Often mentioned redundant power in the server power supply, to understand the difference between where redundant power supplies and ordinary power, so we must first understand the redundancy.
Redundancy refers to repeat some of the components, configuration of the system, when the system failure, the redundant configuration of components involved in the work and bear the failure of components, thereby reducing the failure time Redundan, the system automatic backup, when damage occurs on a device, it can automatically as a standby equipment to replace the equipment is so redundant power we can understand so, dual power supply systems generally use the high-end server products, the two power supply is load balancing, i.e. when the system is working both for the system to provide power, when a power failure occurs, another power will bear all the load. Some server system realizes DC redundancy, some other server products such as Micron's NetFRAME 9000 to realize the full redundancy AC, DC.
In fact, the redundant system parts not only power, storage subsystem is the most fault prone places the entire server system. The following methods can realize the redundancy of the system.
Disk mirroring: will the same data are written to the two disk: disk double: adds a I/O controller as the mirror disk, forming disk double, the bus contention will be improved; RAID: redundant array of inexpensive disks (Redundant array of Inexpensive Disks) abbreviation. As the name implies, it consists of several disks, through a controller coordinated motion mechanism that enables a single data stream in order to write this several disk. The RAID3 system consists of 5 disks, one of the 4 disk storage data, 1 disk storage check information. If a disk failure, can be replaced online fault plate, and the other 3 disks and parity disk to create new data on. RAID5 will check the information distribution in the 5 disks, which can replace any disk, the same as RAID3.
I/O card: the server, mainly refers to the redundant network card and hard disk controller card. Network card redundancy is inserted double network card in server. Redundant network technology for large and medium on the technology, now also gradually is owned by the PC server. The PC server such as Micron, NetFRAME9200 up to 4 network card redundancy, network traffic of the 4 cards each bear 25%. All the ProSignia/Proliant servers Compaq has fault-tolerant redundant dual lan. PCI bus: on behalf of Micron, the highest technical level of product NetFRAME 9200 using three equal weight of PCI technology, the optimization of PCI bus bandwidth, improve the speed of data transmission, high speed hard disk card equipment.
In CPU: system the main processor does not often fails, but symmetric multi processor (SMP) can have multiple CPU sharing work to provide some level of fault tolerance. This redundancy can be understood as many spare, storing the same data
|
Home › News › Industry news
2F168, Area B, Gushio Huating, No. 38, Gushio 1st Road, Gushio Community, Xixiang Street, Bao'an District, Shenzhen City
18198483819 
hx168@cwtipcpower.com.cn