families today combine broadband homes with inexpensive wireless LAN technology to produce a powerful home network. Network Core

There are two basic methods of moving data over network links and switches: circuit switching and packet switching .
In a circuit-switched network, the resources for communication between the end systems are reserved during the communication session, along the end-system communication path.
In a packet-switched network, these resources are not reserved and session messages are used on demand, and Zhejiang can only have to wait for access to the communication link.
A simple analogy is that two restaurants, one needs to renew, and the other does not need to be scheduled but not guaranteed. For the restaurant that needs to be booked, we have to take the trouble of having to call the reservation before we leave home. But when we arrived at the restaurant, in principle we were able to contact the waiter and order immediately. For the restaurant that does not need to be booked, we do not have the trouble of booking the restaurant, but may have to wait until there is a free table to get the waiter order.

In a circuit-switched network, when two hosts are communicating, the network creates a dedicated end-to-end connection between the two hosts. Assuming that each link has N circuits, each link is routed to an end-to-end connection that is used during the connection to obtain the 1/n of the link bandwidth.

1. Circuit switching

Circuits in the link are either implemented by Frequency Division multiplexing (frequency-division MULTIPLEXING,FDM) or by Time Division Multiplexing (Time-division MULTIPLEXING,TDM) implementation.

For FDM, the spectrum of the link is shared by connections created across the link. A frequency band is dedicated for each connection during the connection of the link. In the telephone network, this frequency band usually has 4KHz. The width of the band is called bandwidth (bandwidth).

For TDM links, the time is divided into frames of fixed intervals, and each frame is divided into a fixed number of slots. When a network creates a connection across a link, the network specifies a time slot for the connection in each frame. This time slot is used exclusively by this article alone.

Proponents of packet switching argue that the circuit horcrux less efficient because the dedicated circuit is idle during the quiet period. For example, a person on the phone stops talking, (that is, the band or time slot along the link in the linked path), the idle network resources cannot be used by other in-progress connections.

2. Packet switching

Various applications need to exchange messages (message)when completing tasks. Messages can contain anything the protocol designer needs.
In modern computer network, the source host divides the long message into smaller chunks and is called the Grouping (packet). Between the source and destination, each of these groupings passes through the packet switch (packet switch)of the communication link.

Most packet switches use the store-and- forward transport mechanism at the input of the link. The store-and-forward mechanism means that the entire grouping must be received before the interchange can begin to transmit the first bit of the packet to the output link. Therefore, the storage-forward packet switch introduces the store-and- forward delay at the input of each link along the path of the packet.

In addition to the store-and-forward latency, the packet also bears the queued latency of the output cache. These delays are variable and vary depending on the congestion level of the network. Because the cache space is limited in size, an arriving grouping may find that the cache is fully filled with packets waiting to be transferred. In this situation, the packet loss or packet drop occurs.

3. Packet switching and circuit switching

Proponents of circuit switching often argue that packet switching is not suitable for implementing services because of its variable and unpredictable end-to-end latency. Proponents of packet switching argue that it provides better broadband sharing than circuit switching, which is simpler, more efficient, and less expensive to implement than circuit switching.

Packet switching uses on-demand links, and the link transfer capability is shared only among all the users whose groupings are to be transferred on the link. Such on-demand shared resources are sometimes referred to as statistical multiplexing of resources.

The types of delay in packet switching networks include: processing delay, queue delay, transmission delay, propagation delay.

1. Processing delay : The time required to check the header and determine where to orient the group is part of the processing delay, as well as other factors, such as the time required to check for a higher-than-premium error.

2. Queueing delay : In a queue, it undergoes a queuing delay when packets are waiting for transmission on the link.

3. Transmission delay : Is the time required to push all the grouped bits to the link.

4. propagation delay : Once a bit is pushed to the link and the pen needs to propagate to the target route, the time it takes for the link's starting point to propagate to the router is the propagation delay. using a packet sniffer