Computer Network Review summary (c)

Computer Network Review summary (c)

Chapter data link layer

• Channel classification data link layer
  • Point channels: one using the P2P communication mode.
  • Broadcast channel: broadcast communication using one to many.
    • Many host broadcast channel connection, it is necessary to use a dedicated shared channel data transmission protocol to coordinate these hosts.
• Data Link: In addition to the physical line, there must be a communication protocol to control the transmission of such data.
  • "Digital pipe": a data link can be compared to the pipeline, and the pipeline on this digital data transmission unit is a frame.
• Three basic questions
  • Encapsulated in a frame: add header and trailer before and after each piece of data, and constitutes a frame. Determining frame boundaries.
    • An important role of the head and the tail is carried out frame delimitation.
  • Transparent transmission: Regardless of what kind of transmitted data bit combination, should be able to transmit on the link.
    • When preaching bit combinations in the data happens when one control information exactly the same, it is necessary to take appropriate measures to debit such data will not be mistaken for some kind of control information. So as to ensure transmission of the data link layer is transparent.
    • That is: the link layer so that the "invisible" something hinder data transmission.
    • Byte stuffing methods available to solve the problem transparent transmission.
  • Error detection: bit errors may occur during transmission: 1 may become 0 and 1 may become 0.
    • BER: a period of time, as a percentage of transmission error bits total number of bits transmitted.
      • Bit error rate and signal to noise ratio of a great relationship.
• A cyclic redundancy check CRC: a frame transmitted in a data link layer, widely used cyclic redundancy check CRC error detection technique.
  • 1, the transmitting side, data is divided into first group. Assuming each k bits.
  • 2, assuming that a set of data to be transmitted is M = 101001 (i.e., k = 6).
  • 3, a binary arithmetic modulo 2 ^ n 2 M multiplication, which corresponds later M adding n-0.
  • 4, the number (k + n) bits obtained by dividing the length of a good pre-selected (n + 1) bits of the divisor P, Q is obtained commercially and the remainder is R, R 1 remainder bits less than the divisor P, i.e., R is an n-bit.
  • 5, the remainder R is added after the transmitted data M as a redundant code. Data transmission is: 2 ^ n * M + R
  • 6, the receiving end of each frame received CRC tested
    • (1) If the resulting remainder R = 0, it is determined that no frame errors, accept (accept).
    • (2) If the remainder R ≠ 0, it is determined that the frame is in error, it is discarded.
  • This detection method can not determine whether which one or a few bits of error has occurred.
  • As long as undergo a rigorous selection, and use of a sufficient number of bits of the divisor P, the probability of undetected errors that occur on very, very small.
• Frame check sequence FCS: i.e., after the redundancy code added data.
  • FCS can be obtained by this method CRC, but CRC is not the only method used to obtain the FCS.
• Point to Point Protocol PPP
  • The world's most used data link layer protocol.
  • When the user access to the Internet using dial-up telephone line, PPP protocol is generally used.
  • PPP agreement should meet the needs of:
    • simple
    • Framing package
    • Transparency
    • Multiple network layer protocols
    • Various types of link
    • Error detection
    • Detecting a connection state
    • The maximum transmission unit
    • Network layer address negotiation
    • Data compression negotiation
  • PPP protocol does not require features:
    • Correction
    • flow control
    • No.
    • Multi-line
    • Simplex or half-duplex links
  • Composition of PPP
    • To a method for encapsulating IP datagram serial link.
    • Link control protocol LCP (Link Control Protocol)
      • Data Link Control
    • Network Control Protocol NCP (Network Control Protocol)
      • Network control data
  • Transparent transmission problems
    • When used in the synchronous PPP transmission link (SDH), a predetermined protocol to accomplish hardware bit padding (and practices as HDLC).
      • Zero padding bits
        • It occurred when the information field and a flags field F exactly the same 8-bit combination;
          • The sender 1 5 even after the re-fill zero bits transmitted;
          • 0 to 5 and even 1 bits after deleting at the receiving end;
    • When used in a PPP asynchronous transfer, to the use of a special character filling.
  • PPP protocol does not use numbers and acknowledgments
    • When there is less probability of errors in the data link layer, using relatively simple PPP protocol is more reasonable.
    • In the Internet environment, PPP information field into the data IP datagram. Reliable transmission of data link layer can not guarantee the transport network layer is also reliable.
    • A frame check sequence FCS field can ensure error acceptable.
  • Working status of PPP
    • When users dial the ISP, modem dial-up router made to confirm and establish a physical connection.
    • PC sends a series of LCP packets (PPP encapsulated into a plurality of frames) to the router.
    • The PPP packets and their responses select some parameters, and the network layer configuration, the NCP is assigned a temporary IP address to access a new PC, that the PC as a host on the Internet.
    • When communication is completed, release the NCP network layer connection, to recover the original IP address assigned out.
    • LCP then releases the data link layer connection.
    • Finally, the release of the connection to the physical layer.
• LAN data link layer
  • LAN: The network is owned by a unit, and the geographical scope and the number of sites are limited.
    • It has a broadcast function, easy access to the whole network from one site. Hosts on the LAN can share the LAN connection on a variety of hardware and software resources.
    • Easy to extend the system to evolve gradually and the position of each device can be flexibly adjusted and changed.
    • Improve the reliability, availability and survivability systems.
  • Three factors determine the performance of the LAN
    • Network topology
    • Transmission medium
    • Medium access control method
• Dynamic media access control (multiple access)
  • Random access
    • With Collision Detection Carrier Sense Multiple Access CSMA / CD ---- Ethernet
  • Controlled Access
    • Token ring token ring network token ring -----------------------------
• Two sublayers of the Data Link Layer
  • Logical Link Control LLC (Logical Link Control) sub-layer;
  • Media access control MAC (Medium Access Control) sub-layer.
• Role adapter (NIC) of
  • Serial / parallel conversion;
  • The data cache;
  • Ethernet protocol (IEEE802.X).
• Ethernet service provided
  • Unreliable: Ethernet services are not provided reliable delivery that best effort delivery;
  • Error-free control: When the destination station received erroneous data frames to discard this frame, nothing else to do. Error correction is determined by the high level;
  • High-level repair: If the executives discovered missing some data retransmission, but Ethernet does not know this is a retransmission frame, but as a new data frame to transmit.
• CSMA / CD (carrier sense multiple access with collision detection)
  • Multiple Access: represented in many ways multiple access computer is connected to a bus.
  • Carrier Sense: Each stand must first detection before sending the data to see if there are other computers to send data on the bus, if there is, for the time being not to send data, in order to avoid a collision.
  • Collision detection: the transmission signal is the computer side edge detecting voltage magnitude data channel.
    • When several stations simultaneously transmit data on the bus, the signal voltage swing will increase the value on the bus (superimposed with each other).
    • When a station detects a signal voltage swing exceeds a certain threshold, considered in at least two stations simultaneously transmit data indicating the resulting impact on the bus.
  • Using CSMA / CD Ethernet protocol can not be full-duplex communication, only way alternate communication.
  • Contention period
    • The first data frame transmission station, after transmitting a data frame up to the elapsed time 2τ (twice-end round trip delay) can know the transmitted data frame is subjected to impact.
    • Ethernet-to-end round-trip delay 2τ called contention period, or collision window.
    • After this time the contention of the collision has not been detected, this can certainly send do not collide.
  • Binary exponential backoff algorithm type
    • It found that there was a collision on the bus should stop sending data immediately and defer (retreat).
    • Determining backoff time base, usually taken as contention-2 [tau];
    • Definition of retransmission times k, k ≤ 10, i.e., k = Min [number of retransmissions, 10];
    • From the set of integers [0,1, ..., (2 ^ k) -1] in a random number taken, denoted as r. Retransmission delay is desired substantially backoff time r *;
    • When the weight 16 can not successfully convey ie discard the frame, and high-level report.
• Some features hub
  • The electronic device used to simulate the actual work of the cable, so that the entire system is still run as a traditional Ethernet;
  • Use an Ethernet hub logically is still a bus network, each workstation or CSMA / CD protocol, and a shared bus logically;
  • Much like a multi-repeater hub interface operates at the physical layer.
• Ethernet channel utilization
  • Ethernet situations channel is occupied by:
    • Contention period length 2τ, that is twice the propagation delay of end to end. The interference signal is not transmitted after a collision is detected;
    • Frame length is L (bit), the data transmission rate C (b / s), and thus the transmission time frame is L / C = T0 (s);
  • To improve Ethernet channel utilization, it is necessary to reduce the ratio of τ and T0.
  • A parameter defined in the Ethernet, which is more than one way end to end Ethernet transmission delay τ time T0 of the frame;
    • a → 0 indicates a collision can be detected immediately, and immediately stops transmission, and thus a high channel utilization;
    • a greater, indicating that the proportion of occupied contention is increased, a collision occurs on every lot of waste of channel resources, so that channel utilization significantly reduced.
  • Smax is the maximum channel utilization
    • In the idealized case, the stations on the Ethernet transmission data will not collide;
    • A transmission line is occupied time T0 + τ, and the frame itself is the transmission time T0, so Smax = 1 / (1 + a).
• Ethernet MAC layer
  • Hardware address MAC layer: In the LAN, the hardware address is also called the physical address, or MAC address.
  • Each adapter receives a MAC frame from the network it first checks the MAC address in the MAC frame hardware:
    • If the frame is destined for the own station to accept, then perform other processing;
      • Frame sent to the site "includes the following three frames:
        • Unicast (unicast) frame (one to one)
        • Broadcast (Broadcast) frame (a pair whole)
        • Multicast (multicast) frames (one to many)
    • Otherwise, this frame is discarded, no longer perform other processing.
• In the extended LAN physical layer
  • Host uses an optical fiber and a pair of fiber optic modem is connected to the hub;
  • Extended LAN hub with:
    • Advantages: causing a computer on the LAN originally belonging to different collision domain is able to communicate across the collision domain. Expanding the geographical scope of coverage of local area network;
    • Cons: collision domain increases, but did not improve overall throughput. If different collision domains using different data rates, then the hub can not be used to interconnect them together.
• In the extended Ethernet data link layer
  • Extended Ethernet more common method is carried out in the data link layer;
  • Early use of the bridge, now use Ethernet switches.
  • Ethernet switch features:
    • Ethernet switch is essentially a multi-interface bridge, usually have a dozen or more interfaces;
    • Each interface directly to a single host, or connected to another Ethernet switch, and generally operate in full duplex mode;
    • Ethernet switches having parallelism, can simultaneously communicate with multiple interfaces on the plurality of hosts can simultaneously communicate;
    • Hosts communicate with each other are exclusive transmission media, collision-free data transmission.
  • Ethernet switch advantages:
    • User exclusive bandwidth, increased the total capacity;
    • When switched Ethernet to Ethernet from the shared bus, access to all hardware devices and software, and so does not require any changes to the adapter;
    • Ethernet switches generally have a multi-rate interfaces, to facilitate a variety of different circumstances of users;
  • Ethernet switch exchange:
    • Store and forward mode: the entire frame of data be processed before the first buffer;
    • Straight (cut-through) mode:
      • Receiving a data frame immediately while the data frame destination MAC address forwarding interface determines the frame, thereby increasing the speed of the frame forwarding.
      • The disadvantage is that it does not check the error will be forwarded directly to the frame out, it is also possible to forward some invalid frames to other stations
  • Ethernet switches self-learning function:
    • Start switching table is empty;
    • If not find a corresponding host in the table at the time of the interface which will be broadcast to all interfaces;
    • Whenever the host sends a frame, the switch will automatically maintain the exchange list, i.e., recording or updating the source address, the incoming interface and the effective time.
      • In exchange for each item in the table has some effective time expired items will be automatically deleted.
    • If the switching interface is given in the table of the frame entering the switch interface, this frame is discarded because no forwarding.
    • Switch Spanning Tree Protocol:
      • It does not change the actual topology of the network;
      • Logically cutting off certain links, such that a path from the host to other hosts are all loop-free tree structure.
• Split collision domains: switches, routers, bridges;
• Split broadcast domains: a router.
• Fast Ethernet: meet or exceed the rate of 100 Mbit / s Ethernet
  • 100BASE-T Ethernet: Also known as Fast Ethernet, code-named IEEE 802.3u;
    • 100 Mbit transmission over twisted-pair / s star topology Ethernet baseband signal, still using the IEEE 802.3 CSMA / CD protocol;
    • Can work in full duplex mode without conflict, when operating in full-duplex mode, do not use CSMA / CD protocol.
    • MAC frame format is still prescribed 802.3 standard.
  • Gigabit Ethernet
    • Allows full duplex asynchronous half-duplex mode at 1 Gbit / s;
      • Using CSMA / CD protocol in half;
      • Full-duplex mode without using the CSMA / CD protocol
    • Using a predetermined frame format of IEEE 802.3 protocol;
    • Gigabit Ethernet operates in half-duplex mode, it is necessary to collision detection:
    • In order to maintain a minimum frame length of 64 bytes and a maximum segment length of 100 meters, Gigabit Ethernet adds two functions:
      • Carrier extension: that the minimum frame length is still 64 bytes (which can maintain compatibility) while contention time is increased to 512 bytes.
        • Where the transmitted MAC frame length is less than 512 bytes, it is filled in the following frame with special characters, the transmission length of the MAC frame is increased to 512 bytes;
        • After receiving the receiving end of an Ethernet MAC frame, to the filling of special characters after deleting delivered to senior.
      • Packet bursts
        • When much shorter frame to be transmitted, the first frame to a short extension method using a carrier is filled;
        • Then some short frames may be transmitted one by one, leaving only the necessary minimum interframe gap can;
        • This form may be a string of burst packets, until 1500 bytes or slightly more up.
    • When working in the Gigabit Ethernet full-duplex mode (i.e., the communication parties can send and receive data simultaneously), without using a carrier extending and a packet burst.
  • 10 Gigabit Ethernet: work only in full-duplex mode, so there is no contention, nor the use of CSMA / CD protocol.
• VLAN switches across multiple broadcast domains separated.