[Small machine]
Small machine refers to operational principles Similar to PC (Personal Computer) and servers, but performance and use have a very different high-performance computers, which is a high-grade of DCE (Digital Equipment) Company. Performance calculation products.
The small machine has the unique architecture of the PC and its servers, as well as their own patented technology, and some small machine dedicated processors, such as the US SUN, Japan Fujitsu (Fujitsu) small machines based on SPARC The processor architecture, and the US HP is based on the PA-RISC architecture; Compaq is an Alpha architecture. In addition, I / O bus is also different, Fujitsu is PCI, Sun is SBUS, and so on. This means that the cards on the small machine machines, such as network cards, display cards, SCSI cards, etc. may also be dedicated. In addition, the operating system used by the small machine is generally based on UNIX, like Sun, Fujitsu is SUN Solaris, HP is HP-UNIX, IBM is AIX. Therefore, the small machine is a closed-specific computer system. Users using small machines are generally the high-speed calculation capabilities of the security, reliability, and dedicated servers of the Unix operating system.
The manufacturer of small machines now has mainly IBM and HP and wave, dawn. The IBM typical machine has RS / 6000, AS / 400, and the like. Their main features are only a few hours old, so it is also known as the Z Series (Zero Zero). The AS / 400 is mainly used in banks and manufacturing, as well as Domino, main technologies lies in TIMI (technology independent machine interface), single-stage storage, with TIMI technology can do hardware and software independent of each other. The RS / 6000 is more common, used for scientific computing and transaction processing.
【redundancy】
Redundancy is some parts of the configured system. When the system fails, the component of the redundant configuration involves and assumes the operation of the faulty component, thereby reducing the fault time of the system. The redundant system accessories mainly include: Power: High-end server products, the two power systems are widely used, which is loaded, which provides power to the system when working, when a power is faulty, another power supply Take all loads. Some server systems have realized DC redundancy, and other server products such as Micron's NetFrame 9000 implements full redundancy of AC and DC. Storage subsystem: The storage subsystem is the most prone to failure across the server system. The following methods can realize redundancy of the subsystem. Disk mirror: write the same data in two disks, respectively: Disk double: an I / O controller is added to the mirror disk, forming a disk dual-join, making bus contention improvement; RAID: inexpensive redundant Abbreviations of the Redundant Array of Inexpensive Disks. As the name suggests, it consists of several disks, and the individual data streams are written to these disks from a controller to coordinate motion mechanisms. The RAID3 system consists of 5 disks, 4 of which store data, 1 disk storage check information. If a disk is faulty, you can replace the faulty online and re-create the data on the new disk through another 3 disks and the checkboard. RAID5 distributes the check information on 5 disks, which can replace any disk, and the rest is the same as RAID3. I / O card: For the server, mainly refers to the redundancy of the NIC and the hard disk control card. NIC redundancy is to plug in a dual network card in the server. Redundant NIC technology is originally a technique of large machines and medium-sized machines, and now it is gradually being owned by the PC server. PC servers such as Micron's NetFrame9200 implement redundancy of 4 NICs, these 4 NIC bear the 25% network traffic. All Prosignia / ProLiant servers of Kangbo have a fault-tolerant redundant double network card. PCI bus: Represents Micron's highest technical level product NetFrame 9200 adopts triple peer PCI technology, optimizes the bandwidth of the PCI bus, enhances the data transmission speed of high-speed equipment such as hard disk, network card. CPU: The main processor in the system does not often fail, but the symmetrical multiprocessor (SMP) allows multiple CPU sharing to provide some degree of fault tolerance.
