SCSI, SAS, SATA - RAID

 Back  Permanent Memory  Next


A widely used standard for connecting disk (and other) devices on a separate bus is SCSI (Small Computer System Interface) standard. The joint connecting cable connects to multiple devices, each of which has a different address and that are related to connecting water controller - adapter (controller) inserted in a slot connected to the computer bus. In the version of SCSI-1 and SCSI-2 (no longer used) can connect up to eight devices, and SCSI-3 version of the device four p.m., with a transfer rate of 40 Mb/s to 5 Gb/s depending on the revision of the interface. The controller is a separate and distinct unit typically has an external connector and allows connection of devices outside the computer chassis as printers, scanners and the like. The system is faster, but more expensive. Is supported by nearly all computing platforms. Controller has its own BIOS that can be adjusted so the computer's BIOS and in some ways is an extension of your computer's BIOS. Of course it is a complete hardware solution, which is better than the cheaper devices that rely to the drivers and the operating system.

Every device connected to the SCSI series is characterized by logical identifier, which is addressed by the system. This label is known as the SCSI ID is determined at the time the device is connected to the system and must be unique. Mostly defined switches on the device, although newer drives your id furniture automatically. And I have another SCSI controller device, and it has an identification number, usually the number 7, is often fixed. Electronics these devices is more complex than PATA and SATA system and is significantly more expensive. Besides, the devices are manufactured to withstand the loads and long lasting discs revolve almost twice the speed compared to SATA drives. PATA disk system identification is performed jumper (jumper) which, in principle asks the 'M' (master) or 'S' (slave) marked position. For one control over the connecting cable can be controlled by two disk devices have the same IRQ, but over the jumper specifies different identifiers (device '0' and '1'). It should take into account that both drive on the same cable same options regarding transfer rates. With SATA controller system identifier assigned to each channel and the associated port to which it can be connected to only one disk (no jumper on the drive.)

SCSI system that allows one device to connect to two (or more) computers simultaneously. This is not illogical because the machine does not know that the computer is connected, but only executes the commands it receives over the highway. Property SCSI system connectivity to devices not being treated through physical attributes, but through logical. If you use two controllers in two computers and one drive is no reason not to do so, provided that all other requirements are met as specified. SCSI Device Manager tab in the computer usually has a connector (interface) for external peripherals on a tin facade cards, and some of the particular problems in connecting no. Stated one of the prerequisites to create CLUSTER, interconnected computers in order to increase resource utilization and total computing power.

The cable that connects the device length of 10 cm to 25 m, depending on the technology of cable and cable ways reasoning. Passive inference is simply a 'resistant' conclusion at the end of the cable where the resistance value corresponds to the cable impedance, and actively supervised inference impedance cables at the end of the chain, and adapts to her and therefore more possible cable length. All previous-generation SCSI interfaces are designed on the principle of parallel data transfer. Technology is a logical successor to the serial communications SAS (Serial Attached S CSI). Apart from changing the interfaces and modes of communication, the higher the data rate (v3.0 - 12 Gb/s), provides far but the number of devices (16'256) in one domain, compatibility with the preceding standards, and has the ability to work with concerted SATA disks. SAS and SATA compatibility (using the same physical connectors) allows the user the flexibility to choose the devices that will be used, which is of great importance for the development of both standards and their application in a diverse range of computer systems, from home computers to data storage systems such as a network for connecting computers into a single data repository Storage Area Network - SAN and Network Attached Storage - NAS. The difference is in the way the operating system sees disks store. SAN operating system are presented as one large drive and are usually more complex and larger devices that can have a large number of internal disks. NAS has its own operating system that controls internal disks and provides users controlled the resources at our disposal. Operating systems and NAS may not be concurrent but must have adequate software to support each other 'understanding'.

SCSI devices are professional devices and are not subject alphabet. But what is important to mention is the support for RAID (Redundant Array of Independent Disks) system, the ability to be more physical disks can be linked into a single logical unit (volume) in order to increase performance and resistance to failure. Server without connecting disks RAID systems is unreliable. Modern motherboards for home and semi-professional use enabling RAID with PATA and SATA disks.


RAID

In early 80-ies of the capacity of hard drives has been limited and disks with large capacity were rare and expensive. Thus began experimenting with arrays of smaller, cheaper hard drives. Researchers at Berkeley (USA - California) presented the basic rules for disk arrays. They introduced the term RAID (Redundant Array of Inexpensive Disks) to emphasize the benefits of properly implemented disk arrays can provide. Presented guidelines for the development of six different models of disk arrays, RAID 0 to RAID 6. Number does not indicate a hierarchy so that does not mean that the RAID 6 are better or worse than a RAID 0 disk array. As technology has advanced hard disk drives and the cost per megabyte is falling in, the term ' Inexpensive' has been redefined 'Independent' to highlight the advantages of a disk array. RAID disk arrays has become an inevitable part of the server. Acceptance of RAID technology have resulted in the four most important facts:

Characteristics and principles of operation of some methods for creating disk array are:

Possible combinations are more fields in the new field of other species so formed RAID 7, RAID 10, RAID 0+1, RAID 30, RAID 50 From the above description shows that the purpose of RAID is not only to increase the volume but also increase safety. If there is a failure of one of the disks in the RAID 3, 4, 5, 6 already at the level of signaling circuitry follows (optical or acoustic or both) that the disk performance and canceled the fall because of extensive XOR record as to reconstruct the data, but the system is still works. By replacing the drive performs the reconstruction of data on the new drive, which can take several hours, and warning signs will stop when the reconstruction finishes. New replacement drive must be formatted before adding if it is already being used in a RAID array to erase blocks of possibly created before, otherwise it could cause system crashes due to misunderstanding of the existing blocks. At formatted (erased) disk by its inclusion in the RAID array reconstruction follows already mentioned blocks. Of course, the entire RAID technology is based on replacing disks "on hot", so no downtime. Commonly in use RAID 0, RAID 1 and RAID 5 disk system connectivity. RAID in any of these combinations using SAN (Storage Area Network) and NAS (Network Attached Storage) systems for data storage. SAN architecture based on the availability of all disks all the servers in the network regardless of where the server (and operating system) are located. NAS architecture is based on installation of drives in a separate device that has its 'operating system' and that is visible across the network as a data storage device, as in the example of home (SOHO) multimedia system in the Chapter of 'Multimedia'.

Not the same kind or the driver circuit of RAID; embedded chip on the computer's motherboard or a separate autonomous unit on the optional card. Completely autonomous controller with its own program support has full control of disks, 'extends' in some way their computer system BIOS BIOS, and represents the operating system as a single disk. Usually contains a diagnostic program support and complete its management of disks that monitors. If you are relying on the RAID drivers incorporated into the operating system and some incomplete support through the motherboard BIOS, this solution is certainly cheaper, but it means that in the case of replacement motherboards computer to another kind of motherboard disc content is no longer visible, which is not very okay. The operating system is easy to install, and the data should be stored, so maybe our better solutions regarding data storage. Serious firm (bank) data storage system will develop into complementary and synchronized systems that are physically different and sufficiently remote locations.




SUMMARY:

RAID system in a server with one or more microprocessors is not sufficient prerequisite for safe storage or quickly large amounts of data or their processing. For this purpose, it is very efficient fast local network of computers connected in a cluster (a group - a cluster), which is presented to the user as a single entity and basically its resources are managed from one place. Its aim is to provide higher availability and reliability (High Availability - HA) or higher-performance (High Performance Computing - HPC) as compared to a stand alone computer, and the result is a super-computer. Specific program support provides a high degree of integration of computers (nodes in the cluster), provides for their coordinated work together and turns them into a single multi-processor system. Cluster is usually based on one type of OS (Windows, Linux ...) and is physically realized as a group of at least two computers, or servers, or mixed. In this way it is possible to ensure the use of significant computational resources for intensive data processing, for example as part of research projects. As this is a group of computers that individually have limited physical resources, extensibility, scalability physical clusters is much higher than expandability computer singletons. Increasing the number of computers in the cluster are almost the same percentage increase in its mean performance. When special connectivity technologies such independent clusters and connect to the network and are grouped in order to obtain the system for collecting and processing large amounts of data or solve very complex tasks (weather forecast, predicting earthquakes, etc.), get the grid - the network again but unified by Tasks which the performance of its nodes, client OS or way of managing resources of individual nodes are not essential.

But it seems that time passes rotating disks. Technology, although still expensive, based on flash memory, slowly penetrates into use and replaces the standard electro-mechanical drives. Such 'Discs' called SSD ((Solid State Drive), the speed of writing and reading is roughly twice as high as for conventional disk drives with 10 times less power consumption. So, no moving mechanical parts. Basic use in portable devices where used unit capacity of 32 GB and larger.

 PATA SSD - chips
Figure*** 3.3.18 SSD disk. ( + / - )

It is not necessary that the form be as in the previous Figures. Performance without enclosures will surely occupy less space, which is essential for notebooks. Besides, it is significant and greater reliability compared to conventional drives were to consider a mechanical or electronic failure resistance. But no matter what the future holds, the rotating disk and magnetic tape not so quick to 'retire'. In fact, the realization of large-capacity hard disk drive with a standard SDRAM buffer uses FLASH memory in which a large quantity of data is copied from disk to make their availability was faster - SSHD (Solid State Hybrid Drive).

 Back
 Search
 Next

 Content - Home
 Content  Informatics Alphabet
 
Citing this page:
Radic, Drago. " IT - Informatics Alphabet " Split-Croatia.
{Date of access}. <https://informatics.buzdo.com/>.
Copyright © by Drago Radic. All rights reserved. | Disclaimer