What is a server?
The server is a high-performance computer, as a node, storage, and 80% of data on the network, and information, and therefore also referred to as the soul of the network. Be a metaphor of an image: the server is like a post office's switch, while fixed or moving network terminals such as microcomputer, notebook, PDA, mobile phone, such as scattered a telephone in a family, a variety of office, public places, etc. We communicate with the daily life of the outside world, communication, communication, must be switched to reach the target phone; the same, network terminal equipment such as homes, microcomputer in the enterprise, access to external communication, entertainment, etc. You must pass the server, so it can also be said that the server is in "organization" and "leaders".
The composition of the server is substantially similar to the microcomputer, with processors, hard disks, memory, system bus, etc., which are specifically developed for specific network applications, so servers and microcomputers in processing capacity, stability, reliability, security, scalable There is a lot of differences in sex, manageability. In particular, with the advancement of information technology, the role of the network is increasing, and the requirements of data processing capabilities, safety, etc. of the information system are also increasing, and if you are stealing in hackers during e-commerce Password, loss of key business data; if you are not working properly on the ATM, you should consider behind the scenes of these equipment systems --- servers, rather than complaining about the quality of staff and other objective conditions.
Server technology EMP technology
At present, the technical hotspots of the server include: IRISC and CISC technology, processor technology, multiprocessor technology (AMP technology, SMP technology, MPP technology, COMA technology, cluster technology and NUMA technology), SCSI interface technology, intelligent I / O technology , Fault-tolerant technology, disk array technology, hot-swap technology, and two-machine thermal backup.
The server assumes the task of transmitting and processing a large amount of data in the network, with high scalability, high reliability, high availability, and high management. The IA-64 system will drive the improvement of the technical characteristics of the server, such as high performance CPU, multiprocessor technology, bus and memory technology, fault tolerant technology, cluster technology, hardware management interface, equalization server platform technology, etc.
EMP (Emergency Management Port)
EMP (EMERGENCY Management Port) technology is also a remote management technology that allows the client to implement an off-site operation on the client through telephone line or cable, such as shutting down the operating system, starting power, turning off the power, capturing The server screen, configuring the server BIOS, etc. is a technical means for realizing fast service and saving maintenance costs. Two techniques for applying ISC and EMP can implement remote monitoring management of the server.
RAID redundant disk array technology
At present, the technical hotspots of the server include: IRISC and CISC technology, processor technology, multiprocessor technology (AMP technology, SMP technology, MPP technology, COMA technology, cluster technology and NUMA technology), SCSI interface technology, intelligent I / O technology , Fault-tolerant technology, disk array technology, hot-swap technology, and two-machine thermal backup.
The server assumes the task of transmitting and processing a large amount of data in the network, with high scalability, high reliability, high availability, and high management. The IA-64 system will drive the improvement of the technical characteristics of the server, such as high performance CPU, multiprocessor technology, bus and memory technology, fault tolerant technology, cluster technology, hardware management interface, equalization server platform technology, etc.
RAID (Redundant Array Of Independent Disks) Redundant Disk Array
technology
RAID technology is an industrial standard, and each manufacturer does not have the same definition of the RAID level. There are currently four, RAID 0, RAID 1, RAID 0 1, and RAID 5, which are currently widely recognized in the industry.
RAID 0 is a non-data redundant storage spatial strip, which has a low cost, extremely high read and write performance, and high storage space utilization rate. It is very strict for speed requirements for Video / Audio signal storage, temporary files. Special application. However, since there is no data redundancy, the security is greatly reduced, and the damage of any hard disk constituting the array will bring catastrophic data loss. Therefore, if more than 4 pieces of hard drives are configured in RAID 0, it is unmoxive for general applications. RAID 1 is completely mirrored in two hard disk data, safe, simple technology, convenient management, and good read and write performance. But it cannot be expanded (single hard disk capacity), the data space is wasted, and it is strictly called "array".
RAID 0 1 combines RAID 0 and RAID 1 features, and stand-alone disk is configured to RAID 0, two sets of complete RAID 0 mirror images. Its read and write performance is excellent, high security, but the cost of building arrays is large, and the data space utilization rate is low, and it is not possible to be economically efficient.
Load balancing technology overview
At present, whether in the enterprise network, the park is still on the web, such as the Internet, the development of traffic is exceeding the most optimistic estimate in the past, the Internet is booming, the new application is endless, even in accordance with the network at the time, it is also very I will feel it. In particular, the core part of each network, the data flow and calculation intensity have caused a single device to not bear, and how to achieve reasonable traffic distribution between multiple network devices that complete the same function, so that there is no one When the equipment is too busy, the other equipment has not fully utilized the ability to deal with handling capacity, it has become a problem, and the load balancing mechanism has also emerged.
Load balancing is built on existing network structures, which provides a cheap and effective way to extend server bandwidth and increase throughput, enhance network data processing capabilities, improve network flexibility and availability. It mainly completes the following tasks: solve the network congestion problem, service is available, realize geographical location; provide users with better access quality; improve server response speed; improve server and other resources; avoid network key parts Single point failure.
For a network of load balancing applications, you can start from different levels of the network. To see the specific analysis of the network bottleneck, it is generally no more approval from the transmission link aggregation, using higher-level network switching technology and setup servers. Three angles of cluster strategy.
■ Transport link aggregation
In order to support the growing high-band wide application, more and more PCs use more fast links into the network. The distribution of traffic in the network is unbalanced, the core is high, the edge is low, the key department is high, and the general department is low. With the significant increase in computer processing capabilities, people have higher requirements for the processing capabilities of multi-work group LAN. When the interior of the enterprise is increasing (eg web access, document transmission, and internal network connection), the data interface of the local area network core part will generate bottleneck problems, and the bottleneck extended the response time of the customer application request. And the local area network has a dispersion characteristic, the network itself does not protect the server, a unintentional action (plug in a pin kicked off) allows the server to disconnect from the network.
Typically, the countermeasures to solve the bottleneck problem are to improve the capacity of the server link, which exceeds the current demand. For example, you can upgrade from fast Ethernet to Gigabit Ethernet. For large companies, upgrade technology is a long-term, promising solution. However, for many companies, when the demand has not taken a lot of money and time to upgrade, the use of upgrade techniques is small. In this case, link aggregation technology provides a cost-effective solution to eliminate bottlenecks and unsafe factors on the transport link.
Link aggregation techniques, combining multiple lines of transport capacity into a single logical connection. When the original line meets the needs, and the upgrade of a single line is too expensive or difficult to implement, it is necessary to use multi-line solutions. There are currently four link aggregation technologies that can "bundle" multiple lines. The synchronous IMUX system works in the bit layer of T1 / E1, and utilizes multiple synchronized DS1 channel transmission data to achieve load balancing. IMA is another multi-line reverse multiplexing technique, working in the cell level, can run on a platform using the ATM router. Using the router to implement multi-line is a popular link aggregation technology, the router can assign the packet to each parallel link according to the cache size of the known destination address, or the method of cyclic allocation can also be used. Line distribution packet. Multiple link PPP, also known as MP or MLP, is a router load balancing technology applied to the PPP package data link. The MP can break the large PPP packet into a small data segment, and then distribute it to a plurality of lines, and can dynamically allocate dial lines based on the current link utilization. Doing so although the speed is slow, because the data packet segmentation and additional buffers increase the delay, it can run well on the low speed line.
The link aggregation system adds the complexity of the network, but also improves the network reliability, enabling people to use redundant routes on the lines of the critical LAN segments such as servers. For IP systems, VRRP can be considered (virtual routing redundant protocol). VRRP can generate a virtual default gateway address. When the primary router cannot be turned on, the standby router will use this address to continue the LAN communication. In summary, when the performance of the main line is necessary, link aggregation techniques can be employed when the upgrade of single lines is not feasible.
Higher layer exchange
Large-scale networks are generally composed of a large number of dedicated technical equipment, such as including firewalls, routers, Layer 2/3 floors, load balancing equipment, buffer servers, and web servers. How to organically combine these technical devices, is a key issue that directly affects network performance. Now many switches provide a fourth floors to map an external IP address to a plurality of internal IP addresses, and dynamically use one of the internal addresses for each TCP connection request to achieve load balancing. Some protocols support functionality related to load balancing, such as the redirectivity in the HTTP protocol.
Web Content Switching Technology, that is, URL exchange or seven-layer exchange technology, provides a high-level control method for access traffic. Web Content Switching Technology Checks all HTTP headers, performs load balancing decisions according to information within the header, and can determine how to provide services such as personal homepage and image data based on this information. It is not controlled according to TCP port numbers, so it will not cause retention of access traffic. If the web server has been optimized for image service, SSL conversation, database transaction services, then flow control, using this level will increase the performance of the network. Currently, products and programs of seventh floors exchange technology include the ISWITCH, Switch, Cisco CDN (Content Switching Network System) of the Dawn Network.
Server cluster solution
In some cases, for example, the internal staff and external customers are available at the same time, and the company should connect the internal staff's service request to a slower server to provide more resources for external customers. You can use it. Web content exchange technology. Web host access control devices can also use this technique to reduce hardware costs because it can easily transfer user traffic accessed to the same web server. If the user accesses increase to a certain extent, these traffic can also be transferred to a dedicated web server device, although this dedicated device has a higher cost, but since the same web content exchange technology is used to control traffic, the network The structural framework is no longer changed.
However, the number of standards and rules that use web content exchange techniques is limited, and the flexibility of the standards and rules adopted is limited. In addition, the depth of the load balancing device can monitor the HTTP header is also a factor limiting the content exchange capability. If the information you want is in the field that the load balancing device cannot monitor, the role of the content is unable to play. Moreover, the content exchange is also limited by the number of TCP connections that can be simultaneously opened and the establishment of the TCP connection and the disconnection ratio. In addition, Web content exchange technologies also take up a large number of system resources (including memory occupancy and processor occupation). Tests for web content exchange techniques show that the throughput of the WEB content is very laborious, sometimes it can only get small performance improvements. Therefore, network administrators must carefully consider the issues of investment and returns. ■ Server clusters with balance strategies
Today, the server must have the ability to provide a large number of concurrent access services, and its processing capabilities and I / O capabilities have become bottlenecks that provide services. If the increase in the customer leads to traffic over the server, then its result is inevitable - downtime. Obviously, the limited performance of single servers cannot solve this problem, and one ordinary server can only reach hundreds of thousands of requests per second, and will not handle millions of even more than one second. Request. But if you can make 10 such servers into a system, all requests are allocated to all servers via software technology, then this system completely has millions of or more requests per second. This is the initial basic design idea for realizing load balancing using server clusters.
Early server clusters are usually backed up in the optical fiber mirror card. Let service operators headache is a key server or more applications, and the data traffic is generally not too low, and the service operators have spent two servers, but they often only get a server performance. The new solution is shown in the picture, translates the different IP addresses of multiple server network cards into a VIP (Virtual IP) address through LSANT (LOAD SHARING NETWORK ADDRESS TRANSFER), so that each server is always in operation. It turns out that the work that needs to be done with a small machine is completed by multiple PC servers. This elastic solution is quite obvious to investment protection - avoiding the huge equipment investment brought by small machine rigid upgrades, and avoiding Repeated investment of personnel training. At the same time, the service operator can adjust the number of servers at any time according to the needs of the business.
Network load balancing improves the availability and scalability of Internet server programs on a web server, FTP server, and other critical task servers. A single computer can provide a limited level of server reliability and scalability. However, by connecting the host of two or more advanced servers into a cluster, network load balancing can provide the reliability and performance required for critical task servers.
In order to establish a high-load web site, multiple servers must be used. The combination of proxy servers and web servers mentioned above, or two web servers collaborate, and in the structure of multiple servers, in the structure of these multi-server, each server is different, It belongs to an asymmetric architecture. The role of each server in an asymmetrical server structure is different, such as a server for providing a static web page, while the other is used to provide a dynamic web, and the like. This makes it necessary to consider the relationship between different servers when designing the web page. Once the relationship between the servers will make some web pages, it is not conducive to maintenance, and scalability is also poor.
The network design structure that can carry load balancing is a symmetrical structure, and each server has an equivalent position in a symmetrical structure, and the service can be provided separately without the auxiliary of other servers. Then, the external transmission request can then be uniformly assigned to each server in the symmetric structure evenly, and the server receiving the connection request is independently responding to the customer's request. In this configuration, the Web server is not difficult because the establishment of a Web server is not difficult, so load balancing technology has become a key technique for establishing a high load web site. In summary, load balancing is a strategy that allows multiple servers or multiple links to jointly assume some heavy calculations or I / O tasks to eliminate network bottlenecks at lower cost, improve network flexibility and reliability.
High-end server technology
Server performance indicators are represented by system response speed and job throughput. The response speed refers to the time that the user gives the task from the input information to the server. The job throughput is the amount of task completed throughout the server in unit time. Assuming that the user is not interpretially input request, in the case of abundant system resources, the throughput of a single user is inversely proportional, that is, the shorter the response time, the greater the throughput. In order to shorten the response time of a certain user or service, you can assign more resources. Performance adjustment is based on application requirements and server to run environment and status, changing system resources allocated by each user and service program, and give full play to system capabilities, with as little resources to meet the user's requirements, and reach the purpose of more user services.
Technical goal
The high scalability, high availability, easy management, and high reliability required by the server are not only the technical goals pursued by the manufacturer, but also the needs of users.
The scalability is specifically manifested in two aspects: First, there is a surplus chassis available space, and the second is a plenty of I / O bandwidth. As the processor calculation speed increases and the increase in the number of parallel processors, the bottleneck of server performance will be attributed to PCI and their affiliates. High scalability is that users can increase the relevant components at any time as needed to meet the system operation requirements, and protect investment.
The availability is the time ratio of the device as a measurement indicator, for example, 99.9% availability indicates that an 8-hour time device does not function properly, and 99.99% availability indicates that a 5-minute time device does not work normally. Component redundancy is the basic method of improving usability, usually adding redundant configurations to those components (such as power, hard drives, fans, fans, and PCI cards) that have hazards to the system, and design convenient replacement structures (such as hot plugging) ), Thereby ensuring that these devices do not affect the normal operation of the system even if the fault occurs.
Manageability is designed to use specific technologies and products to increase system reliability, reduce system purchase, use, deployment, and support costs. The most significant role is reflected in reducing the work of maintenance staff and avoiding the loss of system shutdown. The management performance of the server directly affects the ease of use of the server. Manageability is the largest proportion in TCO's various costs. Studies have shown that the deployment and support cost of the system far exceeds the cost of the first time, and paying the management and support personnel is the highest share. In addition, the reduction in work efficiency, the financial loss brought about the loss of business opportunities and the decline in business income can not be ignored. Therefore, the management of the system is both an urgent requirement of the IT department and a very critical role in business efficiency. Manageable products and tools can achieve the purpose of simplifying system management by providing information within the system. Remote management through the network, technical support staff can solve problems on their desktops, do not have to go to the fault site. The system components can automatically monitor their work state. If the failure can be found, it can make a warning at any time, reminding the maintenance personnel to take immediate steps to protect enterprise data assets, and the operation of the faulty component is also very simple and convenient.
Speaking of reliability, simply means that the server must run stable, which is low in down. The key is to cooperate with the hardware device. If the resource to be processed is controlled on the CPU and the operating system, it will avoid the system unable to run due to an error in an error, and the server downtime will be great. Reduce, and this is precisely one of the advantages of UNIX / Linux system. The interruption of daily maintenance work is: host upgrade, hardware maintenance, or installation, operating system upgrade, application / file upgrade, or maintenance, file reorganization, full system backup, etc. Accidental disasters include hard disk damage, system failure, software failure, user error, power supply, human damage, and natural disasters. SMP
SMP (Symmetrical Multi-Processor) is a symmetrical multiprocessor. In a symmetrical structure, the status of each processor in the machine is the same, and they are connected together to share a memory. There is an operating system in the memory, each computer can run this operating system, can respond to the requirements of the external device, ie the position of each memory is equal, symmetrical. The processor of such models in the domestic market is generally 4 or 8, and there are a small number of 16 processors. However, in general, the Machine scalability of the SMP structure is poor, it is difficult to do more than 100 multi-processors, and conventional generally 8 to 16, but this is already enough for most users. The advantage of such a machine is that its use method and the different differences of the microcomputer or workstation. The programming changes is relatively small. If the program written by the microcomputer workstation is to be used to use it to the SMP machine, it is relatively easy. . The model availability of the SMP structure is relatively poor. Because 4 or 8 processors share an operating system and a memory, once the operating system has problems, the entire machine is completely paralyzed. And because this machine is scalable, users' investment is not easy to protect users. However, this type of model technology is mature, and the corresponding software is also more, so the parallel machine launched in the domestic market is now this kind.
Cluster technology
It is popular that the cluster is such a technique: it connects to at least two systems together, so that the two servers can work or look like a machine. Using a cluster system is usually in order to improve system stability and data processing capabilities and service capabilities of the network center. Since the 1980s, various forms of cluster technology have emerged. Because the cluster can provide high availability and scalability, it rapidly becomes the pillars of enterprises and ISP calculations.
Common cluster technology
Server mirroring technology
Server mirroring technology is a mirroring of two servers on both servers on the same LAN through software or other special network devices (such as mirror cards). Among them, a server is designated as the primary server, and the other is from the server. Customers can only read and write on the mirrored volume on the primary server, that is, only the primary server provides services to the user through the network, locked from the server, locked to prevent access of the data. The master / slave servers monitors each other's operating state through heartbeat monitoring lines. When the primary server is downtime, the main server will take over the primary server in a short period of time.
Server mirroring techniques are low cost, improve system availability, ensuring that the system is still available in the case of a server downtime, but this technique is limited to clusters of two servers, and the system does not have scalability.
2. Application error takeover cluster technology
Error Tube Cluster technology is to connect two or more servers built in the same network through cluster technology, each server in the cluster node runs different applications, with its own broadcast address, providing service to front-end users, At the same time, each server monitors other servers' running status, providing a hot backup role for the specified server. When a node is downtime, the server specified in the cluster system will take over the data and applications of the faulty machine in a short period of time, and continue to serve the front-end users.
/ Error Tube Cluster technology typically requires external storage devices - disk array cabinets, two or more servers are connected to disk array with disk array via SCSI cable or fiber, and data is stored on disk array. In this cluster system, two nodes are typically backed up, rather than several servers simultaneously, and the nodes in the cluster system via the serial port, shared disk partitions or internal networks to monitor each other's heartbeat.
Error Take Overcrow cluster technology is often used in a cluster of database servers, Mail servers, and the like. This cluster technology has increased peripheral costs due to shared storage devices. It can realize the cluster of 32 machines, greatly improves the availability and scalability of the system.
3. Fault tolerant cluster technology
A typical application of fault-tolerant cluster technology is a fault tolerant machine, in a fault-tolerant machine, each component has a redundant design. Each node of the cluster system in the fault-tolerant cluster technology is closely linked to other nodes, they often need to share important subsystems such as memory, hard drives, CPUs, and I / O, and each node in the fault-tolerant cluster system has become a common image. A separate system, and all nodes are part of this image system. In a fault-tolerant cluster system, various switches between various applications can be done well without switching time.
The implementation of fault-tolerant cluster technology often requires special hardware and software design, so the cost is high, but the fault tolerance system maximizes the availability of the system, the best choice for financial, financial and security departments.
At present, the availability of the system is used in a wide range of applications that the application error takeover technology, that is, the dual-machine we usually use the cluster technology of the SCSI cable sharing disk array, this technology is currently being trained by various cluster software vendors and operating systems. Software vendors further expanded, forming a market-colored cluster system.
IA architecture high performance cluster system technology
With the high-performance server cluster system of the latest 4, 8-way IA server architecture, the leading VI (Visual Interface) technology effectively eliminates the bottleneck communication between nodes in the system; at the same time, the load balancing technology used by the system makes User equipment can be fully utilized and reached 4 9 reliability, with extremely high product cost performance, providing a powerful database server platform for Internet applications.
1. System Overview
The data center area has always been the world of high-end RISC servers. In many years, people have only selected small machines such as AS400, E10000, HP9000, etc., the price is expensive, and the maintenance cost is high. The IA architecture server is low, easy to use and maintain, and the supercomputer can be constructed by Cluster technology. Its superb processing capability can replace the expensive medium and large machines, and open up new directions for the industry's high-end applications.
For large users in growth, the company operated in the data center or data warehouse is very amazing, and these data is very important to the role of large users. The data under the development of several years is a valuable wealth. By analyzing the data of these Hao Ruohai, the operators can get an intuitive business chart, curve, can provide powerful for future development of large users. policy support. However, because such data continues to expand over time, it has brought huge pressure to the IT system management of large users.
What kind of server does the user need to meet the needs of current and future development?
First of all, there must be super-computing power, you can withstand a long time, and the large number of users can be accessed.
Second, the high availability of the server system is extremely important for users. If the system fails, the interruption of the service, or the loss of important information, it will cause the user to save the loss. Therefore, users must consider highly available system scenarios when selecting a server system.
Third, with the continuous accumulation of data, the query and statistics of the data will make the system becomes slower and slower, and the update of hardware devices can be said to be a developed large user must be indispensable.
Using the high-performance server cluster system [1] of the latest 4, 8-way IA server architecture, the leading VI (Visual Interface) technology effectively eliminates the bottleneck communication between the nodes in the system; at the same time, the load used by the system Balancing technology allows the user equipment to be fully utilized, and reaches 4 9 reliability, which has a very high product cost price. Since the launch of users since 1999, users have provided a strong database server for domestic users. platform. 2. System principle
The high-performance server cluster system is a cluster of 2 nodes or 4 nodes, up to 32 CPUs, and memory maximum supports 32GB. 4 nodes constitute a working unit, up to 16 working units.
Each node is an IA server, which supports 4 or 8 Pentium III Xeon CPU parallel processing. Insert a high-speed exchange card for a Gigabit NIC or VI structure on each server, connect to a high-speed switch (which can be a Gigabit Ethernet card switch or some special high-speed switches, such as high-speed switches of VI structures) as servers Data exchange, called SAN (Server Area Network) switches.
Each server is inserted with a 100 megaphone or Gigabit Ethernet card, which is connected to a switch or hub for a local area network to provide a connection service for the client's access.
Four servers share a Fiber Channel disk array cabinet. Each server has two Fiber Channel cards, which are connected to the hubs of the two Fiber Channels, each connected to two controllers of the Fiber Channel disk array cabinet. For Fiber Channel Disk Array Cabinets, as long as one controller can work normally, the entire array cabinet will work properly, so this configuration is redundant to prevent single point failure. For the most important data stored in the disk array cabinet, you can also take the cluster system to ensure data security, and the Fiber Channel allows the cluster system to 10 km from the disk array.
Each server has a hard disk area for mounting the management section of the boot system of this unit and database system. User data is stored in a shared disk array cabinet.
In the LAN, there is a client as a management console. It is installed on the management console to manage parallel databases, which can monitor the database instance on the four nodes, start, stop, monitor the operation performance. Waiting for functions.
In addition, there is a network management system, the management console, the management console, the UPS management console, etc. of the disk cabinet, and the unified management of the cluster system is implemented. Some management functions require only the TCP / IP protocol to be installed, and some need to install the SNMP protocol to work properly.
In addition to excellent performance indicators, a good cluster system requires support for operating systems and databases, our current cluster system supports Windows NT 4.0 and Windows 2000 operating systems [2], support Oracle and DB2 in databases . It does not run on a single machine, and its performance can only be displayed when multiple nodes work, and the system can truly load balancers.
2.1 Two-Node Cluster System
From the configuration of the program, the user can do different options as needed, you can use two high-end servers to implement a virtual host. At this time, high-speed switching devices with VI structures are more superior, and there is no need to use VI switches to directly connect high-speed data exchange between servers with high-speed switching cards on each server. If you use a Gigabit Ethernet card, you also need a Gigabit switch with a higher cost.
Figure 1 Two-node system connection diagram
2.2 Four Node Cluster System
The design goal of the cluster system scheme is to support the 4 node cluster, and the corresponding storage device and switching device constitute a work unit in four nodes. Each work unit consists of four separate 4 or 8-way servers works as a virtual fault tolerant host, and four servers share a Fiber Channel disk array cabinet. Each server has two Fiber Channel cards, which are connected to the hubs of the two Fiber Channels, each connected to two controllers of the Fiber Channel disk array cabinet. For Fiber Channel Disk Array Cabinets, as long as one controller can work normally, the entire array cabinet will work properly, so this configuration is redundant to prevent single point failure. Figure 2 Four node system connection diagram
3. System features
High-performance server cluster system solutions have enabled 4 nodes of clusters, exceeding the limits of traditional Cluster two knots. If each node is used in the cluster, the 4 node's cluster system can support 32 processors, which can be competent with traditional RISC miniatles and medium-sized machines.
3.1 Load Balancing:
Load balancing concept: Multiple servers are symmetrical methods, each server has an equivalent position, can provide services separately without providing other servers. Then, through some kind of load sharing, the external transmission request is evenly assigned to a certain server in the symmetrical structure, and the requested server is independently responding to the client's request.
1. Puzzle access to a large number of users is parallel to multiple node machines, shorten the time of the user waiting for a response, and the processing capability of the system is improved, and more user-friendly access can be accepted. When the client applies for a database connection, which node is automatically allocated by the database to which node is connected to implement the load balancing of this method does not need to modify existing applications.
2. Sharing a single user heavy load to multiple node machines to do parallel processing, each node machine has multiple CPUs, also parallel processing, after the processing is completed, returns the result, return to the user. On a large number of users, a large user query is over multiple nodes, and then the result is merged, then the result is given to the user, and the system processing capacity is greatly improved. A load balancing of this way requires modifying existing applications, but only needs to modify the SQL query statement.
3.2 High availability:
High availability means that the maximized server is turned on, which is the minimized server-planned downtime, so that critical data can be protected to improve production efficiency. 99% of high availability means that downtime outside the year is 5,000 minutes, while 99.99% means 50 minutes.
High Performance Server Cluster System Solution is a True Cluster system. When a node machine fails, other node machines still work normally, and the client's access will not be interrupted.
Figure 3 High availability schematic
A large number of redundant devices in the high-performance server cluster system, such as multiple servers, multiple UPS, multiple switches, disk redundancy (RAID), and independent two sets of CPU / memory subsystems, even configured in storage devices. Two sets of power systems for disk cabinet buffer buffer, effectively shielding the risk of single point failure, so that the reliability of the entire system has reached 99.99%.
The advantage of this program is that in the normal operation, four server parallel processing tasks, no resource idle, and parallel processing, greatly improve the overall system processing power; and when there is one or several failures, other servers have to execute themselves Outside the task, it also takes over the work of other nodes, so the load is increased, but the entire cluster system is still working properly, and the client's access will not be interrupted, because this is specifically designed for users who cannot tolerant shutdown, fundamentally eliminating Stop factor. A failed server simply removes it from the array. When this server is repaired, you can re-add to the server array without manual intervention. This is the meaning of our parallelism and never stop. 3.3 Centralized management performance
In the LAN, there is a client as a management console. It is installed on the management console to manage parallel databases, which can monitor the database instance on the four nodes, start, stop, monitor the operation performance. Waiting for functions. In addition, there is a network management system, the management console, the UPS management console, etc., and realize unified management of the cluster system. Some management functions require only the TCP / IP protocol to be installed, and some need to install the SNMP protocol to work properly.
Applications for management
Figure 4 Concentrated management schematic
3.4 Scalable performance.
The high-performance server cluster system is a cluster of 2 nodes or 4 nodes, up to 16 or 32 CPUs, and the memory maximum supports 32GB. 2 nodes or 4 nodes constitute a work unit.
First use a work unit to perform data processing, with the development of the business, when the discovery ability is insufficient, the second work unit can be reacted, and the processing performance can be improved. This reflects that the Cluster system is well scalable.
Can the four-node server parallel processing power to reach four times that of single server processing?
No. The two-node server parallel processing capability can basically reach twice the single server processing power, but the four-node server parallel processing capability can only reach three times more than the single server processing power. Since the number of resources used for the CLUSTER system management is increased, the total processing power of the system is increased.
3.5 VI high-speed transmission performance.
VI is the abbreviation of Visual Interface, which is directly translated into "virtual interface", is a communication technology between servers and servers. Strictly speaking, VI is an industrial standard that supports at least 100 companies and institutions, including internationally renowned manufacturers such as Intel, Compaq, IBM, HP, Microsoft. It does not mean some hardware or software equipment, but a special network communication protocol, where we are called VI devices, which is in line with this communication protocol. Sometimes this device is also known as the VIA (Visual Interface Architecture) device, the "Virtual Interface Architecture" device.
The VI device is characterized by high transmission rate - greater than 1.25Gb / s, even higher than the current Gigabit switch network device (the transmission rate of Gigabit switch is 1Gb / s), and because it uses a special communication protocol, Not a traditional OSI7 layer transfer protocol, which is much higher than the normal Gigabit switching device on the transmission speed of the data. The high-performance server cluster system uses the latest communication technology to exchange data between the servers, that is, the VI structure is used to implement high-speed data exchange between servers.
Traditional data transmission (UDP / IP) adopts VI technology
Figure 5 Solving the bottleneck of software transmission
Of course, you can also use Gigabit NIC and Gigabit Switch to implement data exchange between servers, but the most important difference is that the data exchange rate between the servers is much slower than the VI structure. Gigabit network card and Gigabit switch implementation data exchange must pass TCP / IP's seven-layer protocol, and the VI structure of the communication card can bypass the TCP / IP's seven-layer protocol, implement the application directly accessing the VI structure communication card, ie direct Data exchange with hardware devices, very high, 1.25Gb / s, greatly reduce CPU resource occupation. In the Cluster system, the speed of data exchange between the servers directly affects the overall system performance. In the case of the number of nodes, the more significant, even becomes the bottle diameter of the system performance, because the process is occupied between multiple nodes Most of the system resources. A communication card currently supporting a VI structure has Oracle and DB2 for applications that directly perform data exchange in application layers.
When implementing two nodes, the communication card of the VI structure is more superior, because the communication cards of the two VI structures can be directly connected, and there is no need to switch, good performance, cheap. And if you use Gigabit Ethernet, you must use a Gigabit switch.
Internet era, domestic users have more and more demand for high-end applications of servers, using the continuous availability of the IA server architecture, online expansion, remote management, etc. A series of excellent features such as many users have changed many users. may.
Ia architecture server - true pseudo recognition
Currently, the server can be divided into the IA architecture server of the RISC (Reprimanting Instruction Set Calculation) architecture server and CISC (complex instruction set calculation) in accordance with its architecture. The IA architecture server uses an open system and implements industrial standardization technology. Since 1996 Intel launched the Pentium Pro processor, the IA architecture servers have greatly developed. Now, the server with Pentium III Xeon can support 8 channels and have high performance. Company IBM, Dell, SGI, HP, Compaq, etc., high-end server areas have turned to use IA architectures. Since the server of the IA architecture uses an open architecture, it has been favored by domestic server manufacturers, and the manufacturer of major IA architecture servers in China has a wave, Lenovo, Space Journal, Tongfang, etc.
Pseudo IA architecture server
Throughout the domestic server market: In 2000, domestic brands have created amazing growth results, and the market share has accounted for 41%. However, entering 2001, the global economic development speed slowdown, foreign server manufacturers increased the investment in the Chinese market, and domestic large IT companies have also added, making the new dawn domestic server manufacturers face severe challenges. In front of cruel competition, local brands have to take price war. In low-priced competition, it can be profitable and not false, which is closely related to the technical strength and production scale of the company. Taking Beijing Aerospace Terrace as an example, the reason why it is possible to launch a low-cost server product, and most importantly based on its own strength. At present, the only domestic manufacturer of production server products is actually only several aerospace situations, and those server manufacturers that have no weakness or weak server do not have corresponding technical strength and industrial production scale, in order to get a place in low-cost competition, have to play A text game that is "stealing the column", replacing the Intel's IA architecture chipset with high-end PC, other non-IA standard chips, pretending to install the IA architecture, and then push the market with the so-called "promotional price". The pseudo IA architecture server in the market has attracted Intel and some domestic manufacturers, Intel decided to join hands with server manufacturers around the world to organize large-scale promotion activities, and cooperate with Beijing Aerospace LAN, order in mid-November. Activities will be carried out in 13 cities across the country.
What is the true IA architecture server
The server of the Intel processor is usually referred to as an IA (Intel Architecture) architecture server. Since the architecture server adopts an open system, industrial standardization technology and support from a large number of hardware and software suppliers at home and abroad, based on high-volume production ratio, in a global scale, Especially in our extensive application. The real server must meet the following conditions: 1. Adopt Intel CPU;
2. Using Intel's server motherboard;
3. Adopt Intel's RAID technology;
4. Use Intel certified hardware and software;
5. Free service management feature ism.
Intel is the formulator of IA architecture standard, which has more than 80% of the CPU market, and the IA architecture has become a criteria for users. Intel (China) Co., Ltd. Technical Consulting Manager Luo Bin said: "From the current network development situation, the IA architecture characterized by" small, smart, stable "is more than reliable performance, low price, and has been more A wide range of applications. Many people call a server as NT server or PC server, in fact, I should change the name, called the IA architecture server or standardization, mass server, because this server is not only running on NT. And there is not much relationship with the PC. The real IA architecture is a strict industrial standard, the server as the cornerstone of the entire network, playing a pivotable role, if the PC or other pseudo IA standard motherboard or chip is used as the IA architecture The key components of the server will inevitably affect the quality of the IA architecture, and ultimately damage the interests of users. "
Really, the difference between the server of the pseudo IA architecture
1. Interior PCI total line number
The number of internal buses of the real Ia architecture server is corresponding to the number of CPUs. There are several bus, and the bandwidth can be increased, the system's concurrent throughput is large, and the server is multi-task concurrency. The internal bus structure of the true IA architecture server ensures high efficiency of multi-CPU parallel and maximizes the advantages of multi-processor systems.
The pseudo IA architecture server has only one internal bus, the CPU parallel processing efficiency is low, the internal bus crowding system; when the system load is heavy, the South Bridge chip is easily blocked, causing downtime.
2. RAM
The real IA architecture server uses fault tolerance memory ECC (ERROR Checking and Correhensing). The ECC has strong error detection and correction functions, which uses 4 bitits to detect 8 bit data methods, so that more errors are detected, and 1 bit error in the 8 bit is correct, so that the system Safety and stability are greatly improved, thereby eliminating system instability caused by data noise.
Most of the pseudo IA architecture server uses a low price of ordinary SDRAM memory. The SDRAM memory error detection function is relatively simple, such as using parity, use 1 bit to detect 8-bit data, this error correction ability is weak, only It can probably detect errors.
3. Management architecture
The management chip cured by the real Ia architecture server is cured on the motherboard, implemented by hardware, strong error correction ability, when the system has an error, the server can do not depend on the operating system to correct the number of times.
Pseudo IA Architecture Server is software to implement management functions, low robustness, and subject to stability of the operating system.
4. Hot-swappable
The so-called hot-swappable function is to allow users to remove and replace parts of the system without shutting down the system without shutting down the power, thereby improving the timely recovery capacity, scaiferative and flexibility of the system on the disaster. Wait. The real Ia architecture server uses a dual redundant power supply, which can make the hot-swap replacement of the power and the fan; use the hot-swappable SCSI controller to support the hot swap of the hard disk, and ensure that the server works uninterrupted by RAID function, maximum Improve server stability and security, when there is a power supply and hard disk failure, there is no need to interrupt the server's work, minimize people a number of people. The pseudo IA architecture server usually uses a single source, and the power supply cannot be replaced without shutdown, and the IO chipset does not support hot swapping. Therefore, its continuous working time is greatly dependent on the life and burst failure of the immersion equipment in the system. For some servers applied to key tasks, if the damaged RAID card or Ethernet card cannot be replaced without shutdown, it will greatly increase the system shutdown due to hardware failure.
5. I / O quality
Typically, due to the processing speed of the CPU due to the access speed of the disk, the disk technology is a bottleneck that improves the server I / O capabilities. The real Ia architecture server uses a SCSI architecture that supports RAID (redundant disk arrays): RAID0, RAID1, RAID5. RAID 5 uses a special algorithm that can calculate the storage location of any one of the check blocks. This ensures that any read and write operations on the check block will be balanced in all RAID disks, thereby eliminating the possibility of generating bottlenecks, and avoids the serious failure of the system lost fault tolerance due to the damage of the checkboard. In addition, the SCSI hard drive is high, the speed is high (up to 10,000 rpm), the speed of the SCSI standard (160M bandwidth), the large amount of data is strong, and the life is long.
Pseudo IA Architecture Server generally uses an IDE architecture, where the VIA chipset only supports IDE-RAID0: RAID1 and RAID0. RAID0 is simple and cost is low, but RAID0 has no redundancy or error repair capability, and as long as there is a fault in the disk that makes a RAID, the data of the entire RAID system will be lost, and any remedy cannot be performed. Under RAID 1, the failure of any hard disk does not affect the normal operation of the system, but the space that can be used is only half of the total disk capacity, which increases the cost of the system. In addition, the IDE hard disk speed is low (up to 7200 rpm), the highest peak is 100m (DMA100 standard), only suitable for single process IO, and has a sharp decline in performance; IDE high-speed hard drive is extremely difficult.
From the above analysis, the IA architecture has a large difference between data processing capabilities, reliability, security, extension capabilities, and manageability between the server and the pseudo IA architecture server. Typically, users can use the server if there is an IDE RAID card to simply defective, if there is an authentic Ia architecture server.
Rack server new growth point
As the desktop computer has developed to a certain stage, there will be a laptop, and the tower server has developed to a certain stage, and the ultra-light ultra-thin rack server is also appeared. Since the rack server allows the user to place the maximum number of servers in a limited space, easy to manage and maintain, and provide high-density calculation power, high bandwidth and system redundancy, good scalability, and thus Received by vendors and users, and rapidly replacing traditional tower servers in some areas.
For the Idc (Internet Data Center) of the Tower Server (Internet Data Center), the space occupied by the IDC as the space occupied by the business management and charges, the larger the space occupied, the higher the charge. The server is the most common and maximum rental equipment in IDC. Traditional server covers an area, should not be managed. Enterprises hope to rent a higher technology to a smaller space, thereby reducing the cost of renting. Therefore, the rack server products have always been objects that the rack server products have been in the new telecommunications industry represented by ISP (Internet Service Providers) and IDC. The rise of Internet intensive computing modes such as IDC, XSP makes the rack server hot. In the Internet era, IT has injected infinite vitality, but not only the traditional ISP, but also contributed to ICP (Internet Content Providers), more pregnant ASP. ASP (Application Service Provider) is increasinglycoming the focus of attention, the service business provided by the ASP is also increasingly complex, from e-commerce, online transactions, online operation, online outsourcing services, supplier chain management, inventory management, etc. The ASP service is provided, showing a very beautiful picture in front of people. However, ASP and the rise of e-commerce have brought the requirements and update requirements of hardware equipment and promote the development and improvement of hardware providers. ASPs require a large number of servers to meet customer needs, mainly e-commerce, online transactions, financial analysis, and communications. In this way, rack servers have sprouting in this field.
Just because the rack server has room for space, high-density, facilitating centralized maintenance, management, etc., a large-scale application of a rack server in the traditional industry has quietly opened.
The advantage of the rack server
As a server mode designed for the Internet, the rack server is a server that is designed in accordance with unified standards, with a unified use of cabinets. The width of the rack server is 19 inches and the height is in units of U (1U = 1.75 inches = 44.45 mm), usually 1U, 2U, 3U, 4U, 5U, 7U several standard servers. The size of the cabinet is also used universal industrial standards, usually from 22U to 42U, and the high level of the U.Strate with the height of U., the user can flexibly adjust the height according to the high-altitude of the server, to store the server, hub, disk, disk Array cabinet and other network devices. After the server is placed, all of its I / O lines are taken from the back of the cabinet (all the interfaces of the rack server are also rear), and it is unified to be placed in the cabin of the cabinet, which is generally labeled, which is easy to manage.
At present, Compaq, IBM, HP, Dell, Lenovo, Space Journal and other domestic and foreign computer manufacturers have continuously organized, product lines are not only in the replacement of the tower server, but launched Racklined products of the independent series. Compass, Compass, which is the largest delivery of the global rack server, as an example, Compag Proliant DL300 series (including DL320, DL360, DL380, and DL380 G2, etc.) meets the needs of small databases and applications and virtual hosts; ProLiant DL500 Series Ideally apply to complex Internet applications and medium-sized databases, which are built for critical business applications in the rack environment. The ProLiant DL700 series is ideal for establishing a data warehouse and a size data acquisition for a user. The series adopts the Proliant8 road architecture with breakthrough expansion, thereby providing a new level of performance and almost linear scalability.
The rack server has a lot of innovation in the shape of the shape, but is doubts about the inner management, heat dissipation, and chassis design of the rack server in the user's heart. For this issue, the reporter visited the product manager of Kangbiai Industrial Standard Server Division. Mr. Liu Jingwei, he believes that Compass Proliant DL product line optimizes the design of the space being restricted, it uses the rack mounting design, support the exterior of the cluster storage. Among them, the unique manner, the tool-free chassis design and solving the heat dissipation problem is the three highlights of the Compass ProLiant DL series rack server products. It is particularly worth mentioning that "easy management". The traditional practice of IDC and large and medium-sized enterprises in server maintenance is that the staff must be used in front of the server every day. You can neither do always maintenance work, which is not conducive to timely discovery, and should not allow the staff to contact the equipment frequently. This needs to access all servers through a console in the data center, and maintain the normal operation of the server. Like all Kangbo ProLiant servers, ProLiant DL product line is designed with built-in management characteristics, which simplifies the maintenance of IT infrastructure, allowing users to set their energy to business operations. The first is Kangbo's Insight Manager XE enables users to efficiently manage Compass and non-Kangbo server groups, clusters and network products through a browser. Secondly, Kangbia's Lights-Out version of Remote Insight remote control card and fault prevention software are used with remote management software, and it is possible to provide remote control functions for access to the remote server or data center server via a standard browser. . The Remote Insight Remote Control Card uses a 32-bit PCI slot, which is installed on the server, you need to install Insight Manager, Server Environment Tracking, DISK System Tracking, and Fault Warranty software on the server, so regardless of the user The client is still a normal client in the Internet, as long as Outlook Express or Netscape is installed, you can use the server device hardware by viewing the home page by using Outlook Express or Netscape, you can use the IP address of the browser. It is also possible to start or cut off the server power. Another benefit that the Remote Insight remote control card can save external devices (such as displays, switches, etc.), which can save a relatively large tool device's expenses in large data centers. Not only that, the Flexible hot-swap feature of the Compag Proliant DL series also demonstrates his high availability, such as drives, redundant fans, power supplies, and PCI cards, which provide fast hot-swappable capabilities, thus improving The normal running time and accelerated the maintenance. In addition, the user can quickly browse the server status through the LED indicator, quickly recognize the failed machine orientation, thereby achieving a faster warning.
Toolbox design and easy access to components: Manager does not need to open the chassis with tools, disassemble any parts that can be disassembled. Thereby achieving fast upgrade and component replacement. Cable management schemes are controlled to ensure proper thermal management and achieve maintenance in the rack. Kangbia has developed various rack options for a variety of rack environments, including insert rack slides. In order to help the Best efficient use of the Households of the ProLiant DL product line, Compaq provides a complete industry-leading rack product family, including rack configuration software, innovative options, accessories, and power management products. The rack of 42U, 36U and 22U models have been listed.
In addition, in terms of heat dissipation, the Compass ProLiant DL rack server has built a unique heat dissipation partition system in the industry. The system enables the server to "high fever" three parts ----- CPU, hard disk and I / The O area is completely separated by the exhaust passage, making the internal dispersion of the server. Not only that, the Kangbia rack server also implemented from the top, from the most ideal heat dissipation method from the back, and the meticulous bundled cable design after Kangbia server cabinet and no disk design It also greatly improves the heat dissipation efficiency of the server. In this way, Kangbai's "perfect central air conditioner" built by the server will relieve the worries of the user. Rack server market competition is fierce
At the same time as the competition in the desktop server has reached the degree of white heat, the rack server market has been quite fierce.
Since server manufacturers have to see the seductive prospects in this field of rack servers, products based on rack design are launched. The well-known international brand relies on its own technical force to first entered this market. IBM's NetFinity, HP's netserver, Kangbo's Proliant, Dell's Poweredge is a rack server product, domestic brand, such as Lenovo's Wanquan and Wave Yingxin also launched a rack server, a large number of brands.
According to IDC's statistics, in the Chinese market, the market share of rack-based servers is 14% of the total number of servers, while Compaqi's rack server giant's market share has reached 20%. (In the second quarter of 2001, the global market share of Kangbia was more than 50%). This shows that domestic server companies need to work and unremitting efforts to catch up with foreign manufacturers in the rack server market.
If you want to take a place in this field, you really need to think about it. HP has recently adopted the most commonly used most effective tricks: price cut. Mr. Liu Jingwei, introduced Compaq (China) strategy: 1. Provide a complete full range of rack server products; Second, high-end servers provide bundled integrated solutions; three, low-end market promotion Brand, expand market share; four, continue to introduce new products, such as higher density "blade" servers that meet Data Center requirements.
High-end also value added
In fact, for enterprise-class users, not very concerned about what architecture used, who uses who is strong, how is the function, how to exceed performance, can they only care about whether the server meets their own application needs? Is the system run stable and reliable? Is it easy to maintain? These factors will become increasingly becoming the reason for the user to choose the purchase server. In addition, more and more high-end products are turned from hardware to software, service sector. In order to meet the needs of the market, each server manufacturer also transforms the role, and transforms the transition from a simple server product provider to provide high and reliable value-added service providers to enterprise users.
As an example, Kangbo provides customers with comprehensive lifetime service support. More than 30,000 service professionals in Kangbia will serve 114 countries around the world. Compag Carepaq service provides a comprehensive customer service that provides three levels of services for each ProLiant. The services provided for the ProLiant system are extended from basic warranty to the hardware to install to the startup service, and then to the highest level of key support. If there is an unpredictable problem, Compaq service professionals ensure that the user is solved for the user for users within a few hours. Kangbia also provides users with a relaxed transition to a higher technology. The Corpine Server Folding scheme is a relaxed method for transplanting users in the earlier Intel architecture of Inlet and non-Kangbo servers to improve the performance and reliability improvement of the newcomer ProLiant server.
Kangbia can provide integrated, in line with hardware and software and services that meet industrial standards, which excellent performance, appropriate price, can help users maximize investment in data center. Kangbia Position is combined with ProLiant optimization server, classification software, storage, and services to provide enterprise-class users with a comprehensive rack environment solution. This means that users will get a more reliable, simple, more powerful and more economical solution. to sum up
Future network development presents the trend of centralized calculations. Centralized management mode with existing dispersion management models, put forward new requirements for the server: save space, easy to expand, and provide uninterrupted services to new requirements for rack servers, rack-based servers And development will be a new growth point in IT industry.
