Architecture of diskless networks mainly includes: hardware configuration and network topology, system operation platform and diskless support software. The hardware configuration mainly includes: the configuration of the server and the workstation, the configuration of the network device, the configuration of the Internet device, and the configuration of the accessory device.
When making a hardware configuration plan, you should learn about the following:
Ø Number of workstations
Ø Applying to what occasions, such as education lessons or Internet cafes, etc.
Ø Need to run what software
Ø Requirements for network speed
Ø Some reliability
Ø Expected investment scale
According to the number of workstations and applications, the server's hardware configuration can determine the server's hardware configuration; decide to configure what kind of workstation according to the requirements of the application software, if only general applications, such as frequent office software, learning software, Internet browsing Poor workstation configuration can be competent, if you need to use a large scale application such as image processing or 3D game, the requirements for the workstation are much higher. If the customer is relatively high in the network speed, a 1000MBP exchange of the backbone network can be employed, and the Internet's device is generally or provided by the ISP provider. For some of the network's accessory facilities, such as UPS, anti-static facilities, lightning protection facilities, etc., to be determined according to specific feelings.
The topology of the network is basically a star network, but many details in different diskless networks are different, and the following detailed description.
When the transformation of the old computer room is a Windows 2000 terminal network, when a high-performance server is configured, it is also necessary to add a performance of a good performance. As the central node of the network, the secondary node can be more The inexpensive HUB (which can use the original device) constitutes a folded backbone structure. The so-called folded backbone network is that all workstations are connected to the secondary hub through 100Base-T to the secondary hub connected to the center switch, and the central switch provides high bandwidth, typical 100m switches. The bandwidth requires a few thousand megabits, which can improve the two quantities than the network of shared media, and the server connects to the center switch through a 100Mbps link. The high-speed connection can make the server play the greatest potential, while multiple 10Mbps The terminal provides data, which greatly accelerates the start and running speed of the terminal, and can effectively resist the network storm caused by the shared medium, the folded trunk structure is shown in Figure 1.
Figure 1 folded backward structure
In some cases, the application requirements are not too high, such as the school computer room with a Windows 9x network, etc., and the hardware configuration is low (early Sai Yang generation and supporting equipment), it requires higher bandwidth than the Windows 2000 terminal, it requires higher bandwidth, At this time, the folded backbone structure can be replaced with a 10MBPS switch or a fast Ethernet relay (100 Mbps Hub), which constitutes a layer-based 100Mbps link structure, as shown in Figure 2, 100Mbps HUB provides a 100MPS The conflict domain, for all workstations on which they receive, 100 shared links provide a higher, instantaneous, burst of burst, while the central switch provides greater data total throughput for each workstation. Such a network architecture is a high price, and some old Windows 9X-free networks can be set up in this structure.
Figure 2 Hierarchical 100Mbps link structure
Currently, the price of the switch has declined, so when new diskless networks, the full-duplex exchange Ethernet structure should be used because the diskless network is increasingly demanding data exchange. In order to increase the main network bandwidth, the Gigabit full-duplex link can be used between the server and the central node switch. All network components use switches to connect the switch to each workstation, and all the workstation network cards are all using the full-duplex fast Ethernet card, which constitutes a full-duplex exchange Ethernet, which is shown in Figure 3. The total frequency of the network providing system of the full switch configuration has reached several thousand megabits, which can meet the large-scale applications. Figure 3 Full duplex exchange Ethernet structure
At present, the 1000M network equipment has been mature, and the price has also dropped to the scope of the diskless network, and the Gigabit Ethernet has two standards of copper wire and cable. The copper wire standard is 1000BASE-CX, the maximum transmission distance is 25 feet, using 150 ohm shielded twisted pair STP, transmitted on a dedicated cable TWINAX with a serial rate of 1.25GBPS, and the network cable currently used is non-shielded twisted. (UTP), because UTP's anti-electromagnetic interference ability is weak, can only shorten the transmission distance in a Gigabit environment, which can be used normally. When 4 pairs of twisted pairs, the communication distance can also reach 100 meters. The physical signal system based on the Gigabit Ethernet of the cable transmission is the same physical signal system in the optical fiber communication, and the short wavelength of 850 nm is 1000BASE-XS, the maximum transmission distance is 300 meters. With 1300 nm wavelength, the standard is 1000BASE-LX, and the maximum transmission distance is 550 meters. For longer transmission distances can be supported using a single mode fiber.
In a diskless network, when the workstation is more, the network structure of the internal bridge is often installed in the server, as shown in Figure 4. The server multi-network card can be divided into data, which makes the server network bandwidth increase, but this structure also has some drawbacks, that is, the sub-network workstations cannot communicate directly. Generally, it can be adopted without the need to go to the Internet and each subnet can be employed.
Figure 4 Interior bridge network structure
When installing a large-scale diskless network and a disk-free hybrid local area network, for the convenience of network daily management and network extension, the virtual LAN management mode (VLAN) can be used, and the simple method is to use the switch in the center node. Switch with VLAN features, this switch works in the network layer, provides more management control, providing greater configuration flexibility, very suitable for large-free networks, described in detail below.
In the traditional LAN, the channel conflicts and broadcast storms caused by each site shared transport channel are an important factor affecting network performance. In order to solve the channel conflict in the second layer of the network and the broadcast storm problem that occurs in the third layer of the network, the bridge and routers are widely used in the LAN. The network connected by the bridge is the same logistic subnet, and the logical subsidiary means that the network site in the network has the same network layer address, such as the same IP network number or IPX network number. Communication with different logical subsystems by the router, the communication between the logical subnet must be performed by a router.
In this network structure, the physical network composed of hubs, cables, and cables corresponds to the logical subnet. Usually an IP subnet or IPX subnet belongs to a broadcast domain, so the broadcast field in the network is divided according to physical networks. Such network structures are lacking regardless of the efficiency and security perspective. At the same time, since the site in the network is bound in the physical network of the network, it is not possible to divide it to the corresponding logical subnet as needed, so the structure of the network lacks flexibility. For example, sites that are divided into different hubs cannot belong to the same logical subnet. In order to solve this problem, the concept of a virtual local area network (VLAN) is thus triggered. The so-called VLAN refers to the physical location of the site in the network, and can flexibly add a network technology in different logical subnets as needed.
I. Switching technology is a web technology of VLAN technology in recent years, and the exchange mode mainly has a frame exchange based on a LAN switch and a cell exchange based on asynchronous transmission mode (ATM). The emergence of exchange technology has laid a solid base for the implementation of VLAN.
LAN switch mode
The traditional way to solve the network bandwidth is network fuel, segmentation through the bridge / router, but due to the increase in the bridge / router, the network cost is increased, and the network administrator has brought a significant burden, and when crossing When the service is increased, the bridge / router will become a bottleneck in the network, causing network blocking. The traditional hub is passive, only forwarding information, and does not process information frames. The LAN switch is similar to the multi-port bridge, and the information frame address from the LAN segment is used to read the information frame address from the LAN segment, and a parallel switching path is provided in the Source and destination. The system is composed of an Ethernet Packet Processor (EPP), a swap matrix, and system control module. From the information frame from the Ethernet, enter the EPP. According to the address table provided by the system module, the hardware exchange is performed in the switched matrix, and the generation and update of the address table are completed by the system control module. Such a LAN switch provides a wide-port and network-swap feature that utilizes a value-added software configured in the Network Management, and a logical work group can be formed at any port, and the user can combine into a VLAN in any way.
2 based on ATM
The core technology of ATM is an ATM switch. The main function of the ATM switch is to provide a fast and efficient forward and efficient forward and efficient to the output port. During forwarding, the single cell input, cell header conversion, and output are processed. The letter must be processed according to the requirements of the output port. Output processing is to ensure that the cell enters the appropriate physical link. The ATM switch consists of three substantially portions of the switching mechanism module, the control module, and the input / output interface module. The input / output interface module completes the speed transceiver and some of the required processing, including light / electrical conversion, unpacking / mounting, cell delimitation, virtual channel (VP), and virtual pass (VC) identification identification and conversion , Head error control and routing identification formation, etc. The switching mechanism module completes the transfer from the input date to the specified output port based on the routing identifier. The control mechanism completes the call / connection control, assigns the identification number of the virtual connection to achieve routing information (routing table) generation and update, and the management and assignment of resources. All the ATM switches use hardware processing instead of software processing, determine the output port based on the input cell header, the cell is small (fixed to 53 bytes), so the exchange delay is very school
Second, VLAN work mode
1. VLAN based on the exchange port
This method is a collection of some ports of the LAN switch as a member of the VLAN. These collections are sometimes only on a single LAN switch, sometimes across multiple LAN switches. The management application of the virtual local area network, according to the ID ID of the switch port, distributes different ports to the corresponding packet, all sites assigned to each port of a VLAN are in a broadcast domain, they can communicate with each other, different Communication between the VLAN site needs to be done by a router. The advantage of this VLAN mode is that it is simple, easy to implement, broadcast from one port, directly sent to other ports within the VLAN, which is also easy to monitor. Its disadvantage is that the degree of automation is low, flexible is not good. For example, if you support more than one VLAN on a given end; when you move from one end day to another, if the new port is not the same VLAN, the user must have the site. Return the network address configuration.
2. MAC address-based VLAN
The VLAN of this approach requires a switch to track the MAC address of the site and the switch port. When the new site is included, it will be divided into a VLAN as needed. Regardless of how the site moves in the network, the user does not need to reconfigure the network address because the MAC address remains unchanged. However, all users must be explicitly assigned to a VLAN, and after this initialization work is completed, the user's automatic tracking is possible. In a large network, it is very cumbersome of network managers to divide each user-one. 3. Network-based VLAN
It is a network layer's business attribute from moving a VLAN, which is divided into corresponding VLANs using different route protocols. IP subnet 1 is the first VLAN, IP subnet 2 2nd VLAN, IPX subnet 1 is the third VLAN ... With this class. By checking all broadcast and multipoint broadcast frames, the switch can automatically generate a VLAN.
The VLAN constructed in this way can belong to the same VLAN on a different LAN network segment, and the site on the same physical port can also be divided into different VLANs, thereby ensuring that the user is completely free to increase. Move and modify the operation. This approach to network protocols and network addresses applied based on the network is very effective for network managers who want to organize users for specific applications and services. It reduces manual participation in configuring VLAN, making VLANs more flexible, more easily for automation than MAC-based VLANs.
Third, VLAN application value
1. Added the flexibility of network connection
Network administrators can group on the network, without having to group in a geographic location. VLANs can reduce management fees for mobile or change workstation, especially those who have frequent changes, which use VLANs, and this part of management costs are greatly reduced.
2. Controlling broadcast storm on the network
As the network transitions to the switching structure, people lose the router to provide firewall function. In this way, the broadcast storm will be sent to every exchange port, which is often the entire network is a broadcast domain. The advantage of using the swap network is to provide low delay and high throughput. The disadvantage is to increase the broadcast storm of the entire switching network. VLANs can provide a mechanism for building a firewall to prevent excessive broadcast storms of the switching network. With VLANs, you can assign a switching port or user to a particular VLAN group, which can be in a switched network or across multiple switches in a VLAN not sent to the VLAN. . Similarly, adjacent ports do not receive broadcast storms generated by other VLANs. In this way, the broadcast traffic can be reduced, and the bandwidth to the user can be reduced to reduce the generation of the broadcast storm.
3. Increase network security
Some confidential, critical data are often transmitted on the LAN. Confidential data should provide security means such as access control. A effective and easy-to-realization method is to segment a network into several different broadcast groups, and the network administrator limits the number of users in the VLAN, and it is forbidden to access the applications in the VLAN without permission. Exchange ports can group based on application types and access privileges, and restricted applications and resources are typically placed in security VLANs.
4. Added centralized management control
By centralized VLAN management procedures, network administrators can determine the VLAN group, assign specific users and switch ports to these VLAN groups, set security levels, limit the size of the broadcast domain, and share network traffic through redundant link loads, crossover switches Configure VLAN communication, monitor traffic flow and network bandwidth used by VLAN. These capabilities effectively improve the controllability, flexibility, and monitoring of network management procedures, and reduce management costs.
Fourth, VLAN in large, medium-sized diskless network
Large-type diskless network refers to the diskless network of 200 or more workstations. For such free networks, a single server is definitely unqualified. At this time, multiple servers need to be equipped with multiple servers. If each server has each server each workstation Form a physical separated subnet, each subnet independently form a plurality of small and medium-sized diskless networks independently. At this time, it is not a large-scale diskless network, which can only be called multiple diskless networks. The structure of diskless network mode is not applicable in many cases, such as only one line of the Internet, and data between each subnet requires communication. At this point, multiple server clusters can be used, that is, all servers and workstations are all connected in a network. We can complete multi-server clusters through a variety of technical means, but resolve problems such as conflicts, sharing conflicts, and network storms. , More complicated relative to a single server network. With VLAN-based switches to make the network node, better solve this problem, planning the ports used by each subnet, each server is independent, even if a machine name is used, and there is no conflict, and each subnet The work can be accessed each other. In some-------------------------- - - - - page-Video card, the multi-network card, the transfer workstation data is configured in the server, and a switch with a VLAN function can be used as a network. The center node is planned to plan the ports used in each subnet, while the work of each subnet can be accessed each other.
VLAN technology is still in development, and some technical standards have been developed and improved. As its price continues to decrease, the function is constantly enhanced, and it is expected to have great development in the next few years.
