Cache memory is a memory chip on a hard disk controller with extremely fast access speeds that acts as a buffer between the internal storage of the hard disk and the external interface. Because the internal data transfer speed of the hard disk and the external interface transfer speed is different, the cache plays a buffer role in it. The size and speed of the cache is directly related to the transmission speed of the hard disk is an important factor, can significantly improve the overall performance of the hard disk. When the hard disk accesses fractional data, it needs to constantly exchange data between the hard disk and the memory, and if there is a large cache, then the fractional data can be temporarily stored in the cache, which reduces the load on the external system and also improves the data transfer speed.
The cache of the hard disk mainly plays three roles: one is pre-reading. When the hard disk is controlled by the CPU command to start reading data, the control chip on the hard disk will control the magnetic head to read the next cluster or several clusters of data in the cache (due to the hard disk data storage is more continuous, so read the hit rate is higher), when you need to read the next cluster or several clusters of data, the hard disk does not need to read the data again, directly to the cache data transfer to memory can be. Data transfer to the memory can be, because the speed of the cache is much higher than the speed of the magnetic head read and write, so it can achieve the purpose of significantly improving performance; the second is to write the action of the cache. When the hard disk receives the command to write data, it will not immediately write the data to the disk, but first temporarily stored in the cache, and then send a "data has been written" signal to the system, then the system will consider that the data has been written and continue to carry out the following work, while the hard disk in the idle (not read or write time) and then the cache of the When the hard disk is idle (not reading or writing), it writes the cached data to the platters. While this is a performance improvement for writing data, it inevitably poses a security risk - if power is suddenly lost while the data is still in the cache, that data will be lost. Hard disk manufacturers naturally have a solution to this problem: when power is lost, the magnetic head will use inertia to write the data in the cache to a temporary storage area other than the zero track, and then wait until the next startup to write the data to the destination; the third function is to temporarily store the recently accessed data. Sometimes, certain data is accessed frequently, and the internal cache of the hard disk will store some of the more frequently read data in the cache, so that when it is read again, it can be directly transferred directly from the cache.
The size of the cache varies from brand to brand and model to model. Early hard drive caches were basically very small, only a few hundred kilobytes, and could not satisfy the user's needs. 2MB and 8MB caches are used in today's mainstream hard drives, and there are also larger caches in servers or special applications, even up to 16MB, 64MB, etc.
This is the first time I've seen a hard drive that has been used for a long time, and I think it is a good idea to use a hard drive that is more compact than a hard drive.
A large cache, while allowing more data to be stored in the cache to improve the access speed of the hard drive, does not mean that the larger the cache, the better it is. Cache application there is an algorithm of the problem, even if the cache capacity is large, but not a highly efficient algorithm, that will lead to the application of cached data hit rate is low, can not effectively play out the advantages of large-capacity cache. Algorithm is complementary to the cache capacity, high-capacity cache requires a more efficient algorithm, otherwise the performance will be greatly discounted, from a technical point of view, high-capacity cache algorithm is directly affecting the performance of the hard disk to play an important factor. A larger cache is a definite trend in future hard drive development.
Internal Transfer Rate
The Internal Transfer Rate (ITR) is the rate of data transfer between the hard disk head and the cache, and is simply the rate at which the hard disk reads data from the platters and stores it in the cache. The internal transfer rate clearly shows the read and write speed of the hard disk, and its level is the decisive factor in evaluating the overall performance of a hard disk, which is the real standard for measuring hard disk performance. Effectively improve the internal transfer rate of the hard disk in order to have the most direct and obvious improvement in the performance of the disk subsystem. At present, hard disk manufacturers strive to improve the internal transfer rate of the hard disk, in addition to improving the signal processing technology, increase the speed of rotation, the main thing is to continue to increase the capacity of a single disk in order to improve the linear density. As the larger the single-disk capacity of the hard disk linear density, the head of the tracking frequency and travel distance can be correspondingly reduced, thereby reducing the average tracking time, the internal transfer rate is also improved. Although hard disk technology is developing rapidly, the internal data transfer rate is still at a relatively low (relative) level, and the low internal data transfer rate has become the biggest bottleneck in hard disk performance. Currently the mainstream home-grade hard drives, the internal data transfer rate is still basically stuck at around 70~90 MB/s, and this data will drop to even lower when working continuously.
The data transfer rate is generally measured in MB/s or Mbit/s, with Mbit/s being the more common unit used in official data, especially for internal data transfer rates.
| high">high">High-quality, durable, and reliable. Byte is the number of bytes, bit is the number of bits, in the computer every eight bits for a byte, that is, 1Byte = 8bit, is a 1:8 correspondence. Therefore, 1MB/s is equal to 8Mbit/s. Therefore, when writing units, we must pay attention to the case of the letter B. In particular, some people also abbreviate Mbit/s as Mb/s, and the size of the letter B can be called a tiny bit, a thousand miles. The above is the general case of MB/s and Mbit/s correspondence, but in the hard disk data transfer rate of the two can not use the general MB and Mbit conversion relationship (1B = 8bit) to convert. For example, the official internal data transfer rate of a product labeled 683Mbit / s, at this time can not simply think that 683 divided by 8 to get 85.375, that 85MB / s is the internal data transfer rate of the hard disk. Because 683Mbit contains many bits of auxiliary information, not exactly the hard disk transfer data, simply use 8 to convert, will not be able to get the real value of the internal data transfer rate. External data transfer rate The hard disk data transfer rate is spelled Data Transfer Rate, or DTR. The hard disk data transfer rate shows the data transfer speed when the hard disk is working, which is a specific performance of the hard disk, it is not static, but with the work of the specific circumstances and changes. In reading hard disk data on different tracks, different sectors; data stored whether continuous and other factors will affect the hard disk data transfer rate. Because of the uncertainty of this data, the manufacturer in the labeling of hard disk parameters, more is the use of external data transfer rate (External Transfer Rate) and internal data transfer rate (Internal Transfer Rate). The External Transfer Rate, also commonly referred to as the Burst Data Transfer or Interface Transfer Rate. It refers to the data transfer rate between the hard disk cache and the computer system, that is, the computer through the hard disk interface from the cache to read out the data to the corresponding controller rate. Usually the hard disk used by the ATA66, ATA100, ATA133 and other interfaces, is the hard disk's theoretical maximum external data transfer rate to express the ATA100 100 represents the hard disk's theoretical maximum external data transfer rate is 100MB / s; ATA133 represents the theoretical maximum external data transfer rate is 133MB / s; and SATA The maximum theoretical external data transfer rate of a SATA hard drive can be up to 150MB/s. These are the theoretical maximum external data transfer rates of the hard drive, which cannot be reached in actual daily operation. Rotationl Speed Rotationl Speed is the speed at which the motor spindle inside the hard drive rotates, the maximum number of revolutions the hard drive's platters can complete in one minute. Rotation speed is one of the most important parameters of a hard drive, and it is one of the key factors in determining the internal transfer rate of the hard drive, which directly affects the speed of the hard drive to a large extent. The faster the RPM of the hard disk, the faster the hard disk can search for files, and the relative transfer speed of the hard disk is also improved. Hard disk rotation speed is expressed in revolutions per minute (RPM), which is an abbreviation of Revolutions Per minute (RPM), the larger the RPM value, the faster the internal transfer rate, the shorter the access time, and the better the overall performance of the hard disk. The spindle motor of a hard disk drives the platters to rotate at high speeds, generating a buoyant force that causes the magnetic heads to float above the platters. The faster the RPM, the shorter the wait time for the sectors to be brought underneath the heads to access the data. So the RPM largely determines the speed of the hard disk. The speed of the ordinary hard disk for home use is generally 5400rpm, 7200rpm several kinds of high speed hard disk is now the first choice of desktop users; and for notebook users is 4200rpm, 5400rpm is mainly, although there are companies have released a 7200rpm laptop hard disk, but in the market is still relatively rare; server users of hard disk performance The SCSI hard drive used in servers has a speed of 10,000rpm, and even 15,000rpm, which is a lot more than the performance of a home appliance. Higher speed can shorten the average seek time and actual read/write time of the hard disk, but with the increasing speed of the hard disk also brought about by the rising temperature, motor spindle wear and tear, and the negative impact of the work of the noise increase. Notebook hard disk rotation speed is lower than desktop hard disk, to a certain extent, is affected by this factor. Notebook internal space is small, the size of the notebook hard disk (2.5-inch) is also designed than the desktop hard disk (3.5-inch) is small, the speed of speed caused by the rise in temperature, the notebook itself puts forward higher requirements for the cooling performance; noise becomes larger, and must take the necessary noise reduction measures, these are on the notebook hard disk manufacturing technology put forward more requirements. At the same time the increase in rotational speed, while others remain unchanged, it means that the power consumption of the motor will increase, the more electricity consumed per unit of time, the shortening of the battery's operating time, so that the laptop's portability will be affected. This is why notebook hard drives generally use a relatively low speed 4200rpm hard drive. RPMs are changed as hard drive motors improve, and now Fluid dynamic bearing motors have fully replaced traditional ball bearing motors. Fluid dynamic bearing motors are typically used in the precision machinery industry, where they utilize a viscous film of liquid oil in place of ball bearings. This avoids direct friction between metal surfaces, minimizing noise and temperature; at the same time, the oil film effectively absorbs vibration, improving vibration resistance; it also reduces wear and tear, improving life expectancy. Average Seek Time Average Seek Time, spelled Average Seek Time, is one of the most important parameters for understanding hard drive performance. It refers to the average value of the time it takes for the magnetic head to move from the beginning to the track where the data is located after the hard disk receives a system command. It reflects the ability of the hard disk to read data to a certain extent, and it is an important parameter that affects the data transfer rate inside the hard disk, measured in milliseconds (ms). Different brands and models of products have different average seek time, but the lower the time, the better the product, the mainstream hard disk products are now in the average seek time of about 9ms. The average seek time is actually a parameter determined by a combination of factors, such as speed, single-disk capacity, and so on. Generally speaking, the higher the RPM of the hard disk, the lower its average seek time; the larger the single-disk capacity, the lower its average seek time. When the capacity of a single disk increases, the head's seek action and travel distance decrease, thus reducing the average seek time and accelerating the speed of the hard disk. Of course, manufacturers also artificially adjust the average seek time of the hard drive for market positioning and noise control. On a hard drive, data is stored in channels and clusters, and after frequent read and write operations, data is often not continuously arranged on the same channel, so the magnetic head tends to need to move between channels repeatedly when reading data, so the average seek time plays a very important role in data transmission. Average seek time also plays a crucial role in reading and writing a large number of small files. When reading and writing large files or storing large amounts of data continuously, the average seek time advantage is not reflected, and at this time, the size of the single-disk capacity, rotational speed, cache is a more important factor. Number of heads Hard disk heads are the key components of the hard disk to read data, its main role is to store in the hard disk platters on the magnetic information into electrical signals to the outside world, and its working principle is the use of special materials, resistance will be with the principle of the magnetic field changes in order to read and write the data on the platters, the head of the good and bad to a large extent, determines the hard disk platters of the storage density. Currently more commonly used is the GMR (Giant Magneto Resisive) giant magnetoresistive head, GMR head of the use of magnetoresistive effect of better materials and multi-layer film structure, which is more sensitive than the previous traditional magnetic head and MR (Magneto Resisive) magnetoresistive head, the relative change in the magnetic field can cause a large change in the resistance value, so as to realize a higher storage density. Magnetic heads are more sensitive than conventional heads and MR (Magneto Resistive) heads. Heads are the tools that read and write to the platters in a hard disk, and are one of the most delicate parts of a hard disk. Heads are made of coils wrapped around a magnetic core. When the hard disk is in operation, the magnetic head reads data by sensing changes in the magnetic field on the rotating platters; it writes data by changing the magnetic field on the platters. To avoid wear and tear on the heads and platters, the heads are suspended above the high-speed rotating platters during operation without direct contact with the platters, and only after the power is turned off do the heads automatically return to a fixed position on the platters (known as the landing zone, where the platters don't store data, and where the platters start). Because of the nature of the work of the magnetic head, the requirements for its magnetic induction sensitivity and precision are very high. Earlier heads used ferromagnetic substances, in the magnetic induction sensitivity is not ideal, so the early hard disk single-disk capacity are relatively low, single-disk capacity is the disk on the high density of magnetic channels, the magnetic head induction degree is not enough, you can not accurately read out the data. This resulted in the early hard disk capacity are very limited. With the development of technology, the magnetic head in the magnetic induction sensitivity and precision have made great progress. Initially, the magnetic head is read, write function together, which is the head of the manufacturing process, technology are very high requirements, and for personal computers, in the process of exchanging data with the hard disk, read data is much faster than the write data, read, write the operation of the two characteristics are also completely different, which led to the read, write the separation of the magnetic head, the two are working separately, do not interfere with each other. Thin-film induction (TEI) heads Between 1990 and 1995, hard disks used TFI read/write technology, where the TFI head is actually a wire-wound core. The TFI read heads reached the limits of their capabilities because they increased their magnetic sensitivity while decreasing their write capability. Anisotropic Magneto Resistive (AMR) heads AMR (Anisotropic Magneto Resistive) In the mid-1990s, Seagate introduced hard disks that used AMR heads.AMR heads use TFI heads for write operations, but use thin strips of magnetic material as the read element. In the presence of a magnetic field, the resistance of the thin strips varies with the magnetic field, which in turn produces a very strong signal. The hard disk interprets the change in resistance of the thin strip due to the change in polarity of the magnetic field and improves the read sensitivity. the AMR head further increases the surface density and also reduces the number of components. Since there is a limit to the amount of resistance change in the AMR film, and AMR technology can support a maximum recording density of 3.3 GB/square inch, there is a limit to the sensitivity of AMR heads. This led to the development of the GMR magnetic head. GMR (Giant Magneto Resistive) The GMR head inherits the read/write technology used in TFI heads and AMR heads. However, its read head exhibits greater sensitivity to magnetic changes on the disk.GMR heads are composed of four thin films of conductive and magnetic materials: a sensing layer, a non-conductive intermediary layer, a magnetic peg layer, and an exchange layer.The sensitivity of the GMR sensors is three times greater than that of the AMR heads, so they are able to increase the density and performance of the disk. The number of heads on a hard drive depends on the number of platters in the drive, and data is stored on both sides of the platters, so a platter corresponds to two heads to work properly. For example, a hard disk with a total capacity of 80GB, using a single disk with a capacity of 80GB, that is only one disk, the disk has data on the front and back, it corresponds to two heads; and the same total capacity of 120GB hard disk, using two disks, there are only three heads, one of which has no head on one side of the disk. Hard disk and disk array common technical terms Hard disk rotational speed (Rotational Speed): that is, the hard disk motor spindle speed, rotational speed is one of the key factors in determining the internal transfer rate of the hard disk, it is fast or slow in a large degree of impact on the speed of the hard disk, at the same time the speed of the speed of the hard disk is also one of the important symbols of the distinction between hard disk grades. The spindle motor of the hard disk drives the platters to rotate at a high speed, generating buoyant force that makes the magnetic head float above the platters. The faster the RPM, the shorter the wait time for the sectors to be brought under the heads. Therefore, the rotational speed largely determines the speed of the hard disk. The most common hard disk speeds on the market today are 5400 rpm, 7200 rpm, and even 10000 rpm; theoretically, the faster the speed, the better. In theory, the faster the speed, the better, because a higher speed can shorten the hard disk's average seek time and the actual read and write time. However, the faster the speed, the greater the heat generation, is not conducive to heat dissipation. Nowadays, the mainstream hard disk rotation speed is usually more than 7200rpm. Average seek time: The time it takes for a hard drive to move the read/write heads on the platters to the designated track to find the corresponding target data, which describes the ability of the hard drive to read data in milliseconds. When the capacity of a single disk increases, the head's seeking action and moving distance decreases, thus reducing the average seek time and speeding up the hard disk. The average seek time of mainstream hard disks on the market today is generally below 9ms, and hard disks larger than 10ms belong to earlier products and are generally not worth buying. Ping 狈 奔?Average latency time): refers to the time when the magnetic head moves to the track where the data is located, and then waits for the block of data you want to continue to rotate to the head under the time, generally between 2ms-6ms. Average access time: The average time it takes for the heads to find a given piece of data, usually the sum of the average seek time and the average latency time. Average access time is the most representative of the hard disk to find a certain data time, the shorter the average access time is better, generally between 11ms-18ms. Note: Many hard drive advertisements nowadays refer to the average access time as the average seek time instead. Burst data transfer rate: The highest rate at which a computer can read data from a hard disk's internal cache via the data bus. It is also called external data transfer rate (EDR). Currently, hard drives with UDMA/66 technology have an external transfer rate of 66.6MB/s. Internal data transfer rate: The maximum data transfer rate between the heads and the hard drive's cache, which is generally dependent on the hard drive's platter rotation speed and platter line density (the spacing of data on the same track). Also called sustained transfer rate. General use of UDMA / 66 technology hard disk internal transfer rate is only 25-30MB / s, only a very small number of products more than 30MB / s, due to the internal data transfer rate is the real bottleneck of the system, so we have to distinguish between these two concepts when buying.In general, however, hard drives with higher single-disk capacity have higher internal They have a higher internal transfer rate when the single-disk capacity is the same. Self-Monitoring Analysis and Report Technology (S.M.A.R.T): Hard drives shipping today basically support S.M.A.R.T technology. S.M.A.R.T technology can monitor the hard disk's head unit, platter motor drive system, internal hard disk circuitry, and platter surface media materials, etc. When S.M.A.R.T monitors and analyzes the hard disk for possible problems, it will alert the user in time in order to avoid loss of computer data.S.M.A.R.T technology must be supported by the motherboard in order to be able to play a role and S.M.A.R.T technology does not guarantee that it can predict the problem. T technology is not guaranteed to predict all possible hard drive failures. Magnetoresistive head technology MR (Magneto-Resistive Head): MR (MAGNETO-RESISTIVEHEAD) that is, short for Magnetoresistive Head. MR technology allows for higher actual recording densities, recording of data, thus increasing hard disk capacity and improving data throughput rates. Currently, there are several generations of MR technology; MAXTOR's Diamond III/IV and others have adopted the latest MR technology. The working principle of magnetoresistive head is based on the magnetoresistive effect to work, its core is a small piece of metal material, its resistance changes with the change of magnetic field, although the rate of change is less than 2%, but because the magnetoresistive element is connected to a very sensitive amplifier, so it can be measured that tiny resistance change.MR technology can increase the capacity of the hard disk by more than 40%.GMR ( GiantMagnetoresistive) Giant Magnetoresistive Heads GMR heads, like MR heads, utilize the principle that the resistance value of a special material changes with the magnetic field to read data on a disk, but GMR heads use materials with a better magnetoresistive effect and a multilayered thin-film structure, which are more sensitive than MR heads, and the same change in the magnetic field can cause a greater change in the resistance value, which can Achieve higher storage density, the existing MR heads can achieve a disk density of 3Gbit-5Gbit/in2 (gigabits per square inch), while the GMR heads can achieve more than 10Gbit-40Gbit/in2. Currently GMR heads are in a mature promotion period, and in the coming years, it will gradually replace MR heads as the most popular head technology. Cache: The cache is where the hard disk exchanges data with the external bus. The process of reading data from the hard disk is to convert the magnetic signal into an electrical signal, and then fill and empty it through the cache again and again, and then fill and empty it again, and then send it out step by step in accordance with the cycle of the PCI bus, so it is obvious that the role of the cache is quite important. In the interface technology has been developed to a relatively mature stage when the size and speed of the cache is directly related to the hard disk transmission speed is an important factor. At present, the cache of mainstream hard disk mainly has several types such as 512KB and 2MB. Its type is generally EDO DRAM or SDRAM, and SDRAM is generally the main one at present. According to the different ways of writing, there are two kinds of write-through type and write-back type. Write-through type When reading hard disk data, the system first checks the request instruction to see if the desired data is in the cache, and if it is then the cache sends out the response data, a process known as a hit. This way the system does not have to access the data in the hard disk. Since SDRAM is much faster than magnetic media, it also speeds up the data transfer. Write-back is when writing hard disk data is also looking in the cache, if found by the cache on the data written to the disk, nowMost of the hard disk is used in the write-back hard disk, which greatly improves performance. MTBF (Minimum Time Between Failures): The maximum amount of time that a hard disk can operate from the start of operation until it fails. A typical hard disk has an MTBF of at least 30,000 or 40,000 hours. Partial Response Maximum Likelihood (PRML): It enables the platters to store more information, and at the same time effectively improves data reading and data transfer rates. It is one of the advanced technologies currently applied to hard disk data reading channel. PRML technology is to divide the hard disk data reading circuit into two "operation pipeline", the first section of the pipeline will be read by the magnetic head signal digitization and then only select part of the "standard" signal transfer to the second section to continue. The first part of the pipeline will digitize the signals read by the magnetic head and then only select part of the "standard" signals to be transferred to the second part of the pipeline to continue processing, and the second part of the received signals will be compared to the pre-programmed signal model of the PRML chip, and then select the signals that have the smallest difference in the combination of the signal to be output in order to complete the process of reading the data. PRML technology can reduce the hard drive error rate of the read data, and therefore further improve the density of disk data. Single track seek time: The time it takes for a magnetic head to move from one track to another. Ultra DSP technology: Ultra DSP performs mathematical operations 10 to 50 times faster than a typical CPU. With Ultra DSP technology, a single DSP chip can simultaneously provide dual functions of processor and drive interface to reduce the use of other electronic components, which can greatly improve the speed and reliability of the hard disk. The interface technology can greatly increase the maximum external transfer rate of the hard disk. The biggest benefit is that the data can be transferred directly from the hard disk to the main memory without taking up more CPU resources, which improves system performance. Hard disk surface temperature: The temperature of the sealed case of a hard disk that rises due to the temperature generated while the hard disk is in operation. Higher temperatures generated during hard disk operation will affect the data read sensitivity of thin-film heads (including MR heads), so hard disks with lower surface temperatures have better data read and write stability. Max full seek time: The total time it takes for the heads to start moving until they finally find the desired block of data. Hard disk mirroring (Disk Mirroring): The simplest form of hard disk mirroring is a host controller with two hard disks that mirror each other. Data is written to both hard disks at the same time, and the data on both hard disks is identical, so when one hard disk fails, the other hard disk can provide data. Hard disk data across the disk (Disk Spanning): using this technology, several hard drives look like a large hard disk; this virtual disk can be stored across the disk data on different physical disks, the user does not need to care about which disk contains the data he needs Hard disk data segmentation (Disk striping): data is scattered across several disks. The first segment of the data is placed on disk 0, the second segment is placed on disk 1, ...... until it reaches the last disk in the hard disk chain, and then the next logical segment is placed on hard disk 0, and then the nexta logical segment is placed on disk 1, ...... and so on until the completion of the write operation. Dual control (Duplex) Duplexing: This refers to using two controllers to drive a hard disk subsystem. When one controller fails, the other controller immediately takes control of the hard disk operation. In addition, if you write the right controller software, you can achieve different hard disk drives work at the same time. Fault Tolerant: (Fault Tolerant): Machines that are fault tolerant are resistant to failure. For example, RAID 1 mirroring system is fault tolerant, one of the mirrored disks fails, the hard disk subsystem can still work. Host Adapter: This refers to the control component (such as a SCSI controller) that enables the host and peripherals to exchange data Hot Fix: This refers to replacing a failed hard disk with a hot backup of the disk. Note that the failed disk is not really physically replaced. The disk used as a hot spare is loaded with the original data from the failed disk, and then the system resumes working. Hot Patch: A system that has a hot backup of the hard disk and can replace the failed disk at any time. Hot Spare: A hard disk electrically connected to the CPU system that replaces the failed disk in the system. The difference with the cold backup is that the cold backup disk is usually not connected to the machine, hard disk failure to replace the faulty disk. Mean Time Between Data Loss (MTBDL - Mean Time Between Data Loss): The average time between events where data loss occurs. Mean Time Between Failure (MTBF - Mean Time Between Failure or MTIF): The average time that a device operates without failure. RAID - Redundant Array of Inexpensive Drives: A technology that combines multiple inexpensive hard drives into a fast, fault-tolerant hard drive subsystem. System Reconstruction (Reconstruction or Rebuild): After a hard disk failure, the process of recovering the failed disk data from other correct hard disk data and parity information. Reconstruction Time: The time it takes to rebuild data for a failed disk. Single Expensive Disk (SED - Singe Expensive Drive) Transfer Rate: The speed at which data can be accessed under different conditions. Virtual Disk:Similar to Virtual Memory, Virtual Disk is a conceptual disk where the user doesn't have to care about which physical disk his data is written on. Virtual disks generally span several physical disks. But the user sees only one disk. Hot Swap: This refers to replacing devices online without downtime. DAS (direct access storage device) NAS (Network Attached Storage) SAN (Storage Area Networks)Related articles
|