Notes-data communication and Web tutorials-chapter I

1. Networking topology (network topology)

    1. Common bus topologies (Common bus Top ology): They communicate through a single bus (a bunch of parallel lines). Each device listens to the bus through the appropriate interface to check the data transfer. If the interface determines that the data is sent to the device it serves, it reads the data from the bus and transmits it to the device. When a device has data to be transmitted, its interface circuit detects whether the bus is idle, and if so, sends the data immediately. When two devices simultaneously detects the bus idle and transmits the data to produce a conflict signal, the device transmits continuously listens to the bus, therefore can detect the conflict signal, and stops the transmission, waits for a random event to retry. This process is known as carrier-listening multi-access with collision detection. Advantage: Without central control, adding a   new equipment is very simple, because the devices do not interact with each other. In addition, in a bus network, a device fails or is unloaded without causing the network to crash. Ethernet is a public bus network, and the main advantage of Ethernet is that it is easy to add new devices to the network.                                 &NBSP ;                                 &NBSP ;   2. Star topology (star topology): It uses a central computer to communicate with other devices in the network in a centrally controlled manner. A device that needs to communicate transmits data to the center for computing A, and then the computer then transfers it to the target node. Advantages: The central control of the way to make the responsibility clear. Disadvantage: A central computer failure will directly cause the entire system to crash.                                  &NBS P               3. In the Ring topology (ring topology): The device is connected to a ring, and each device can only communicate directly with one or two adjacent nodes of it, toWhen communicating with other nodes, the information must sequentially pass through each device between the two. The ring network can either make one-way (the transmission is in the same direction) or bidirectional (data can be transmitted in two directions). Disadvantage: When one node is sending data to another node, all nodes between them have to participate in the transmission, so long as one node fails, all the nodes in the ring will not communicate properly.                                  &NBS P                  4. Fully connected topology (Fully connected topology): There is a direct connection between each pair of devices. Communication becomes very simple because there is no longer a need to compete for public lines. Cons: Increased costs, many connections are not fully utilized.                                  &NBS P                          5. Composite topologies: Use a combination of various topologies. A compromise scheme is to divide the network into several ring, star, or bus topologies connected to PCs and other devices. The devices in the same LAN communicate according to the communication standards of their respective topologies. If it is a cross-network transmission, it must pass through a network bridge that connects the two networks.                                 &NBSP ; the   2. osi model     OSI model is a 7-tier model. Each layer implements a specific function and communicates directly with the upper and lower layers. High-level protocols tend to handle user services and various application requests. Low-level protocols focus on   to handle the actual transmission of information.            OSI layer function summary     levels               &NBS P     function 7. Application tier             provides user services such as e-mail, file transfer, etc. 6. Presentation layer             Convert data format, data encryption and decryption 5. Session layer             communication synchronization, error recovery, and transactional operations 4. Transport             network decisions for grouping and reassembly 3. Network layer             Route selection, billing information Management 2. Data Link layer     error detection and correction, group frame 1. Physical layer             Data physical transfer      1. The physical layer (physical layer) physical layer consists of two main components: transport media and connection policies. The transmission media determines how the signal is transmitted. There are three kinds of connection strategies: circuit switching, message exchange, packet switching. Circuit switching   (circuit switching): The line between two nodes is maintained until one of them terminates the communication. That is, the line is used only for communication between the two nodes. Channels that have already been allocated cannot be used for other communications. Circuit switching requires routing and connectivity before starting the transmission of information. Moreover, the network will maintain this line until one party terminates the communication. This connection is most useful when two of nodes require continuous communication.   Message exchange (message switching): When a message (unit information) needs to be transmitted, the network uses the message exchange to establish the route. Because the messages are stored by each passing node, the network using this method is also known as the storage and forwarding network (store-and-forward network). In the message exchange, the messages are temporarily stored by each node. Different messages may go through different routes, and different messages can share the same public line on a single tick.   Packet Switching (PaCket switching): Minimizes the impact of problems caused by transmitting long messages in message exchange. Long messages may exceed the buffer capacity of the nodes and may cause the lines between adjacent nodes to be occupied for a long time. A line failure can cause the entire message to be lost. In the packet switching network, there are two kinds of common routing methods of datagram and virtual circuit. In the way of datagram, each packet is transmitted independently, which means that the network protocol treats each packet as a separate message and routes it. This approach allows the routing strategy to take into account the actual changes in the network environment, and if a path is blocked, it can change the route. Disadvantage: When the message contains many groupings, the independent routing will result in huge expense; the groupings may not reach the target node in the order in which they are sent. Using virtual circuit, the network protocol establishes a line (virtual circuit) before sending the packet. All groupings are transferred through the same path to ensure that the target node is reached correctly in order. Lines are not dedicated, and different virtual circuits can share a common network line.          Connection policy comparison     strategy                   &N Bsp             benefits                       & nbsp                         downside circuit switching         fast. For cases where transmission delay is not allowed         Because network lines are private, other routes are not      /                                     ,         &NB Sp       &nbsp                         to use. As with telephone calls, both parties must                               &NB Sp                          ,         &NB Sp                   simultaneously participates in   message exchange         routing is non-exclusive, complete a newsletter and nbsp         Usually the message takes longer to reach the goal                       After the loss, it can be reused immediately. Receiver       land. Since the intermediate node must store the message, the quote                       will not need to immediately accept the message     &NBSP ;                         ( ) is too long and can also cause problems. Message end still in use                                 &N Bsp   &NBSp                          ,         &NB Sp             First-set routing, regardless of whether the network condition has been   changed    group switching         When blocking occurs, Number of packet switched networks           Because each packet is routed separately, the cost will be increased                 &NBS P     It is reported that the different grouping options for messages are not         Plus. You must select a route for each grouping. In the datagram                       co-routing, so you can better use the network       &NB Sp  , grouping may not arrive in order       2. The data link layer, when the physical layer sends and receives data, the data link layer is responsible for ensuring that it works correctly. Competition occurs when more than two nodes are simultaneously transmitting on the same media. Solution to the competition: carrier-listening multi-access with collision detection, the device avoids collisions by listening to the bus. The bus is busy, the device temporarily does not transmit, if both devices detect the bus idle, and start transmitting data at the same time, there will be a conflict, usually will later try to transfer. It is not possible to prevent multiple nodes from transmitting at the same time on a common bus, but rather to react to the conflict; token passing is a competitive mechanism to prevent conflicts by calling a series of loops through the specific bitstream of all network nodes called tokens. The node must obtain a token to transmit, attach the token to the end of the sent message, and change the token's control bit to indicate that the token is occupied. The message is then sent to the target node to obtain a token, select the transmission according to different protocols, or pass the token to the next node. Token delivery is not only applicable to ring   nets, it can also be applied to any other network topology   in structure. As long as the node number, you can make the   card in order to cycle through the number.   The receiver's data link layer is error-Detection to determine if an error occurs, and an error occurs   It usually requires a message to be re-routed. The parity bit is the simplest detection method, but it cannot detect all errors. At the same time, the data link layer also makes error correction (Correction),  to correct the corrupted data bits.       3. The Network Layer Network layer provides the transport layer with the ability to establish end-to-end communication. The network layer contains an algorithm for finding the best path between two points. Routing algorithm when looking for the quickest and cheapest route for a particular node, there are many factors that must be taken into account, such as line cost and line effectiveness. The Network layer protocol on the network node is a common route choice. Problems when choosing routes: New sites and nodes are regularly added to the network and have an impact on existing routes and their costs. The algorithm must   have the ability to cope with the conditions of change; the congestion caused by overloading of the computer on the route often causes some communication to be discarded, and the network layer protocol has to be able to notify the sender of the loss of some packets, and when the new best path occurs due to the change of the network environment, Information transmitted through other routes   will be diverted. The   network layer controls the communication subnet (Communications Subnet). The so-called communication subnet is a collection of transport media and switching components necessary for routing and data transmission   transmission. The network layer is the highest level for subnets. This layer may also contain billing software whose cost depends on the amount of data transferred, or it may be related to the time period for using the network. The network layer records this information and is responsible for billing.       4. Transport Layer (Transport layer) Transport layer is a transitional level. One of the functions of the transport layer is to provide a reliable and efficient network connection. It allows the above three layers to perform their tasks independently of the actual network structure   constraints. At the same time, it relies on the following three levels to control the actual network operation. The primary responsibility of the transport layer is to provide reliable and efficient transmission lines for the session layer.   Transport layer can establish multiple connections   (i.e. multiplexing, multiplexing) in the network. After the transport layer divides the data, it transmits each part to different network   nodes, which is called down multiplexing (downward multiplexing). Several transport users share a single node, a process called up multiplexing (upward multiplexing).   The transport layer is also responsible for buffering tasks on network nodes. Both the source node and the target node may need to be buffered. When the transport layer of the sender receives the transmitted data from the session layer, it divides the data into a transport protocol data unit (Transport Protocol data Unit,tpdu), which is then transmitted to the receiver's transport layer (via the lower level protocol). The Transport Layer protocol requires an acknowledgment that is received correctly for each received TPDU. Assume that transport user A wants to teleport   several TPDU to User B. In some cases, these tpdu are cached at the source node and queued for low-level protocol-in   line transfers. After each TPDU is sent, the transport layer does not remove it from the queue, but waits for a confirmation that has been received correctly. If the protocol requires each T P D u to have a confirmation, then the transport layer of the sender must wait. If the   has not received confirmation after a period of time, it will re-send the TPDU, then the TPDU should still be in the buffer queue, has not been removed. When the transport layer of the receiving party receives a TPDU, it must save the TPDU until the session layer takes it away. Remember: The session layer does not necessarily take TPDU immediately. If the receiver knows that the sender caches all the TPDU, it can only leave a buffer position to receive T P D U. The disadvantage of this is that the receiver will have no space to accept subsequent TPDU until the cached T P D U is not taken away by the session layer. In this case, the receiver ignores these TPDU and refuses to send back acknowledgments for them. The sender cannot receive the confirmation and can only resend the message. If the network is reliable (that is, there are few errors), TPDU can also be cached on the target node. If you can   confirm that the receiver will cache the data, the sender does not need to do so again.   Connection Management (Connection Management) is the co-  protocol that the transport layer must follow to establish and release connections between two nodes. Tomlinson developed the three-time handshake (Three-way handshake)   protocol. Its working process is this:  1) User A transmits a TPDU that requests a connection   once. Its serial number is X.  2) User B echoes a TPDU that confirms the request and its serial number. Its serial number is Y. 3) User A sends back an acknowledgment to User B's confirmation frame by including the serial number x and y in the first data TPDU.       5. The session Layer session layer includes the protocols necessary to establish and maintain a connection or session between two end users. If a transport connection is interrupted due to a network failure, the session layer will request another transport connection fromThe session is not interrupted.   has three ways to effectively improve communication quality, and allows the session layer to define the logical units of information, which are session management, synchronization point, transaction. Session management (Dialog Management): Most communications are full Duplex, i.e. the data can be transferred   simultaneously in two directions. In half-duplex (Half-duplex) communication, data can be transmitted in two directions, but it must be rotated in turn. The session layer can also manage half-duplex communication. By exchanging a data token, it coordinates the session,  so that only one user can transfer the data at a time.   Sync points (synchronization point): Session-level users can avoid large losses by defining the same   steps in the data flow. These points divide the data into distinct session units. Also define   recovery points to prevent errors. There are two synchronization points: the primary synchronization point and the secondary synchronization point. At each primary synchronization point (Major synchronization  point), the receiver's session layer must acknowledge that the session unit has been successfully received by the intended user, so that the sender knows that the session unit has reached its original destination. If an error occurs before confirmation, the sender can resynchronize, or re-route the data from the nearest synchronization point. It is the responsibility of the sender to keep a backup of the session unit. However, once the session unit is confirmed, the sender can delete the backup and free up the cache space. The Minor synchronization point is similar to the primary synchronization point. The sender can insert a secondary synchronization point in the session unit. The user can resynchronize in the current session unit based on the secondary synchronization point. The secondary synchronization point does not require confirmation to mitigate the network load. On the other hand, the sender cannot release space for caching session units.      (Activity): Users can also insert another structure in the data, defining the unit of work-the transaction. and session units, transactions are defined and separated from each other, independent of each other. The session layer of the sender identifies the start and end of a transaction using a specific command. When the session layer of the receiving party detects the start flag of a   transaction, all input commands are put into the buffer until the end flag of the transaction is found. Only then will it pass these commands on to high-level protocols, allowing them to perform the necessary tasks.                                  &NBS P        6. Presentation layers (Presentation layer) First, we want to distinguish between information and data, and simply say that the computer stores not information, but data. The information room artificially gives meaning to the data. Fundamentally, data is the classification of bits, bytes, and other things that cannot be expressed. Information is a man-made explanation. The presentation layer must be aware of the system it is serving, as well as the format of the data it receives from other systems. It must ensure that information is transmitted correctly on the network. Another function of the   presentation layer is data compression (Compression). It can reduce the number of bits of the letter   interest on the basis of preserving its original intent. Another reason to change (or encrypt) the bit before sending is security. If someone intercepts the transmitted   information without permission, the information will be unreadable due to the encryption. To understand the meaning of the information, the data must be decrypted.       7. The application layer (aplication layer) is responsible for communicating with users and applications. It is called the application layer,  because it contains network applications. Typical Web applications include World Wide Web applications, e-mail, file transfer, virtual terminal protocols, and distributed systems. The lower layer   Protocol of the model provides a means of representing information and is responsible for transmitting the information to the destination. The e-mail protocol (email Protocol) located at the application level is responsible for defining the architecture of the e-mail system. It stores the message in a mailbox (actually a file). This allows users to organize, read, and respond to information.   File Transfer Protocol (File Transfer Protocol) also allows users to exchange information, but it takes a different approach. It   usually allows users to connect to a remote system, retrieve the above directories and files, and copy them to the user's system.   Virtual Terminal Protocol allows end users to connect to a remote computer across the network. Once the   is switched on, the user can interoperate as if it were the actual terminal.   Distributed Systems (Distributed system) is another developing technology in the field of computer network. It allows many devices to run the same software and access public resources. Typical resources include workstations, file servers, hosts, printers, and so on.

Notes-data communication and networking tutorials-chapter I