
Upgrading And Repairing PCs 21st Edition
- Chapter 3: Processor Specifications
- Chapter 3: Processor Features
- Chapter 10: Flash And Removable Storage
- Chapter 20: PC Diagnostics, Testing, and Maintenance
Alternative Storage Devices
Since the mid-1980s, the primary storage device used by computers has been the hard disk drive. More recently, however, flash-based storage including SSDs (solid-state drives) are increasingly being used as hard drive replacements. Although SSDs can physically replace a hard disk drive (HDD), they operate using a completely different set of principles that may require a treatment unlike that used for conventional HDDs. For data backup, data transport between computers, and temporary storage, secondary removable storage devices such as flash memory devices/drives, optical drives, magnetic tape drives, removable media hard drives, and even floppy drives have been useful supplements to primary storage. Cloud storage, too, now plays a major role in data transfer, storage, and backup.
Flash Memory Devices
Flash memory is a special type of solid-state memory chip that requires no power to maintain its contents. Flash memory cards can easily be moved from digital cameras to laptop or desktop computers and can even be inserted into photo printers or self-contained photo display devices. Flash memory can store any type of computer data, but its original primary application was digital photography. However, more and more digital music players have removable flash memory cards, and so-called thumb or keychain flash memory devices that plug directly into a USB port have helped make flash memory a mainstream storage medium and a popular replacement for some types of magnetic removable-media storage, particularly floppy disks, Zip drives, and SuperDisk drives. Flash memory in the form of SSDs is rapidly increasing in market share as a high-speed alternative to conventional magnetic hard disk storage.
Flash memory was invented by Fujio Masuoka at Toshiba in the early 1980s, with the original patents filed in late 1981. At the time Toshiba unfortunately didn't know how important this invention was, and by 1988 Intel had introduced competitive versions and quickly took the lead in flash memory development and production.
Flash memory is a type of nonvolatile memory that can be electrically programmed and erased. It was originally used in PC motherboards as EEPROM (electrically erasable programmable read-only memory) chips for storing the motherboard basic input/output system (BIOS). Flash ROMs replaced the EEPROM, which could only be programmed or erased by special equipment outside of the motherboard.
Flash memory can be considered sort of a cross between random access memory (RAM) and read-only memory (ROM). Just like RAM, flash memory can be written directly in the system, yet just like ROM it is completely nonvolatile, meaning that it retains data after the power has been turned off (and without a battery like the Complementary Metal Oxide Semiconductor [CMOS] RAM). Besides being nonvolatile, there is one other big difference between flash memory and conventional RAM: The system cannot rewrite Flash memory; it must always erase it first.
When erased, flash memory cells are in a low-voltage state that carries a logical 1 value. The act of writing to (or programming) flash places a charge in the transistor’s floating gate, which changes the 1 to a 0. Once a flash cell is programmed (that is, changed to a 0), the only way it can be changed back to a 1 is by erasing it. The problem with this is that, although you can program individual cells or pages, you can only erase cells or pages on a block basis, and a block usually consists of thousands of cells (512 KB in most cases). The actual programming and erasing process coaxes electrons into and out of the transistor’s floating gate by a process known as Fowler-Nordheim tunneling.
The two major types of flash memory technology are called NOR (Not OR) and NAND (Not AND). Both use the same basic transistor (cell) design, but they differ in how the cells are interconnected. NOR flash works more like dynamic RAM (DRAM), providing high-speed random-access capabilities with the ability to read or write data in single-byte quantities. NOR flash is the type of memory used for flash ROMs, such as those found in motherboards, cell phones, and most devices that have updatable firmware.
On the other hand, NAND flash works more like a storage device, reading and writing data in pages or blocks instead of individual bytes. NAND flash is used in devices that store file-oriented data, such as SSDs, USB key or thumb drives, digital cameras and digital film media, music players, and more. NAND flash is denser than NOR flash, storing more data in a given amount of die space and costing less overall for a given amount of storage.
The speed, low power requirements, and compact size of recent flash memory and SSD devices have made flash memory a perfect counterpart for portable devices such as laptop computers and digital cameras, which often refer to flash memory devices as so-called “digital film.” Unlike real film, digital film can be erased and reshot. Ultra-compact, USB flash memory drives have all but replaced traditional floppy drives, Zip/SuperDisk drives, and even optical discs for transporting data between systems.
Several types of flash memory devices have been popular, including the following:
- CompactFlash (CF)
- SmartMedia (SM)
- MultiMediaCard (MMC)
- SecureDigital (SD)
- Memory Stick
- ATA Flash
- xD-Picture Card
- Solid-state drive (SSD)
- USB flash devices
Some of these are available in different sizes (Type I/Type II). The table below shows the various types of solid-state storage used in digital cameras and other devices, listed in order of introduction.
| Type | L (mm) | W (mm) | H (mm) | Volume (cc) | Date Introduced |
|---|---|---|---|---|---|
| ATA Flash Type II | 54 | 85.6 | 5 | 23.11 | Nov. 1992 |
| ATA Flash Type I | 54 | 85.6 | 3.3 | 15.25 | Nov. 1992 |
| CompactFlash (CF) Type I | 42.8 | 36.4 | 3.3 | 5.14 | Oct. 1995 |
| SmartMedia (SM) | 37 | 45 | 0.76 | 1.27 | Apr. 1996 |
| MultiMediaCard (MMC) | 24 | 32 | 1.4 | 1.08 | Nov. 1997 |
| CompactFlash (CF) Type II | 42.8 | 36.4 | 5 | 7.79 | Mar. 1998 |
| Memory Stick | 21.45 | 50 | 2.8 | 3 | July 1998 |
| Secure Digital (SD) | 24 | 32 | 2.1 | 1.61 | Aug. 1999 |
| xD-Picture Card (xD) | 20 | 25 | 1.7 | 0.85 | July 2002 |
| Memory Stick Duo | 20 | 31 | 1.6 | 0.99 | July 2002 |
| Reduced Size MMC (RS-MMC) | 24 | 18 | 1.4 | 0.6 | Nov. 2002 |
| MiniSD | 20 | 21.5 | 1.4 | 0.59 | Mar. 2003 |
| MicroSD | 15 | 11 | 1 | 0.165 | July 2005 |
| Memory Stick Micro | 15 | 12.5 | 1.2 | 0.225 | Sep. 2005 |
SSDs and USB flash drives are not listed because they do not have a single standardized form factor. SSDs, normally used as hard disk drive (HDD) replacements, come in different form factors, including the same form factor as 1.8-inch, 2.5-inch, and 3.5-inch HDDs as well as adapter card–based versions that plug into a slot in the motherboard.
CompactFlash
CompactFlash was developed by SanDisk Corporation in 1994 and uses the ATA (AT Attachment) architecture to emulate a disk drive; a CompactFlash device attached to a computer has a disk drive letter just like your other drives. When CompactFlash was first being standardized, even full-sized hard disks were rarely larger than 4 GB, so the limitations of the ATA standard were considered acceptable. However, CF cards manufactured after the original Revision 1.0 specification are available in capacities up to 128 GiB. While the current revision 6.0 works in [P]ATA mode, future revisions are expected to implement SATA mode.
The original size was Type I (3.3 mm-thick); a newer Type II size (5 mm-thick) accommodates higher-capacity devices. Both CompactFlash cards are 1.433-inch wide by 1.685-inch long, and adapters allow them to be inserted into laptop computer PC Card slots. The CompactFlash Association oversees development of the standard.
SmartMedia
Ironically, SmartMedia (originally known as SSFDC for solid state floppy disk card) is the simplest of any flash memory device; SmartMedia cards contain only flash memory on a card without control circuits. This simplicity means that compatibility with different generations of SmartMedia cards can require manufacturer upgrades of SmartMedia-using devices. Now defunct, the Solid State Floppy Disk Forum originally oversaw development of the SmartMedia standard.
MultiMediaCard
The MultiMediaCard (MMC) was codeveloped by SanDisk and Infineon Technologies AG (formerly Siemens AG) in November 1997 for use with smart phones, MP3 players, digital cameras, and camcorders. The MMC uses a simple 7-pin serial interface to devices and contains low-voltage flash memory. The MultiMediaCard Association was founded in 1998 to promote the MMC standard and aid development of new products. In November 2002, MMCA announced the development of the Reduced Size MultiMediaCard (RS-MMC), which reduces the size of the standard MMC by about 40% and can be adapted for use with standard MMC devices. The first flash memory cards in this form factor were introduced in early 2004 to support compact smartphones. In 2008, the MMCA merged with JEDEC, which is the global leader in developing open standards for the microelectronics industry.
SecureDigital
A SecureDigital (SD) storage device is about the same size as an MMC (many devices can use both types of flash memory), but it’s a more sophisticated product. SD, which was codeveloped by Toshiba, Matsushita Electric (Panasonic), and SanDisk in 1999, gets its name from two special features. The first is encrypted storage of data for additional security, meeting current and future Secure Digital Music Initiative (SDMI) standards for portable devices. The second is a mechanical write-protection switch. The SD slot can also be used for adding memory to Palm PDAs. The SDIO standard was created in January 2002 to enable SD slots to be used for small digital cameras and other types of expansion with various brands of PDAs and other devices. The SD Card Association was established in 2000 to promote the SD standard and aid the development of new products. Note that some laptop computers have built-in SD slots.
Reduced-size versions of SD include MiniSD (introduced in 2003) and MicroSD (introduced in 2005). MiniSD and MicroSD are popular choices for smartphones and can be adapted to a standard SD slot. MicroSD is compatible with the TransFlash standard for mobile phones.
The original SD standard allowed for memory card capacities of up to 2 GB. To support higher capacities the SDHC (High Capacity) standard was created in 2006. SDHC supports cards from 4 GB to 32 GB in capacity. To increase capacity beyond 32GB, the SDXC (eXtended Capacity) format was released in 2009. SDXC supports capacities of up to 2 TB. Note that devices are backward compatible, meaning that a device that supports SDXC also supports SDHC and standard SD cards. A device that supports SDHC also accepts standard SD cards, but such a device does not support SDXC cards. Devices that support only standard SD do not support either SDHC or SDXC cards.
Sony Memory Stick
Sony, which is heavily involved in both laptop computers and a variety of digital cameras and camcorder products, has its own proprietary version of flash memory known as the Sony Memory Stick. This device features an erase-protection switch, which prevents accidental erasure of your photographs. Sony has also licensed Memory Stick technology to other companies, such as Lexar Media and SanDisk.
Lexar introduced the enhanced Memory Stick Pro in 2003. Memory Stick Pro includes MagicGate encryption technology, which enables digital rights management, and Lexar’s proprietary high-speed memory controller. Memory Stick Pro is sometimes referred to as MagicGate Memory Stick.
The Memory Stick Pro Duo is a reduced-size, reduced-weight version of the standard Memory Stick Pro. It can be adapted to devices designed for the Memory Stick Pro.
Sony later released “Mark 2” certified versions of the Memory Stick Pro in 2008. This certification indicated that the cards were suitable for use with AVCHD (Advanced Video Coding High Definition) recording devices. Sony also released a smaller Memory Stick Micro (also called M2) format in 2006, which was designed to compete with microSD. In 2009 Sony announced the Memory Stick XC (eXtended Capacity) format in order to compete with SDXC.
Because the Memory Stick formats are proprietary and only used in Sony devices, I recommend avoiding them wherever possible. In order to avoid using expensive and hard to find proprietary memory, make sure any device you purchase accepts industry standard memory such as SD. Fortunately, Sony’s newer devices are including support for industry standard SD memory formats in response to the negative backlash against its proprietary Memory Stick.
ATA Flash PC Card
Although the PC Card (PCMCIA) form factor has been used for everything from game adapters to modems, SCSI (Small Computer Systems Interface) cards, network cards, and more, its original use was computer memory, as the old PCMCIA (Personal Computer Memory Card International Association) acronym indicated.
Unlike normal RAM modules, PC Card memory acts like a disk drive, using the PCMCIA ATA (AT Attachment) standard. PC Cards come in three thicknesses (Type I is 3.3 mm, Type II is 5 mm, and Type III is 10.5 mm), but all are 3.3-inch long by 2.13-inch wide. Type I and Type II cards are used for ATA-compliant flash memory and the newest ATA-compliant hard disks. Type III cards are used for older ATA-compliant hard disks; a Type III slot also can be used as two Type II slots.
xD-Picture Card
In July 2002, Olympus and Fujifilm, the major supporters of the SmartMedia flash memory standard for digital cameras, announced the xD-Picture Card as a much smaller, more durable replacement for SmartMedia. In addition to being about one-third the size of SmartMedia—making it the smallest flash memory format yet—xD-Picture Card media has a faster controller to enable faster image capture.
Both 16 MB and 32 MB cards (commonly packaged with cameras) record data at speeds of 1.3 MB/s, whereas 64 MB and larger cards record data at 3 MB/s. The read speed for all sizes is 5 MB/s. The media is manufactured for Olympus and Fujifilm by Toshiba, and because xD-Picture media is optimized for the differences in the cameras (Olympus’s media supports the panorama mode found in some Olympus xD-Picture cameras, for example), you should buy media that’s the same brand as your digital camera.
Just as with the proprietary Sony Memory Stick formats, I also recommend avoiding the proprietary xD-Picture card format wherever possible. Instead, I only recommend purchasing devices that use industry standard memory card formats such as SD. Because of the backlash against proprietary formats, Olympus and Fujifilm abandoned xD-Picture card in 2010.
In general, a solid-state drive (SSD) is any drive using solid-state electronics (that is, no mechanical parts or vacuum tubes). Many people believe that SSDs are a recent advancement in computer technology, but in actuality they have been around in one form or another since the 1950s, well before PCs even existed.
Today, solid-state drives are used for many of the tasks magnetic and optical drives have traditionally performed, including system drives, primary and secondary data storage, and removable-media storage.
Virtual SSD (RAMdisk)
Although most people think of a physical drive when they discuss SSDs, these drives are available in both physical and virtual form. A virtual SSD is traditionally called a RAMdisk because it uses a portion of system RAM to act as a disk drive. The benefits are incredible read/write performance (it is RAM, after all), whereas the drawbacks are the fact that all data is lost when the system powers down or reboots, and that the RAM used for the RAMdisk is unavailable for the operating system (OS) and applications.
RAMdisk software has been available for PCs since right after the PC debuted in late 1981. IBM included the source code to a RAMdisk program (later called VDISK.SYS) in the March 1983 PC DOS 2.0 manual, as part of a tutorial for writing device drivers. (Device driver support was first implemented in DOS 2.0.) IBM later released VDISK.SYS as part of PC DOS 3.0 in August 1984. Microsoft first included a RAMdisk program (called RAMDRIVE.SYS) with MS-DOS 3.2 (released in 1986). Versions of RAMDRIVE.SYS were included in DOS and Windows versions up to Windows 3.1, and a renamed version called RAMDISK.SYS has been included with Windows XP and Windows 7/Vista. However, they are not automatically installed, and they are not well documented. These DOS- or Windows-based RAMdisk programs are useful for creating high-speed SSDs using existing RAM. As an alternative to using RAMDRIVE.SYS, you can use a variety of commercial and freeware utilities available for Windows and for Linux on Wikipedia.
Flash-Based SSDs
Shortly after the release of the IBM PC in 1981, several companies developed and released physical solid-state drives that could function as direct hard drive replacements. Many of these used conventional dynamic or static RAM, with an optional battery for backup power, whereas others used more exotic forms of nonvolatile memory, thus requiring no power to retain data. For example, Intel had released “bubble” memory in the late 1970s, which was used in several SSD products. Bubble memory was even included in the Grid Compass in 1982, one of the first laptops ever released. Although SSDs can use any type of memory technology, when people think of modern SSDs, they think of those using flash memory. Flash-based SSDs more recently started appearing in commercially available laptop PCs from Dell, Asus, Lenovo, and others in 2007–2008. Since then, many other laptop and desktop PC manufacturers have introduced systems with flash-based SSDs.
Ever since SSDs first became available for PCs in the early 1980s, many have thought that they would universally replace hard drives. Well, it has been nearly 30 years since I first heard that prediction, and it is just now becoming partially true. Until recently, the principle barriers preventing SSDs from overtaking hard disks has been cost per GB and performance. Early SSDs were slower than HDDs, especially when writing data, and performance would often fall dramatically as the drive filled up. The development of controller hardware and operating systems optimized for SSDs have enabled recent SSDs to surpass conventional hard disk drives in performance. Although SSDs are still more expensive per GB than traditional hard disk drives, SSDs are now widely used for applications where cost is not as important as performance and durability: Tablets, smartphones, netbooks, and Ultrabooks use SSDs.
Many systems now strike a balance between the higher performance of SSDs and the greater capacity of conventional hard disk drives by using both technologies. Many Ultrabooks use a small SSD (32 GB is a typical size) for the operating system and a conventional or hybrid SATA hard disk for applications and system storage. Many high-performance desktop systems also use an SSD from 128 GB to 512 GB as a system drive, and a traditional hard disk for additional storage.
Note: A hybrid SATA hard disk includes a small amount of flash memory used to cache most-frequently-used information.
Virtually all modern SSDs use the SATA (Serial ATA) interface to connect to the PC and appear just like a standard hard disk to the system. Both 2.5-inch and 1.8-inch SSDs are shown in the image below. Some high-performance SSDs come in a card-based form factor, usually designed for PCI Express slots.
SLC Versus MLC
As previously mentioned, SSDs use NAND flash technology. Two subtypes of this technology are used in commercially available SSDs: SLC (single-level cell) and MLC (multilevel cell). SLC flash stores one bit in a single cell, whereas MLC stores two or more bits in a single cell. MLC doubles (or more) the density, and consequently lowers the cost, but this comes at a penalty in performance and usable life. SSDs are available using either technology, with SLC versions offering higher performance, lower capacity, and higher cost. Most mainstream SSDs use MLC technology, whereas more specialized high-end products (mostly for server or workstation systems) use SLC.
One major problem with flash memory is that it wears out. SLC flash cells are normally rated for 100,000 Write/Erase (W/E) cycles, whereas MLC flash cells are rated for 10,000 or fewer W/E cycles. When used to replace a standard hard drive, this becomes a problem because certain areas of a hard drive are written to frequently, whereas other areas may be written to only a few times over the life of the drive. To mitigate this wear, SSDs incorporate sophisticated wear-leveling algorithms that essentially vary or rotate the usage of cells so that no single cell or group of cells is used more than another. In addition, spare cells are provided to replace those that do wear out, thus extending the life of the drive. Considering the usage patterns of various types of users, SSD drives are generally designed to last at least 10 years under the most demanding use, and most last much longer than that. As SSD capacity increases, so does the ability of the wear-leveling algorithm to spread out data among available cells.
Note that, because of the way SSDs work internally, the concept of file fragmentation is immaterial, and running a defragmenting program on an SSD does nothing except cause it to wear out sooner. Unlike magnetic drives, which must move the heads to access data written to different physical areas of the disk, an SSD can read data from different areas of memory without delay. The concept of the location of a file becomes moot with wear leveling, in that even files that are presented as contiguous to the file system are actually scattered randomly among the memory chips and cells in the SSD. Because of this, SSDs should not be defragmented like traditional magnetic drives.
Note: Windows 7 and 8 are SSD-aware, which means they can tell an SSD from a standard magnetic drive. These versions of Windows determine this information by querying the drive’s rotational speed via the ATA IDENTIFY DEVICE command. (SSDs are designed to report 1 RPM.) When Windows detects that an SSD is attached, it automatically turns off the background Disk Defragmenter function, thus preserving drive endurance. When using SSDs with Windows Vista and earlier versions, you should manually disable or otherwise prevent any form of defragmentation program or operation from running on SSDs.
TRIM Command
Another technique to improve SSD endurance and performance is an extension to the ATA interface called the TRIM command. This allows an SSD-aware OS (such as Windows 7 or later) to intelligently inform the SSD which data blocks are no longer in use, thus allowing the drive’s internal wear leveling and garbage collection routines much more space to work with, which allows the drive to maintain a high level of performance especially after all blocks have been written to at least once. For this to work, both the drive and the OS must support the TRIM command. Windows 7 and Server 2008 R2 and later are SSD aware and support the TRIM command, whereas earlier versions of Windows do not. SSDs released in 2009 or later generally support the TRIM command, whereas those that do not may be able to add support via a firmware upgrade. When you are upgrading the firmware on an SSD, it is highly recommended to have a full backup because in some cases a firmware upgrade reinitializes the drive, wiping all data in an instant.
When an OS deletes a file or otherwise erases data from a drive, it doesn’t actually erase data. Instead, the OS simply marks the file allocation or master file tables to indicate that those blocks are available, while leaving the data in them untouched. This works fine on a normal HDD because overwriting is the same as writing, but it greatly hinders a flash drive since a flash drive cannot overwrite data directly. On a flash drive, any overwriting causes the drive to first write any previously existing data to a new block, then erase the block, and finally write the new data. Over time, this results in the SSD filling up and slowing down, even though from the OS point of view there is a lot of empty space.
When TRIM is used, whenever a file is deleted, copied, or moved or the drive is reformatted, the drive is immediately informed of all the blocks that are no longer in use. This allows the drive controller to erase the unused blocks in the background, ensuring that there is always a sufficient supply of erased blocks available to keep write performance at near like-new levels.
To further improve SSD performance, Windows 7 and later disable features such as Superfetch and ReadyBoost as well as prefetching on SSDs with random read, write, and flush performance above a certain threshold.
When running a non-TRIM aware OS (Vista, XP, and earlier), you may still be able to take advantage of TRIM by installing a TRIM-aware application. For example, Intel provides a program called the Intel SSD Optimizer (part of the Intel SSD Toolbox) that you can periodically run to report to the drive which files have been deleted. Other SSD manufacturers provide similar tools (often called wiper.exe) as well. If you are running a non-TRIM aware OS with an SSD, check with the SSD manufacturer to see if it has an optimization tool available.
Partition Alignment
Another issue with SSDs is that they are normally designed to read and write 4 KB pages and to erase data in 512 KB blocks. Windows XP and earlier OSs normally start partitions 63 sectors into a disk, which means that the OS file system components and clusters overlap pages and blocks, resulting in more pages being read or written, and more blocks being erased than necessary, which can cause a noticeable performance hit.
SSDs perform at their best when partitions are created with the SSD’s alignment needs in mind. All the partition-creating tools in Windows 8/7/Vista place newly created partitions with the appropriate alignment, with the first partition starting an even 2048 sectors into the disk. Because this is evenly divisible by both 4 KB (eight sectors) and 512 KB (1024 sectors), there is no overlap between OS file system cluster and SSD page/block operations.
Even if you are using Windows 8/7/Vista or another OS that normally creates aligned partitions, you may still have misaligned partitions if the OS was installed into an existing partition or as an upgrade. Many of the drive manufacturers have free partition alignment tools available that can check and even correct the alignment of partitions on the fly. When creating new partitions on an SSD, you can optionally use the DISKPART command to manually set the offset to the start of the first partition such that all partitions on the drive will be properly aligned. With manual intervention, you can ensure that even Windows XP and earlier will create partitions that are properly aligned for maximum-performance.
SSD Applications
SSDs are ideal for laptops because they are more rugged (no moving parts), weigh less, and consume less power. The weight savings is fairly minor because the difference between an SSD and a conventional drive of the same (or even greater) capacity is generally only a few grams. The power savings is more real—SSDs only draw about a tenth of a watt compared to about 1 watt for an HDD (average). But even that may be overstated. Although drawing one-tenth the power sounds like a considerable savings, compared to other components such as the CPU, GPU, and display, each of which draw 30 watts or more, the overall power savings in going from a standard HDD to an SSD is relatively low in comparison to the total power consumed.
SSDs are ideal as the boot drive for desktop systems because of their performance. Using an SSD can drop boot or resume from hibernation times dramatically. SSDs are less ideal for storing large amounts of data because capacities are less than what is available for conventional HDDs.
Will your next computer contain an SSD? If you buy a tablet, a netbook, or an Ultrabook, the answer is “very likely.” SSDs are big enough to contain the operating system and applications and are rapidly dropping in price per GB compared to magnetic storage. Netbooks, Ultrabooks, and other PCs can use external hard disk drives or cloud-based storage for data storage, and some Ultrabooks include both a small SSD for use by Windows and a larger hybrid hard disk (magnetic storage with a small amount of flash memory) for application and data storage. Tablets can use flash memory slots, cloud-based storage, or both to supplement the capacity of an SSD.
As an alternative to floppy and Zip/SuperDisk-class removable-media drives, USB-based flash memory devices have rapidly become the preferred way to move data between non-networked systems. The first successful drive of this type—Trek’s ThumbDrive—was introduced in 2000, and since then hundreds of others have been introduced.
Note: Some USB flash memory drives are even built into watches, pens, bottle openers, and knives (such as the Victorinox SwissMemory Swiss Army Knife).
Unlike other types of flash memory, USB flash drives don’t require a separate card reader; they can be plugged into any USB port or hub. Any system running Windows XP or later can immediately recognize, read from, and write to a USB flash drive. As with other types of flash memory, USB flash drives are assigned a drive letter when connected to the computer. Most have capacities ranging from 2 GB to 64 GB, but can be as large as 256 GB with even larger capacities planned for the near future. Typical read/write performance of USB 1.1-compatible drives is about 1 MB/s. Hi-Speed USB 2.0 flash drives are much faster, providing read speeds ranging from 5 MB/s to 15 MB/s and write speeds ranging from 5 MB/s to 13 MB/s. SuperSpeed USB (USB 3.0) flash memory drives are now available for USB 3.0 ports common on most modern desktops and laptops. Although some USB 3.0 flash memory drives support read/write performance up to 150 MB/s, the actual interface is designed to support up to 625 MB/s (5 Gb/s). As controllers improve, future USB 3.0 flash memory drives are likely to provide performance closer to the maximum speed of the interface (Ed.: Minus encoding overhead). Because Hi-Speed and SuperSpeed USB USB flash drives vary in performance, be sure to check the specific read/write speeds for the drives you are considering before you purchase one.
USB 3.0 FAQ
Q. Does my computer support USB 3.0? How can I tell?
A. USB 3.0 ports use blue connectors and are typically marked with an SS next to the USB fork icon. In Windows Device Manager, look for an eXtensible Host Controller Interface (XHCI) controller entry in the Universal Serial Bus category.
Q. Will my new 32 GB SupersSpeed USB thumbdrive run at SuperSpeed or HighSpeed if I use 2.0 ports?
A. A USB 3.0 drive must be connected to a USB 3.0 port to run at SuperSpeed (5 Gb/s). If a USB 3.0 drive is connected to a USB 2.0 port, it runs at USB 2.0 speeds (Hi-Speed 480 Mb/s).
For additional protection of your data, some USB flash drives have a mechanical write-protect switch. Others include or support password-protected data encryption as an option, and most are capable of being a bootable device (if supported in the BIOS). Some drives feature biometric security—your fingerprint is the key to using the contents of the drive—whereas others include more traditional security software.
Some companies have produced bare USB flash drives that act as readers for MMC, SD, xD-Picture Card, Compact Flash, and Memory Stick flash memory cards. These USB flash readers are essentially USB flash drives without flash memory storage on-board. You can use them as a card reader or as a USB drive with removable storage.
As with any storage issue, you must compare each product’s features to your needs. You should check the following issues before purchasing flash memory-based devices:
- Which flash memory products does your camera or other device support? Although adapters allow some interchange of the various types of flash memory devices, for best results, you should stick with the flash memory type your device was designed to use.
- Which capacities does your device support? Flash memory devices are available in ever-increasing capacities, but not every device can handle the higher-capacity devices. Check the device and flash memory card’s websites for compatibility information. In some cases, firmware updates can improve a device’s compatibility with larger or faster flash memory card standards.
- Are some flash memory devices better than others? Some manufacturers have added improvements to the basic requirements for the flash memory device, such as faster write speeds and embedded security. Note that these features usually are designed for use with particular digital cameras only. Don’t spend the additional money on enhanced features if your camera or other device can’t use those features.
Only ATA Flash cards can be attached directly to an older laptop computer’s PC Card slot. All other devices need their own socket or some type of adapter to transfer data. Figure 10.2 shows how the most common types of flash memory cards compare in size to each other and to a penny.
The table below provides an overview of the major types of flash memory devices and their currently available maximum capacities.
| Device | Minimum Capacity | Maximum Capacity | Notes |
|---|---|---|---|
| Compact Flash (CF+) | 16 MB | 128 GB | Highest capacity; most flexible format; supported by most DSLRs. Lexar Media and SanDisk also make faster versions of CF+ media; Lexar Media also makes LockTight secured access media. |
| SmartMedia | 16 MB | 512 MB | Popular choice for older Fujifilm and Olympus digital cameras. |
| MultiMediaCard (MMC) | 16 MB | 4 GB | MMC cards can work in most SD slots. |
| RS-MMC | 128 MB | 2 GB | Use adapter to plug in to MMC slots. |
| Secure Digital (SD) | 16 MB | 2 TB | SD cards do not work in MMC slots. Used by most brands of consumer-level digital cameras. SD High Capacity (SDHC) cards have capacities of 4 GB up to 32 GB. Devices that are compatible with SDHC can also use SD cards, but not vice versa. SDXC cards have capacities from 32 GB up to a theoretical maximum of 2 TB. Devices that are compatible with SDXC cards can also use SDHC and SD cards, but not vice versa. |
| MiniSD | 128 MB | 4 GB | Use adapter to plug in to SD slots. |
| MicroSD | 128 MB | 16 GB | Use adapter to plug in to SD slots. |
| Memory Stick | 16 MB | 128 MB | Developed by Sony and licensed to other vendors. Proprietary - not recommended. |
| Memory Stick Pro | 256 MB | 4 GB | The enhanced high-speed version of Memory Stick. Memory Stick with digital rights management support. Proprietary - not recommended. |
| Memory Stick Pro Duo | 256 MB | 16 GB | Reduced-size version of Memory Stick Pro. Proprietary - not recommended. |
| ATA Flash | 16 MB | 2 GB | Plugs directly into a PC Card (PCMCIA) slot without an adapter. |
| xD-Picture Card | 16 MB | 2 GB | Use the same brand as your digital camera for the best results. Proprietary - not recommended. |
| USB flash drive | 16 MB | 256 GB | Some include password-protection and write-protect features. |
I normally recommend only devices (cameras, PDAs, and so on) that use Secure Digital (SD, including SD variants like MiniSD or MicroSD), CompactFlash (CF), or USB flash memory. Any of the others I generally do not recommend due to proprietary designs and higher costs as well as limitations in capacity and performance.
Secure Digital has become the most popular format in modern devices. It is reasonably fast and is available in capacities approaching those of CF, and in smaller MiniSD and MicroSD formats, which are physically compatible with the full-sized SD using adapters. SD sockets also take MMC cards, which are basically thinner versions of SD. Note that the opposite is not true—MMC sockets do not accept SD cards.
CF is the most widely used format in professional devices. It offers the highest capacity, in a wide range of speeds in a reasonably small size.
SD Cards Speed Class and UHS Speed Class Markings
SDHC cards, and some SD and most SDXC cards, are marked with a stylized C icon containing a number (see the image below). This is the speed-class marking. Speed class markings include 2, 4, 6, and 10, with 2, 4, and 10 being the most common. Class 2 cards provide sustained read/write speeds of 2 MB/s or faster, class 4 cards provide sustained read/write speeds of 4 MB/s or faster, and so on.
The higher the speed class number, the faster the card can transfer data.
Speed Class markings on typical SDHC cards.
The table below lists the speed-class recommendations for different types of still photo and video recording tasks.
| Speed Class | Continuous Shooting Still Photography (JPEG) | Continuous Shooting Still Photography (RAW) | HD Video |
|---|---|---|---|
| 2 | No | No | No |
| 4 | Yes | No | Yes |
| 6 | Yes | Yes | Yes |
| 10* | Yes | Yes | Yes |
*Requires high-speed bus interface in supported devices; also supports HD still consecutive recording.
As you can see from the table above, speed-class ratings are a useful guide to selecting suitable cards for use with video recording. However, continuous shooting in still photography is more heavily influenced by maximum card and bus speed. Thus, many vendors of SDHC cards also provide their own maximum speed ratings.
Note: For more information about vendor speed ratings, such as 133X, and how they correspond to speed-class ratings, see this Lexar white paper (PDF format) on “Understanding SD Association Speed Ratings”.
SDHC and SDXC cards might also be labeled with a different marking, the Ultra High Speed (UHS) Class marking. This mark resembles a stylized U with a number inside. U1 indicates a card has a maximum speed of 104 MB/s. All cards with a UHS class 1 marking are also Class 10-compliant.
File Systems Used by Flash Memory
USB flash memory drives typically use the FAT32 file system, which supports up to 2 TB of capacity.
Flash memory cards up to 2 GB in size use the FAT16 file system, while SDHC and Compact Flash memory cards with capacities above 2 GB use FAT32. SDXC cards have capacities larger than 32 GB and use the Microsoft-developed exFAT file system.
SDXC cards typically use the exFAT (FAT64) file system. exFAT uses a 64-bit address table to support capacities of up to 512 TB (recommended) and 64 ZB (maximum size). FAT64 is designed specifically for use with flash memory, is optimized for movie recording, and supports universal time coordinate (UTC) time stamps.
Windows Vista SP1, Windows 7 SP1, and newer include built-in support for SDXC. To add support for SDXC to older versions of Windows, see the service pack requirements.
Note, however, that Windows Vista and 7 may have problems determining the correct size of an SDXC card with a capacity over 32 GB. To correct this, downloadable updates for Windows 7 and Windows Vista are available.
Flash Card Readers
You can purchase several types of devices to enable the data on flash memory cards to be read in a PC. Although it is possible to connect most digital cameras to a PC via USB, in many cases, you must use custom cables.
Card Readers
The major companies that produce flash card products sell card readers that can be used to transfer data from flash memory cards to PCs. These card readers typically plug in to the computer’s USB ports for fast access to the data on the card.
In addition to providing fast data transfer, card readers save camera battery power because the camera is not needed to transfer information. Because computer and electronics device users might have devices that use two or more types of flash memory, many vendors now offer multiformat flash memory card readers, such as the SanDisk 12-in-1 Card Reader/Writer shown in the figure below.
Card readers are also available as internal bay-mounted devices that plug in to the internal front panel USB port connectors found on most modern motherboards. Other than the mounting location, internal bay mounted card readers are functionally identical to external readers. One problem with internally mounted readers is that you usually have to open the PC to disconnect them. Disconnection is normally required when installing an OS to prevent issues with improper drive letter assignments.
Before you purchase an external card reader, check your PC and your photo printer, either of which may already have a built-in reader. The built-in readers in photo printers are especially convenient because you can often print photos directly from the flash card without having to transfer the files to your PC. Many laptops include a single-slot card reader that supports SD, Memory Stick, and xD-Picture Card media. If you use CompactFlash, you still need to use an external card reader.
Note: USB 3.0 card readers are now available and provide transfer rates that are much faster than USB 2.0 card readers. If you use systems with USB 3.0 ports, consider upgrading to a USB 3.0 card reader for faster file copying performance.
ReadyBoost Support
Microsoft Windows Vista and newer all include support for using high-speed flash memory cards and USB drives as a disk cache known as ReadyBoost.
When you configure flash memory as ReadyBoost, it is used to hold information about application files and libraries that has been loaded into memory by SuperFetch. SuperFetch helps improve system performance by providing information from RAM rather than directly from disk.
Using ReadyBoost to hold SuperFetch information can help improve your computer’s performance if it has a slow hard disk (4.0 or lower score on the Windows Experience Index [WEI]) or if you have limited system RAM (4 GB or less).
A flash memory device must have at least 256 MB of free space to be usable with ReadyBoost. Maximum size of the ReadyBoost cache varies by the file system used by the device:
- 4 GB on a device using FAT32
- 32 GB on a device using NTFS
Windows automatically tests an eligible flash memory device for ReadyBoost-compatibility when you plug it in. When a device is tested by Windows to determine if it is ReadyBoost-compatible, the random read/write speed of the media must meet the following minimums:
- 2.5 MB/s throughput for random 4 KB reads
- 1.75 MB/s throughput for random 1 MB writes
ReadyBoost is used only for non-sequential disk reads. To help determine the performance improvement that ReadyBoost provides on a specific system, use the Performance Monitor tool in the Computer Management console and enable the ReadyBoost Cache counters (cache read bytes/sec, cache reads/sec, skipped read bytes/sec, and skipped reads/sec). Note that Windows 7 supports multiple ReadyBoost cache devices.
Note: Conventional USB 2.0-based card readers are usually not fast enough for use with ReadyBoost, but laptop internal card readers or USB card readers designed for use by high-performance Compact Flash (CF) media are typically fast enough.
Note: For the purposes of data storage, USB flash memory drives that are too slow to support ReadyBoost are still compatible with Windows. To enable ReadyBoost on a USB flash drive or flash memory card, open the card’s properties sheet and click the ReadyBoost tab. Windows will test the device’s performance and advise you whether the drive or card is fast enough to support ReadyBoost. Click Use this device (Windows will select the recommended size of the ReadyBoost cache on the drive). Click Apply, and then OK to begin using the drive or card for ReadyBoost.
To disable ReadyBoost, select Do not use this device.
Cloud-Based Storage
Cloud-based storage (remote storage that is accessed by the Internet) has become a popular alternative to flash-based or optical storage for data storage, exchange, and backup.
Although some earlier cloud-based storage services used proprietary interfaces, the trend is increasingly in the direction of making cloud-based storage, sync, or backup look and act like another drive folder.
Although cloud-based storage services mimic a local folder, they use powerful encryption technologies to protect data from unauthorized users. The performance of cloud-based storage depends primarily upon the speed of your Internet connection and the priority level of the service running on your computer. Typically, automatic cloud-based backup services run at a low priority to avoid slowing down users’ normal experience of using their devices. However, the trade-off is that restoration of lost data can take several days or longer.
Before choosing a cloud-based storage, sync, or backup service, look at capacity, prices, and performance. Typically, lower-cost or free services have limits on capacity and run more slowly than paid versions. If you are looking at cloud-based storage for a group of workers or family members, be sure to compare the costs and features of a shared plan over multiple individual plans.
Floppy Disk Drives
Alan Shugart is generally credited with inventing the floppy disk drive in 1967 while working for IBM. One of Shugart’s senior engineers, David Noble, actually proposed the flexible medium (then 8 inches in diameter) and the protective jacket with the fabric lining. Shugart left IBM in 1969, and in 1976 his company, Shugart Associates, introduced the minifloppy (5.25-inch) disk drive. It, of course, became the standard eventually used by personal computers, rapidly replacing the eight-inch drives. Shugart also helped create the Shugart Associates System Interface (SASI), which was later renamed small computer system interface (SCSI) when approved as an American National Standards Institute (ANSI) standard.
Sony introduced the first 3.5-inch microfloppy drives and disks in 1981. The first significant company to adopt the 3.5-inch floppy for general use was Hewlett-Packard in 1984 with its partially PC-compatible HP-150 system. The adoption of the 3.5-inch drive in the PC was solidified when IBM started using the drive in 1986 in some systems and finally switched its entire PC product line to 3.5-inch drives in 1987.
In 2002, many companies started selling systems without floppy drives. This started with laptop computers, where internal floppy drives were first eliminated and replaced with external (normally USB) drives. By 2003, virtually all systems sold, be it desktop or laptop, no longer included a floppy drive, although sometimes you can purchase an external USB model as an option. An optional USB floppy drive can be used as a bootable drive if the BIOS permits it, as is the case with most recent systems.
Tape Drives
Tape drives and media were once a somewhat popular form of magnetic storage for backup use. Although the drives were expensive, the tape media was cheap, allowing multiple backup sets to be inexpensively created. As hard drive capacities increased, however, the capacity of tape media could not keep pace, and using multiple tapes to back up a single drive meant time-consuming and error-prone media swaps. The performance of tape drives also suffered in relation to hard disks, greatly increasing the time it took for a backup to complete. Hard drives also become much less expensive, such that it was cheaper and easier to simply purchase more hard drives for backups. Over time, all these factors have caused tape backup drives and media to no longer be suitable for standard desktop or laptop PC backups. Currently, tape drives and media are only used for high-end server backups.
The most common types of tape backups in use today include LTO Ultrium 5 (with a native/compressed capacity of 1.5/3.0 TB), LTO Ultrium 4 (800 GB/1.6 TB), LTO Ultrium 3 (400/800 GB), and SDLT (160/320 GB).


