Say it in the front.
The hub runs on the first layer of the OSI model, the switch on the second layer, and the router on the third layer.
Network hardware devices, including: network cards, hubs, switches, routers, connected network devices, gateways and bridges, and test network devices.
The network card can be plugged into the computer's motherboard expansion slot or centrally on the motherboard, as well as a wireless card connected via the USB interface.
Desktop network card, also belongs to PCI-E network card laptop network card
Wireless card wired network card, this also belongs to the motherboard integrated network card
Because the current network has ATM network, Token Ring network and Ethernet points, they use their respective network card, so the network card also has ATM network card, Token Ring network and Ethernet points.
According to the connector interface, sub-BNC connector NIC, RJ-45 Connector network card.
According to the different motherboard slots, sub-PCI-E network card, PCI network card, USB card and motherboard integrated network card.
According to the different working objects of the network card, the sub-server special card, ordinary workstation network card, laptop special network card (PCMCIA card specially designed for notebook computer network card ), wireless network card.
How to install a desktop card
Computer use generally need to be able to surf the internet, and the Internet must have a network card. Now the general computer motherboard will have built-in network card, but if the internal network card is damaged, you need to plug in a network card, network cable to this network card to normal use, the following describes how to install the network card.
Step One: First open the computer mainframe box, we will see the host box in the PCI slot, note that the slot on the socket and the network card on the missing tone does not match.
Step two: Before the network card inserted into the PCI slot, should pay attention to the NIC iron plate to avoid scratching to the motherboard.
Step three: You can press the NIC into the PCI slot until the metal pins are fully inserted.
Step four: Then on the network card on the top of the screw hole with the main box docking, screw the screw to make its network card become more stable.
Step Five: Finally, we will test whether the NIC is connected to the motherboard. We use an Internet cable to plug in the network card RJ45 interface, the normal network card indicator lights up.
Note: After installing the network card is not necessarily able to connect the Internet, because some network cards need to install network card driver to access the Internet. Therefore, it is necessary to install the corresponding model card driver.
Install the driver of the network card, there is not much to repeat, the online data too much.
A hub is just a multiport repeater with one port connected to the backbone and multiple ports connected to a group of workstations. In addition to being able to connect to Macintosh and personal computer workstations, hubs can also connect to print servers, switches, file servers, and other devices in the network.
Types of Hubs
Separate hubs, stacked hubs, and modular hubs, depending on how they are configured.
Stand-alone hub stacking hub modular Hub
As its name implies, a stand-alone hub that serves a computer workgroup is isolated from other devices in the network. They can be connected to other hubs via coaxial, optical, or twisted pairs, but they are not generally used for layering or daisy chaining. Stand-alone hubs are best suited for smaller, independent departments, home offices, or lab environments. They can be both passive and intelligent. They are installed for a small group of users and are easy to connect.
A stand-alone hub does not follow some kind of fixed design. The number of ports it provides is also not fixed (although they typically contain 4, 8, 12, or 24 ports). Smaller standalone hubs with only 4 ports (mainly designed for small offices or home offices) are also licensed as "hubby", "Hublet" or "mini hubs". On the other hand, a standalone hub can provide up to 200 connection ports. The disadvantage of using this single hub with so many connections is that it can easily lead to a single point of failure for the network. In general, large networks use multiple hubs (or other connected devices).
A stacked hub is similar to a stand-alone hub. Physically, however, they are designed to be connected to other hubs and placed in a separate telecom cabinet; Logically, a stacked hub represents a large hub. ".". One of the great benefits of using a stacked hub is that the network or workgroup doesn't have to rely on a single hub, so you can avoid a single point of failure. The maximum number of such hubs that can be stacked up is different. For example, some hub manufacturers limit the maximum number of stackable hubs to 5, while other hub manufacturers can stack up to 8 hubs.
Although many stacked hubs use an upstream link port for Ethernet, some prefer to use high-speed cables to stack the hubs for better results. The consequence of this is that the products produced on different production lines are incompatible with each other, and even the products produced by the same manufacturer are incompatible. Hubs with standard Ethernet upstream link ports are easily interconnected with the products of other manufacturers. The general rule is that although it is not necessary to use a stacked hub produced by the same manufacturer, it is clear that people prefer to choose the hardware that is already connected to the inside instead of the externally attached hardware. Like stand-alone hubs, stacked hubs can support connectors and transfer rates for different transport media. They may have special handling functions, perhaps not. Although 6, 12, or 24 ports are usually available, they provide a number of ports that are not fixed. Figure 6-10 shows two different stacked hubs
The modular Hub provides a wide range of optional interface options through the chassis. This makes it easier and more flexible to use than standalone hubs and stacked hubs. Like a personal computer, a modular hub has a motherboard and slots so that you can plug in different adapters. Plugged-in adapters enable these modular hubs to be connected to other types of hubs or to routers, WANs, or to the backbone of a token ring or Ethernet network. These adapters can also connect this modular hub to a management station or redundant device, such as an alternate power supply. Because a modular hub can install redundant components, it has the highest reliability in all types of hubs. Another benefit of using a modular hub is that it provides an expansion slot to connect the increased network devices. In addition, they can connect many different kinds of devices. In other words, depending on the needs of the network, the corresponding modular hubs can be customized. However, the price of modular hubs is also the most expensive one. A small network using this hub is a bit overqualified. Modular hubs are almost always intelligent.
Installation of Hubs
As with the installation of a network interface card, it is ensured that the hub is properly installed according to the manufacturer's instruction manual. Most of the time, installing a hub is simple, and some people even think it's easier than connecting a workstation to a network. First, switch on the power supply. See that the power led on the hub is lit, indicating that the power supply is connected. Most hubs perform self-test programs when they are opened. Flashing lights indicate that the self-test process is in progress. When the self-test is complete (the LEDs for most of the hubs are continuously illuminated at this point), plug one end of the connection cable into the hub port and the other end to a switch or router in the backbone or network. The second step is to connect the hub to a workstation or other hub in the same way. As shown in 6-12, after the workstation is connected to the network via the newly installed hub, refer to the Hub's instruction manual to confirm that the connection and communication LEDs are indicated as normal.
If you are installing a stacked hub or a bracket-mounted hub, you need to secure the hub with the screws and pliers that accompany the hub. In the case of a stacked hub, the hub to be stacked must be connected with its dedicated cable or via an upstream link port. Again, it's best to refer to the documentation that came with the hub.
Basic concepts of switches
The basic purpose of developing LAN bridges was to extend the LAN on the number of distances and stations. With the advent of high-end port density bridges that can operate at wire speed, a new LAN has emerged: "Switched LANs". Switched LANs are an alternative to traditional shared-bandwidth LANs. The only obvious difference from products deployed in a structured cabling environment is that hubs are switched (bridges), not shared (repeaters). However, with a shared LAN or switched LAN, the way the network operates varies greatly. In addition, switched LANs provide users with some configuration that is not available for sharing. And all of this comes at a price.
Over the years, with the increase in hardware technology for connected devices, it has been difficult to differentiate the boundaries of hubs, switches, routers, and bridges from each other clearly. Switches this device can logically divide a network into several smaller segments. Unlike a hub that belongs to the first layer of the OSI model, the switch belongs to the data link layer (the second layer) of the OSI model, and it resolves the MAC address information. In this sense, switches are similar to bridges. But in fact, it is equivalent to multiple bridges.
All ports on the switch use the specified bandwidth. It turns out that this is actually a better price/performance ratio than the bridge. Each port of the switch acts as a bridge, and each device connected to the switch can enjoy their own dedicated channel.
Isolate conflict domains
In a shared Ethernet LAN, use the CSMA/CD Mac algorithm to arbitrate the use of shared channels. If two or more stations have frames in the queue that are waiting to be sent, they will collide (collision). A set of competing channel access stations is called a conflict domain. As shown in 6-16, the contention of a station in the same conflict domain leads to conflict and fallback (Backoff). Stations in different conflict domains do not compete for public channels, so they do not create conflicts.
In a switched LAN, the switch port is the endpoint of the conflicting domain on that port. If a port is connected to a shared LAN, there will be a conflict between all the stations on that port, and there will be no conflict between the station of the port and the other port of the switch. If there is only one end station per port, there will be no conflict between any pair of end stations.
Therefore, the switch isolates the conflicting domains for each port.
Segments and differential segments
A switched hub can be used for traditional shared LAN segmentation (segment), as shown in 6-17. Switches used in this way provide a folding backbone (collapsed backbone). Although switch performance for a folded backbone must be high, the model used is still the original, traditional LAN segmentation model.
In addition, the switch can be used for end-station interconnection, as shown in 6-18. Here, each network segment is connected to only one end station, and the LAN segment has reached its maximum level, called the differential segment (microsegmentation).
Micro-staging environment has some interesting features:
(1) There is no conflict between the end stations. Each end station is within its own conflict domain. However, there may still be conflicts between the
conflicting domains ' Macs in the end-station and switch Ports
(2) You can use full-duplex elimination of all Conflicts
(3) each end station has dedicated bandwidth, that is, a differential segment can be used exclusively by a single station.
(4) The data rate for each station does not depend on any other stations. Devices that are connected to the same switch can be run on the network MB/s, MB/s, or MB/s, which is not possible in networks that use shared hubs. Of course, a shared L-N and a single station (differential segment) can be connected at the same time on a switched hub, as shown in 6-16. A station connected to a switch port via a shared LAN features a shared LAN, and the station directly connected to the switch has the function of a differential segment.
Note: from an Ethernet point of view, each dedicated channel represents a collision detection domain. The conflict detection domain is an Ethernet segment that is logically or physically divided. Within a single segment, all devices detect and handle data transfer conflicts. This potential conflict is limited because the switch has a limited number of devices that can accommodate a collision detection domain. The original switch was used to replace the hub and solve the congestion problem of the local area network. While some people think that it is not the best solution, it is a good temporary solution to use a switch in a congested segment. So recently, network managers have used switches to replace routers on the backbone. This makes the switch's sales business so booming. There are at least two advantages to using a switch on the backbone. First, using a switch is usually more secure because the switch makes the data transfer to each device independent. Second, the switch provides a separate channel for each (potential) device. The result of this is that when transmitting large amounts of data and requiring more stringent time delay signals, such as videoconferencing, the ability to fully exploit the network.
The switch itself is still flawed. Although it has buffers to cache input data and contain bursts of information, a large number of successive data transmissions can overwhelm it. In this case, the switch cannot guarantee that no data is lost. In an environment where many nodes share the same data channel, the device conflicts are increased, and each node uses one port of the switch in a network with all the switches, thus occupying a dedicated data channel, which makes the switch unable to provide an idle channel to detect the conflict. In addition, although some high-level protocols, such as TCP/IP, are able to detect and respond to data loss in a timely manner, some other protocols, such as UDP, do not. When a packet of this Protocol is transmitted, the number of collisions will accumulate and the data will be suspended after the final limit is reached. For this reason, when designing a network, you should carefully consider whether the connection location of the switch matches the capacity and information transfer mode of the backbone network.
Switching mode of the switch
Switches can be divided into several different classes. One is the LAN switch, which is suitable for local area network. Although the Ethernet switch
Common, but the LAN switch can also be designed to be suitable for Ethernet or Token Ring network two types. LAN switching
The machine also varies depending on the mode of exchange it uses, with the shortcut mode and the storage and forwarding mode. As for the Exchange Mode of LAN, the
Described in the following two sections.
(1) Quick mode
A switch with quick mode reads the frame header before it accepts the entire packet and decides where to forward the data. Once said, the first 14 bytes of the frame data is the frame header, which contains the target MAC address. With this information, the switch is sufficient to determine which port will get the frame, and can start transmitting the frame (without caching the data or checking the correctness of the data).
What if there's a problem with the frame? Because a switch with Shortcut mode cannot read the frame's checksum sequence when the frame starts transmitting, it cannot use the checksum sequence to verify the integrity of the data. On the other hand, switches with quick mode can detect fragments of data or packets. When a small piece of data is detected, the switch waits until the entire piece of data has arrived before it starts transmitting. It is important to note that data fragmentation is only one of a variety of data disabilities. A switch with quick mode cannot detect problematic packets; in fact, propagating compromised packets can increase the number of errors in the network. The biggest benefit of using the quick mode is that it has a high transmission rate. Since it does not have to stop to wait for the entire packet to be read, the switch forwards the data much faster than the switch with the store-and-forward mode (which you will find in the next section). However, if the data transfer of the switch is congested, the advantage of this time-saving approach is meaningless for switches that use quick mode. In this case, the switch must cache (or temporarily hold) the data as if it were a switch with storage-forwarding mode.
Switches with quick mode are more suitable for smaller workgroups. In this case, the transfer rate is high and the number of connected devices is relatively small, which minimizes the likelihood of errors.
(2) Storage and forwarding mode
A switch running in store-and-forward mode reads the entire frame of data into memory before sending the message and checks its correctness. Although this is a more time-saving approach than using shortcuts, it is possible to store forwarded data in such a way that guarantees accuracy. Because the switch running in store-and-forward mode does not propagate error data, it is more suitable for large local area networks. Conversely, a switch that uses the shortcut mode forwards the data even if it accepts the error. This can cause network congestion if a large number of data transfer conflicts occur in the part of the switch connection. In a large network, failure to detect errors can cause serious data congestion problems.
Switches with store-and-forward mode can also transfer data between segments at different transmission rates. For example, a high-speed network printer that can serve 50 students at the same time can be connected to a one-Mbps port on the switch or allow all student workstations to take advantage of the same switch's ten Mbps port. With this arrangement, the printer can perform multitasking quickly. This feature also makes the switch with storage-and-forwarding mode ideal for environments with multiple transfer rates.
Build a virtual local area network with a switch
In addition, in order to improve the efficiency of bandwidth utilization, the switch can logically merge some ports into a broadcast domain to build a virtual local area network. A broadcast domain is a combination of ports that comprise the second layer of network segments of the OSI model, and it must be connected to a third-tier device, such as a router or a third-tier switch. These ports are not necessarily within the same switch, and may not even be in the same segment. A virtual local area network includes a server, workstation, printer, or any other device that can connect to a switch. Figure 6-19 is a simple virtual LAN. It is important to note that using a virtual local area network is a big benefit of connecting users who are not in the same geographic location, and can build a smaller workgroup from a large local area network.
Note: above we mentioned that the network connected to the switch is the same broadcast domain, in order to improve efficiency we should be free of the broadcast of the occurrence of the impact of other computer work, then, how can switch connected network into multiple broadcast domain? In this case, we need to VLAN partition the network connected by the switch, by default, the network connected by the switch belongs to a VLAN, each VLAN is a broadcast domain, and the VLAN is unable to communicate with each other, If the implementation of the communication between the VLANs must be done using a third-tier device router (this section is discussed in the course of swapping and routing).
The switch must be properly configured to form a virtual local area network. In addition, in order to identify the ports that each logical LAN belongs to, you can do this by setting security parameters, whether to filter the instructions (for example, when the switch disables the forwarding of frames for a segment), restrict the behavior of certain users, and network management options. It is clear that the switch is very flexible to use. Discussing how virtual LANs are implemented is beyond the scope of this book, but if you are responsible for designing a network or installing a switch, you need to delve deeper into the virtual local area network. Some commercial publications (and many switch manufacturers) boast of virtual LANs as the most advanced solution for building networks and the mainstream of the future. (For more advanced technology on virtual LANs, refer to the Exchange and routing course content).
thinking: What is the difference between a hub and a switch?
A router is a multi-port device that can connect different transmission rates and operate in various environments for local and wide area networks, and can also use different protocols. The router belongs to the third layer of the O S I model. The 2nd chapter has said that the network layer directs the data transmission from one network segment to another, and can also guide data transmission from one kind of networks to another. In the past, routers were slower than switches and bridges because of excessive attention to third-tier or higher-level data, such as protocols or logical addresses. Therefore, unlike bridges and second-tier switches, routers are protocol-dependent. Before they can use a protocol to forward data, they must be designed or configured to recognize the protocol.
As in the case of bridges, traditional standalone LAN routers are slowly being replaced by third-tier switches that support routing capabilities. But the concept of routers is still very important. The remainder of this section covers the application of the third-tier switch. Standalone routers are still an option for connecting remote users using WAN technology.
Note: routers and other devices cannot, it is possible to isolate conflict domains, and can also isolate broadcast domains.
Features and functionality of routers
The robustness of a router lies in its intelligence. The router can not only trace a node of a network, but also, like a switch, chooses the most recent and fastest transmission path between two nodes. For this reason, they can also be connected to different types of networks, making them powerful and very important devices in large LANs and WANs. For example, the Internet relies on millions of routers all over the world to connect.
The protocol for packet routing usually can be routed by the router with TCP/IP, ipx/spx and AppleTalk protocols, and the NetBEUI protocol is not capable of routing packets across routers.
Typical routers are internally equipped with their own processors, memory, power supplies, and interfaces for various types of network connections such as Console, ISDN, AUI, serial, and Ethernet ports, and so on. The prepared input and output sockets typically also have a management console interface as shown in 6-17. Powerful and capable routers that support a variety of protocols have several slot ports to accommodate various network interfaces (R-45, BNC, FDDI, ISDN, and so on). Routers with multiple slots to support different interface cards or devices are called stacked routers. Routers are very flexible to use. Although each router can be assigned to perform different tasks, all routers can do the following: Connect to different networks, parse the third tier of information, connect to the optimal data transfer path from point A to point B, and reroute through other available paths if the primary path is interrupted.
The main features of the router:
(1) Routers can interconnect different MAC protocols, different transmission media, different topologies and different transmission rates of heterogeneous networks, it has a strong heterogeneous network interconnection capabilities.
(2) The router is also a storage and forwarding device for WAN interconnection, it has strong WAN interconnection capability and is widely used in Lan-wan-lan network interconnection environment.
(3) Routers interconnect different logical subnets, each subnet is a separate broadcast domain, therefore, routers do not forward broadcast information between subnets, with strong ability to isolate broadcast information.
(4) The router has the function of flow control and congestion control, and can match the speed of the network at different rate to ensure the correct transmission of the packet.
(5) The router works on the network layer, which is related to the Network layer protocol. Multi-protocol routers can support multiple network layer protocols (such as TCP/IP, IPX, DECNET, etc.) and forward packets of various network layer protocols.
(6) The router checks the network layer address and forwards the network Layer data packet (Packet). As a result, routers can filter packets based on IP addresses, and routers use ACLs (Access control list, the accessing controls lists) to control packets encapsulated by various protocols, as well as to filter the port numbers of TCP and UDP protocols.
(7) to micro-segment the large network, the segmented network segment is connected with a router. This can improve network performance, improve network bandwidth, and facilitate the management and maintenance of the network. This is also the approach that shared networks often use to solve bandwidth problems.
(8) The router can not only be used in medium and small LAN, but also in wide area network and large and complex internet environment.
(9) You can isolate conflict domains and broadcast domains.
Because of its customizable nature, it is not easy to install a router. In general, a technician or engineer must be very familiar with routing techniques to know how to place and set up a router in order to perform its best performance. Figure 6-20 shows some ideas about how routers are connected in the network, although this example is somewhat simplistic. The 7th chapter will cover the knowledge of routers used in the WAN. If you plan to design a private network or configure a router, you should study the router technology more deeply.
In the design shown in Figure 6-20, if a workstation in workgroup C wants to use workgroup A's printer, create a connection that contains the address of the printer in workgroup a. The packet can then be routed to hub C. When router C receives the transmitted data, when the third layer of data is parsed, router C will stage the packet. Once the packet is discovered to be passed to the printer in workgroup A, router C chooses the optimal path to transfer the packet to the printer in workgroup a. In this example, the packet may be passed directly to router a. The router increases the number of relays at the end of the packet before it forwards the packet. Router C then forwards the packet to router A, and router a resolves the destination address of the packet before forwarding it to hub a. This transmission is then propagated by hub A to all users in workgroup a until the printer a responds.
Classification of routers
1. Local router
The so-called local router refers to the 6-20 above, each network segment using a router to connect, but only within a limited area network, not across the remote connection.
2. Remote router
Whether it is a local router or a local router, the nature of the router does not change, or the router, but the remote router refers to the router connected to the network segment is a branch in different regions of the remote network, 6-21 below.
routing protocols (RIP, OSPF, EIGRP, and BGP)
For routers, to find the optimal data transmission path is a more meaningful but very complex work. The optimal path may depend on the number of forwarding times between nodes, the current network operating state, the unavailable connection, the data transfer rate, and the topology structure. To find the optimal path, each router communicates with each other through a routing protocol. The difference is that the routing protocol is not equivalent to a routable protocol. such as TCP/IP and ipx/spx, although they may be at the top of a routable protocol. The routing protocol is used only to collect data about the current state of the network and is responsible for finding the optimal transmission path. Based on this data, the router can create a routing table for future packet forwarding. In addition to the ability to find the optimal path, the routing protocol can be characterized by the time it takes for the router to find the optimal transmission path when the network changes or disconnects. Bandwidth overhead-The running network supports routing
The bandwidth required for the protocol is also a significant feature. Although it is not necessary to know exactly how the routing protocol works, you should still have some knowledge of the most common routing protocols: RIP, OSPF, EIGRP, and BGP (there are many more routing protocols, but they are not widely used) also IGRP routing protocol, It is a Cisco device-specific protocol, and other non-Cisco devices cannot use such protocols. The four common routing protocols are described below.
(1) RIP (Routing Information Protocol) designed for IP and IPX:RIP is one of the most previous routing protocols, but it is still widely used because it only takes into account the reason for the number of relays between nodes when choosing the optimal path between two points. For example, it does not take into account the congestion and connection rate of the network. Routers that use RIP broadcast their own routing tables to other routers every 30 seconds. This broadcast can result in a tremendous amount of data transfer, especially when there are a large number of routers in the network. If the routing table changes, it may take a few minutes for the new information to be transferred to a remote location on the network, so the convergence time of the RIP is very long. Also, rip restricts the number of relays to more than 16 hops (over 16 router devices). So, in a large network, if the data is to be relayed more than 16 hops, it can no longer be transferred. Also, RIP is slower and less secure than other types of routing protocols.
(2) OSPF designed for IP (Open Shortest Path First):This routing protocol compensates for some of the flaws in rip and can coexist with rip in the same network. OSPF uses a more flexible algorithm when choosing the optimal path. The term optimal path refers to the most efficient path from one node to another node. In an ideal network environment, the optimal path between two points is the direct connection to the two-point path. If the amount of data to be transmitted is too large, or the data is lost too much during transmission, the data cannot be transmitted along the most direct path, the router will have to choose another route that is also the most efficient path through the other routers. This scenario requires the router to have more memory and a more powerful CPU. This way, the user does not feel that the bandwidth consumed is minimized and the convergence time is short. OSPF is the second most used protocol following RIP.
(3) eigrp designed for IP, IPX, and Apple Talk (Enhanced Internal Gateway Routing protocol):This routing protocol was developed by Cisco Corporation in the mid 1980s. It has fast convergence time and low network overhead. Because it's more than OSPF. EIGRP is easy to configure and requires less CPU, supports multiple protocols, and restricts redundant network traffic between routers.
(4) BGP (Border Gateway Protocol) designed for IP, IPX, and Apple Talk:BGP is a routing protocol designed for Internet backbone networks. The rapid development of the Internet has driven the development of BGP, the most complex routing protocol, to the increasing demand of routers. BGP developers face not only the prospect of being able to connect 100,000 of routers, but also the problem of how to properly and efficiently route through thousands of internet backbones.
Hardware devices for the network