{"id":53836,"date":"2019-12-11T00:00:00","date_gmt":"2019-12-10T21:00:00","guid":{"rendered":"https:\/\/prohoster.info\/blog\/blog_prohoster\/vvedenie-v-ssd-chast-1-istoricheskaya"},"modified":"2020-02-18T14:01:46","modified_gmt":"2020-02-18T11:01:46","slug":"vvedenie-v-ssd-chast-1-istoricheskaya","status":"publish","type":"post","link":"https:\/\/prohoster.info\/en\/blog\/administrirovanie\/vvedenie-v-ssd-chast-1-istoricheskaya","title":{"rendered":"Yesterday, December 9, there was another Kubernetes release \u2014 1.17.","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"

\"Yesterday,
\n
\nExploring the history of disks is the beginning of understanding how solid-state drives work. The first part of our article series \"Introduction to SSD\" will take you on a journey through history and clearly illustrate the difference between SSDs and their closest competitor, HDDs.<\/p>\n

Despite the plethora of various data storage devices, the popularity of HDDs and SSDs is undeniable in today's world. The difference between these two types of drives is clear to the average user: SSDs are more expensive and faster, while HDDs are cheaper and offer more capacity.<\/p>\n

Particular attention should be paid to the unit of measurement for drive capacity: historically, it has developed that decimal prefixes such as kilo- and mega- in the context of information technology imply both the tenth and the twentieth power of two. To avoid confusion, binary prefixes kibi-, mebi-, and others have been introduced. The difference between these prefixes becomes noticeable as the volume increases: when buying a 240-gigabyte disk, you may store 223.5 gibibytes of information on it.
\n<\/a><\/noindex><\/p>\n

Diving into history<\/h2>\n

\n\"Yesterday,<\/a><\/noindex>
\nThe development of the first hard drive began in 1952 by IBM. On September 14, 1956, the final result of these developments was announced \u2014 the IBM 350 Model 1. This drive could hold 3.75 mebibytes of data with quite impressive dimensions: 172 centimeters tall, 152 centimeters long, and 74 centimeters wide. Inside, there were 50 thin disks coated with pure iron, each 610 mm (24 inches) in diameter. The average access time to data on the disk was about 600 ms.<\/p>\n

As time went on, IBM steadily improved the technology. In 1961, they introduced the IBM 1301<\/a><\/noindex> with a capacity of 18.75 megabytes, featuring read heads on each platter. In the IBM 1311<\/a><\/noindex> removable disk cartridges were introduced, and starting in 1970, an error detection and correction system was implemented in the IBM 3330. Three years later, the IBM 3340<\/a><\/noindex> became known as \"Winchester.\"<\/p>\n

Winchester (from English Winchester rifle)<\/b> is a generic term for rifles and shotguns produced by the Winchester Repeating Arms Company in the USA in the second half of the 19th century. These were some of the first multi-shot rifles that gained immense popularity among buyers. They were named after the company's founder, Oliver Fisher Winchester.<\/p>\n

The IBM 3340 consisted of two spindles of 30 MiB each, which is why engineers referred to this disk as \"30-30.\"<\/b>. This name reminded people of the Winchester Model 1894 rifle that uses .30-30 Winchester cartridges, which is why Kenneth Haughton, who led the development of the IBM 3340, said, \"If it's a 30-30, then it must be a Winchester.\" Since then, the term \"Winchester\" has come to refer not only to rifles but also to hard drives.<\/p><\/blockquote>\n

\nThree years later, the IBM 3350 \"Madrid\" was released, featuring 14-inch platters and a 25 ms access time.<\/p>\n

\"Yesterday,<\/a><\/noindex>
\nThe first SSD was created by Dataram in 1976. The Dataram BulkCore consisted of a chassis with eight RAM memory modules of 256 KiB each. Compared to the first hard drive, the BulkCore was tiny: 50.8 cm long, 48.26 cm wide, and 40 cm high. Its data access time was only 750 ns, which was 30,000 times faster than the most advanced HDD at that time.<\/p>\n

In 1978, Shugart Technology was founded, which changed its name to Seagate Technology a year later to avoid conflicts with Shugart Associates. Two years into its operation, Seagate introduced the ST-506 \u2014 the first hard drive for personal computers in a 5.25-inch form factor with a capacity of 5 MiB.<\/p>\n

In addition to the founding of Shugart Technology, 1978 is also remembered for the release of the first Enterprise SSD by StorageTek. The StorageTek STC 4305 could hold 45 MiB of data. This SSD was developed as a replacement for the IBM 2305, had similar dimensions, and cost an incredible $400,000.<\/p>\n

\"Yesterday,<\/a><\/noindex>
\nIn 1982, SSDs entered the personal computer market. The company Axlon designed an SSD for the Apple II called RAMDISK 320, which used RAM chips. Since the drive was based on volatile memory, it came with a battery to maintain data integrity. The battery was sufficient for 3 hours of operation in case of power loss.<\/p>\n

A year later, Rodime released the first 10 MiB hard drive, the RO352, in the familiar 3.5-inch form factor for modern users. Despite being the first commercial drive in this form factor, Rodime essentially introduced no innovations.<\/p>\n

The first product in this form factor is considered to be the floppy disk drive presented by Tandon and Shugart Associates. Moreover, Seagate and MiniScribe agreed on adopting the 3.5-inch industrial standard, effectively sidelining Rodime, which faced the fate of a patent troll and a complete exit from the storage industry.<\/p>\n

\"Yesterday,<\/a><\/noindex>
\nIn 1980, Toshiba engineer Professor Fujio Masuoka registered a patent for a new type of memory called NOR-type Flash memory. The development took 4 years.<\/p>\n

NOR memory is a classic 2D matrix of conductors<\/b>, where one cell (similar to magnetic core memory) is installed at each intersection of rows and columns.<\/p><\/blockquote>\n

\nIn 1984, Professor Masuoka presented his invention at the International Electronics Developers Meeting, where Intel quickly recognized the potential of the development. Toshiba, where Professor Masuoka worked, did not consider Flash memory anything special, and therefore agreed to Intel's request to create a few prototypes for study.<\/p>\n

Intel's expressed interest in Fujio's development prompted Toshiba to allocate five engineers to assist the professor in solving the commercialization issue of the invention. Intel, for its part, assigned three hundred employees to create its own version of Flash memory.<\/p>\n

While Intel and Toshiba were developing Flash drives, two important events occurred in 1986. First, SCSI was officially standardized\u2014a set of agreements for interaction between computers and peripheral devices. Second, the AT Attachment (ATA) interface was developed, known under the brand name Integrated Drive Electronics (IDE), which moved the disk controller inside the disk.<\/p>\n

For three years, Fujio Masuoka worked on improving Flash memory technology, and by 1987 he developed NAND memory.<\/p>\n

NAND memory is the same as NOR memory, organized in a three-dimensional array<\/b>. The main difference is that the access algorithm for each cell became more complex, the area of the cells became smaller, and the overall capacity increased significantly.<\/p><\/blockquote>\n

\nA year later, Intel developed its own NOR-type Flash memory, and the company Digipro created a storage device called Flashdisk based on it. The first version of Flashdisk, in its maximum configuration, could hold 16 MiB of data and cost less than $500.<\/p>\n

\"Yesterday,<\/a><\/noindex>
\nIn the late 80s and early 90s, hard drive manufacturers competed to reduce the size of disks. In 1989, PrairieTek released the PrairieTek 220 disk with 20 MiB in a 2.5-inch form factor. Two years later, Integral Peripherals created the Integral Peripherals 1820 'Mustang' disk with the same capacity, but in a 1.8-inch form factor. A year later, Hewlett-Packard reduced the disk size to 1.3 inches.<\/p>\n

Seagate has remained committed to the 3.5-inch form factor for hard drives and focused on increasing spindle speeds, launching its famous Barracuda model in 1992, which was the first hard disk with a spindle speed of 7200 RPM. But Seagate was not about to stop there. In 1996, the Seagate Cheetah line of drives reached speeds of 10000 RPM, and four years later, the X15 version spun up to an incredible 15000 RPM.<\/p>\n

In 2000, the ATA interface was renamed PATA. This change came about with the introduction of the Serial ATA (SATA) interface, featuring more compact cabling, hot-swapping support, and increased data transfer speeds. Seagate took the lead here as well, releasing the first hard drive with this interface in 2002.<\/p>\n

Initially, the production of Flash memory was very expensive, but in the early 2000s, costs dropped sharply. Taking advantage of this, Transcend launched SSDs ranging from 16 to 512 MiB in 2003. Three years later, companies like Samsung and SanDisk joined the mass production scene. In the same year, IBM sold its disk division to Hitachi.<\/p>\n

Solid-state drives were gaining traction, and an obvious problem arose: the SATA interface was slower than the SSDs themselves. To address this issue, the NVM Express Workgroup began developing NVMe \u2014 a specification for accessing SSDs directly over the PCIe bus, bypassing the 'middleman' in the form of the SATA controller. This would allow data access at PCIe bus speeds. Two years later, the first version of the specification was ready, and a year after that, the first NVMe drive was released.<\/p>\n

Differences Between Modern SSDs and HDDs<\/h2>\n

\nAt a physical level, the difference between SSDs and HDDs is quite noticeable: SSDs lack mechanical components, and data is stored in memory cells. The absence of moving parts allows for quick access to data stored in any part of memory; however, there is a limitation on the number of write cycles. Due to the limited number of write cycles for each memory cell, there is a need for a balancing mechanism \u2014 wear-leveling is performed by relocating data among cells. This task is handled by the drive's controller.<\/p>\n

For load balancing, the SSD controller needs to know which cells are occupied and which are free. The controller can track the writing of data to a cell by itself, but this is not the case for deletion. As we know, operating systems (OS) do not erase data from the disk when a user deletes a file; they mark the corresponding memory areas as free. This solution eliminates the need to wait for a disk operation when using HDDs, but is completely unsuitable for SSDs. The SSD controller works with bytes rather than file systems, and therefore requires a separate message regarding file deletion.<\/p>\n

This led to the creation of the TRIM command (from English \u2014 to trim), which allows the OS to notify the SSD controller about the release of a specific memory area. The TRIM command irreversibly erases data from the disk. Not all operating systems are aware of the need to send this command to solid-state drives, and hardware RAID controllers in disk array mode never send TRIM commands to the drives.<\/p>\n

To be continued\u2026<\/h2>\n

\nIn the following sections, we will discuss form factors, connection interfaces, and the internal organization of solid-state drives.<\/p>\n

In our laboratory, Selectel Lab<\/a><\/noindex> You can test modern HDD and SSD drives yourself and draw your own conclusions.<\/p><\/blockquote>\n

Only registered users can participate in the survey. Please log in<\/a><\/noindex>, please.<\/p>\n

Do you think SSD will replace HDD?<\/h2>\n