Edit this paragraph Profile
Bridge will connect two similar networks, and network data flow management. It works on data link layer, not only to extend the network distance or range, and can improve network performance, reliability and security. Network 1 and Network 2 through a connecting bridge, the bridge receives a packet sent network 1, checking the address of the packet, if the address belongs to the network 1, it will abandon its contrary, if the address of the network 2, it will continue to send to the network 2, so that information can be isolated using the bridge, with a network number will be divided into multiple segments (belonging to the same network number), to isolate the security segment, to prevent unauthorized access to other users of the network segment. As the network segments, each segment is relatively independent (of the same network number), a segment of the fault will not affect the operation of another segment. Bridge can be dedicated hardware devices, you can also install the bridge by computer software to achieve, then the computer will be installed on multiple network adapters (network cards).
Edit this paragraph function
Bridge function is similar to an extended network of repeaters on the span, but it can provide intelligent connectivity services that, according to the destination address in the frame segment to which to forward and filter. Bridge the understanding of the segment in which the site is relying on "self learning" to achieve, a transparent bridge, the conversion bridge, bridge package, source routing bridges. When using the bridge to connect two LAN, the bridge segment 1 from the MAC frame, first check the destination address. If the frame is sent to a station on the segment 1, the bridge is not forwarding the frame to segment 2, but will filter out; if the frame is sent to a station on the segment 2, the bridge It will be forwarded to the segment 2, which shows that if one each on LAN1 and LAN2 in Ben Wang segment users to communicate simultaneously, is clearly achievable. Because the bridge played a role in isolation. It can be seen, under certain conditions, the bridge has increased the role of network bandwidth. Bridges store and forward repeaters compared with the advantages and disadvantages, its advantages are: use bridges to interconnect to overcome the physical limitations, which means Goucheng LAN segment of the total number and the number of data points it is easy to expand. Bridge into the store and forward capabilities to adapt them to use different MAC protocol for connecting two LAN, LAN and thus constitute a different mix together the mixed network environment. Bridge relay function depends only on the MAC address of the frame, and thus high-level protocol is completely transparent. Bridge will be a large LAN into segments, will help to improve reliability, availability and security. The main drawback is the bridge: a bridge in the implementation due to receive before forwarding the frame and the buffer, compared with the repeater will introduce more delay. As the bridge does not provide flow control, so the flow is large has the potential to overload, resulting in the loss frame. Bridge the advantages outweighed the disadvantages is the reason for its widespread use. Bridge work in the data link layer, to link the two LAN, forwarding frames based on MAC addresses, can be seen as a "low-level router" (routers work at the network layer, such as the IP address of the network address forwarding). Usually through a slow remote bridge link (such as telephone lines) to connect two remote LAN, the local bridge, the performance is more important, while the remote bridge is concerned, can be in the long run is more important. Comparison of bridge and bridge router forwards the frame does not understand high-level protocol, which allows it to handle both the same IP, IPX and other protocols, it provides a non-network routing protocols (such as NetBEUI) sub-functions. As the router network layer processing of data, so they are easier to interconnect different data link layer, such as Token Ring and Ethernet network segment segment. Bridges are usually difficult to control than the router. Protocols such as IP routing protocols are complex, so easy to manage network routing; IP and other protocols also provides more information on how the network segment (even if its address is also provided such information). The bridge only with the MAC address and physical topology to work. Therefore, the bridge is generally suitable for small simple network.
Edit this paragraph with the reasons
Many units have more than one local area network, and hope to connect them. The reason why a unit has more than one local area network, the following six reasons: First, many university departments or the company's departments have their own local area network, primarily used to connect their personal computers, workstations and servers. As each department (or departments) of different nature of work, so use a different LAN, the system (or sector) to be interaction between the morning and evening, so need to bridge. Second, a more geographically dispersed units, and the far distance, with installation of a coaxial cable network across all locations, as a local area network at each location, and use the bridge and linking the infrared link, such costs may be lower. Third, it may be necessary on a single logical LAN into multiple LAN to adjust the load. For example, the number used by the bridge to connect LANs, each LAN has a group of workstations, and has its own file server, so most communication limited to a single LAN, a backbone to reduce the burden. Fourth, in some cases, a single local area network from the load point of view there is no problem, but the farthest away from the physical distance between the machine is too far (for example, more than 802.3 required 2.5km). Even if the cable laying is not a problem, but because of round-trip delay is too long, the network will not work. The only way is to segment the LAN, the bridge is placed between paragraphs. By using the bridge, the work can increase the total physical distance. Fifth, the reliability issues. In a single LAN, a defective node continuously output of useless information flow will seriously damage the operation of the LAN. Bridge can set a key component in the LAN, like fire doors within the building as a single node prevent disorders and undermine the entire system. Sixth, the bridge will help security. Most have a mixed LAN interface work (promiscuousmode), in this way, the computer receives all the frames, including those not addressing the frame sent to it. If the network set up in many places and carefully block the bridge to forward important information do not, then the network can be separated to prevent information theft.
Edit this paragraph compatibility
One might naively think that a 802 from the LAN to another LAN bridge 802 is very simple, but it is not. 802.x to 802.y in the nine combinations, each of which has its own special problems to solve. In discussing these particular issues, try to look at these general problems facing the bridge. First, a variety of local area network using a different frame format. This incompatibility is not caused by the technical reasons, but simply because the company supports the three standard (Xerox, GM and IBM), no one is willing to change their standards are supported. The result: copy between different LAN frame to rearrange the format, which requires CPU time, recalculate the checksum, but also may produce memory errors caused by the bridge can not detect the error. The second problem is connected to the LAN is not required according to the same data transfer rates. When the slow speed of the LAN to the LAN to send a long list of consecutive frames, the bridge handle frames slower than the frame into. Bridge must be too late to deal with the frame buffer memory, and have to guard against running out memory. Even 10Mb / s of 802.4 to 10Mb / s of the 802.3 bridge, to some extent the existence of such problems. Because the bandwidth consumed by 802.3 parts of the conflict. 802.3 It is not actually 10Mb / s, and 802.4 (almost) does is 10Mb / s. Bridge bottlenecks associated with a subtle but important issue is the value of the timer on each floor. If the 802.4 LAN network layer to send a long message (frame sequence). After the issuance of the last one, it opens a timer, waiting for confirmation. If this message must go through the slow 802.5 network bridge, then the last one before being forwarded to the low-speed LAN, timer, there may be time to. Network layer may think that the entire frame is lost and re-send the message. After the failure of several transmission, the network layer will give up and tell the transport layer transmit the destination site has been shut down. Third, all the problems, it may be the most serious problems are three 802LAN have different maximum frame length. For 802.3, the maximum frame length depends on the configuration parameters, but the standard 10M/bs system has a maximum payload of 1500 bytes. 802.4 maximum frame length is fixed at 8191 bytes. 802.5 no upper limit, as long as the site of transmission time does not exceed the token holding time. If the token time defaults to 10ms, the maximum frame length is 5000 bytes. An obvious question arises: When a long frame must be forwarded to the LAN can not receive long frame, it will happen? In this layer does not consider the frame into small pieces. All agreements are assumed to either reach or not reach the frame, no provision for the restructuring of smaller units framing. This does not mean you can not design such an agreement, you can design and have this agreement, but 802 does not provide this functionality. Basically, the problem can not be solved, must be discarded too long to forward frames. The level of transparency so it was.
Edit this paragraph two kinds of bridges
The first bridge is transparent to the 802 bridge (transparentbridge) or Spanning Tree Bridge (spanningtreebridge). Support of this design were the primary concern is completely transparent. According to their point of view, with multiple LAN buy units in the IEEE standard bridge, you can simply plug to connect the bridge, everything will be fine. Without changing hardware and software, no need to set the address switch without load routing tables or parameters. Short, doing nothing, just insert the cable on the bin, not subject to the operation of the existing LAN bridge in any way. This is really incredible, they finally succeeded. Transparent bridge to work in promiscuous mode, it receives all the LAN connection with transmission of each frame. When one arrives, the bridge must decide whether to discard or forward. If you want to forward, you must decide which sent LAN. This need to query a large bridge in the hash table in the destination address to make a decision. The table lists each possible destination, and it belongs to which an output line (LAN). Insert the beginning of the bridge, all of the hash table are empty. As the bridge does not know the location of any destination, thus the use of diffusion algorithms (floodingalgorithm): the arrival of each destination unknown even in this frame output to all the LAN bridge (except to send the frame of the LAN). Over time, the bridge will learn the location of each destination. Once you know the destination location, where the frame is sent to the appropriate place on the LAN only, and no longer distributed. Transparent bridge algorithm is used in reverse learning (backwardlearning). Bridge work by mixed mode, so it can see the connection of any frame on the LAN. View source address on the LAN can know which machines can access which, so add in a hash table. When the computer and the bridge power, power, or migration, network topology will change. To deal with dynamic topology problem, increase the hash table entry each time, are indicated in the frame arrival time. Whenever the destination table in the frame arrived, will update the current time. Thus, from the table each time the machine can know the arrival time of the final frame. Bridge, there is a process periodically scans the hash table, clear the time is earlier than the current time a number of minutes of all entries. So, if removed from a computer on the LAN, and elsewhere to re-connect to the LAN, then within minutes, it can start to work without human intervention. This algorithm also means that if no action within a few minutes the machine, then sent to the frame will have to distribute it, until it sends out an own up. Route selection process to reach the frame sent on the LAN (the source LAN) and destination of the LAN (the purpose LAN), as follows: (1) If the source and destination LAN the same LAN, then discard the frame. (2) If the source and destination LAN LAN is different from the forwarding of the frame. (3) If the destination LAN is unknown, to spread. In order to improve reliability, it was set up between the LAN in parallel two or more bridges, but this configuration caused other problems, because the topology produced a loop, may lead to an infinite loop. The solution is to talk about the following Spanning Tree (spanningtree) algorithm. To solve the above mentioned infinite loop problem is to bridge communicate with each other, and each LAN with a reach spanning tree topology of the actual coverage. Spanning Tree, you can ensure that any two there is only one path between the LAN. Once the agreement is good spanning tree bridge, LAN transmission among all spanning trees are to comply with this. Due from each source to each destination there is only one path, it is impossible to have cycles. To build the spanning tree, we must first choose a bridge as the spanning tree root. This is accomplished by each bridge broadcasts its serial number (the serial number set by the factory and ensure the world's only), choose the smallest sequence number as the root bridge. Then, to each according to the root bridge to construct the shortest path spanning tree. If a bridge or a LAN failure, then re-calculated. Bridge through the BPDU (BridgeProtocolDataUnit) communicate with each other, in the bridge configuration to make their own decision, each bridge and each port requires the following configuration data: bridge: bridge ID (unique identifier) port: port ID (unique identifier) the relative priority of each port's port cost (high-bandwidth = low cost) each configured bridge, the bridge will be determined automatically according to configuration parameters spanning tree, this process has three stages: (1) Select the root bridge's bridge with the smallest bridge ID is selected as the root bridge. Bridge ID should be unique, but if two bridges have the same minimum ID, the MAC address of the small bridge is selected as the root. (2) In all other bridges on the selected port eradicate the root bridge election outside of each bridge needs a root port, it should be the most suitable communication port with the root bridge. Each port by calculating the cost to the root bridge, take the smallest as the root port. (3) Select the LAN of each "designated (designated) Bridge" and "designated port" If there is only one bridge connected to a LAN, it must be the designated bridge of the LAN, if more than one, then to the root network Bridge was selected to spend the minimum specified in the LAN bridge. Designated bridge and designated port to connect the corresponding LAN (if more than one such port, then the low priority to be elected). A port must be one of the following: (1) Root Port (2) a designated LAN port (3) blocking the port when a bridge is powered on, it assumes that he is the root bridge, sending out a CBPDU (ConfigurationBridgeProtocolDataUnit), inform it that the root bridge ID. A bridge receives a root ID bridge ID than his knowledge of CBPDU, it updates its own table, if the frame is from the root port (upload) to reach, then to all the specified ports (downstream) distribution. When a bridge receives a root bridge ID is greater than the known ID of CBPDU, the message is discarded if the frame arrived from the specified port, a frame is sent back to tell the true roots of the lower bridge ID. When intentionally or due to line faults caused by network reconfiguration, the process will repeat, resulting in a new spanning tree.
Source routing bridges
Transparent bridge has the advantage of easy to install, simply inserted into the cable and you're done. But on the other hand, this bridge is not the best use of bandwidth, because they only used the topology of a sub-set (spanning tree). The two (or other) the relative importance of factors led to a split within the 802 Committee. Support CSMA / CD and token bus people choose a transparent bridge, and Token Ring's supporters called source routing preference (sourcerouting) the bridge (by IBM's encouragement). The core idea of source routing is assumed that each frame of the sender knows the receiver is on the same LAN. When sending a LAN to another, the source machine to the destination address set to 1 as the high mark. In addition, it added in the header of this frame should follow the actual path. Source routing bridge is only concerned about the destination address to a high frame when the frame to see this, it scans the routing header, sent this frame to find that the LAN number. If sent to the frame that number followed by the Ben Wang LAN bridge number, then this frame is forwarded to the routing table in the back of their own that LAN. If the LAN is not Benwang number followed by the bridge is not forwarding the frame. This algorithm has three possible implementations: software, hardware, mixed. Concrete realization of these three different price and performance. The first is no interface hardware overhead, but requires fast CPU processing all incoming frames. Finally, implementation requires a special VLSI chip that share a lot of bridge work, so the bridge can use a slower CPU, or you can connect more LAN. Source routing on the premise that the Internet each machine knows the best path for all other machines. How to get the source routing algorithm routing is an important part. Routing algorithm to obtain the basic idea is: If you do not know the location of the destination address, source machine on the release of a radio frame, asking where it is. Each bridge is to forward the search frame (discoveryframe), so that the frame can reach the Internet, each LAN. When the reply came back, via the bridge to record their own identity in the reply frame, then, radio frame sender can get the exact route, and select from the best route. Although this algorithm can find the best route (it finds all routes), but also facing the problem of the explosion frame. Transparent bridge is somewhat similar situation will occur, but not so serious. The spread is based on spanning tree, so the total number of frames transmitted is a linear function of network size, rather than as the source routing is an exponential function. Once the host to find a route to a destination, it will be deposited into the high-speed buffer, look no further. While this approach greatly curb the explosion frame, but it adds to all of the transactional burden on the host, and the whole algorithm is certainly opaque.
Comparison of two kinds of bridge
Transparent bridge is generally used to connect an Ethernet segment, and source routing bridges are generally used to connect Token Ring network segment.
Edit this paragraph remote bridge
Bridge is also sometimes used to connect two or more far away from the LAN. For example, a company located in multiple cities, each city the company has a local LAN, the ideal situation is that all of the LAN are connected, the whole system is like a large LAN. This goal can be achieved by the following method: each set in both a LAN bridge, and use point to point connection (such as leased telephone company phone lines) to connect them together two by two. Point to point connection can be a variety of different protocols. One way is to use a standard point to point data link protocol, the MAC frame will be added to complete the payload. If all of the LAN are the same, this approach works best, it's the only problem is that the frame must be sent to the correct LAN. Another approach is to remove the source MAC of the bridge head and tail, and the remaining part of the added point to point protocol in the payload, then the purpose of generating new bridge in the head and tail. The disadvantage is that the destination host is not the source host checksum calculated by checksum, so the bridge damage arising from certain memory errors may not be detected.
Edit this paragraph the basic working principle
Data link layer device is connected bridge (bridge), it played in the networking of data reception, data forwarding address filtering and the role of multiple networks for the exchange of data between systems. The basic characteristics of a bridge. Bridge on the data link layer LAN interconnection; 2. Bridge to interconnect two different data link layer protocol, different transmission media and different network transmission rate 3. Bridge to receive, store, address filtering and forwarding means to achieve the interconnection between the communications network; 4. Bridges need to be interconnected network over the data link layer protocol using the same 5. Bridge can separate the traffic between two networks, interconnection networks will help to improve the performance and safety.
Edit this paragraph is different from the repeater and bridge hub
Bridge is determined by the logic to determine how the transmission frame. This logic is based on the Ethernet protocol, in line with the second layer of the OSI standard. So the bridge can be seen as the second layer of the device. Used to determine when to forward frames. Bridge edited as follows: 1. Check the received signal, to explain the meaning of 0 and 1, and find the destination MAC address of the frame. 2. If the purpose of the MAC address with the frame through different interfaces on the bridge to reach the destination (not the frame to reach the bridge that interface), then re-generate the signals to transmit through this frame. (This process is called forwarding.) 3. If the frame reaches the destination address of the interface is up to the port, then discard the frame. (This process is called filtering.)