COVER FEATURE A Survey of Storage Options

Over the past 50 years, technological innovations have continually driven down storage costs while dramatically increasing performance and value—and the trend shows no signs of slowing.

John P. f you were a personal computing enthusiast in In 1970, Fairchild Corp. invented the first 256- Scheible 1984, you could buy an IBM PC for $5,000 to Kbyte SRAM; by 1972 it was the best-selling semi- IBM $6,000—about $8,000 to $9,500 in today’s cur- conductor memory chip in the world, superseding rency. For this you would get a 0.004-GHz 8- the magnetic core memories that had I processor with 0.064 Mbytes of RAM, a 12- used since 1947. inch monochrome text-only display, a 0.16-Mbyte Both types of RAM are volatile—they lose their floppy drive, and no possibility of a hard drive. contents when the power shuts down. DRAM Compare that with a PC purchase today: a 2- requires thousands of refreshes per second. SRAM GHz 32-bit processor with 2,000 Mbytes of RAM, is faster because it does not need to be refreshed. If a 128-Mbyte video graphics card, a 700-Mbyte all things were equal, main memory would use CD-RW (and maybe a DVD-R), a 17-inch 1600 × SRAM, since it is faster than DRAM and does not 1200 display with 32 million colors—all for maybe require refresh circuitry to protect from cor- $1,500. The old machine cost six times as much for ruption. However, SRAM is physically larger than 1/2000th of the processing power. The purchasing DRAM and it costs considerably more, making it power gain over 18 years is 1,200,000 percent. impractical for main memory, although it is still use- ful for relatively small caches. THE FORMS OF STORAGE Asynchronous and synchronous DRAM. DRAM can be Very few industries have shown such dramatic either asynchronous or synchronous. With an asyn- increases in value and function and simultaneous chronous interface, the processor must wait idly for reductions in cost as the computer storage indus- the DRAM to complete its internal operations, which try. The technology has evolved continuously over typically takes about 60 nanoseconds. With syn- the past 50 years both across different media and chronous control, the DRAM latches information within each one. from the processor under the system clock’s control. We can categorize primary storage as volatile ran- Years ago, most PCs came with asynchronous dom access memory, which loses its data upon fast-page-mode DRAM, which ran at speeds be- power loss, and nonvolatile memory, such as flash tween 80 and 100 ns. Extended-data-out DRAM and read-only memories, in which data persists. improved speed by about 20 percent. However, Data transfer forms include floppy disks, CDs, escalating CPU and speeds out- , and flash memory cards. Fast random access stripped the ability of both FPM and EDO to deliver forms include disk drives, , and the data in a timely manner. As system speeds increased, slower DVD-RAM. Data archive forms include especially when 66-MHz memory buses became tape, CD, and DVD. standard on PC , FPM and EDO DRAMS dragged effective speeds down by forcing Random access memory CPUs to wait to receive data from memory. The two basic types of RAM are dynamic RAM Once it became apparent that bus speeds would and static RAM. IBM’s Robert H. Dennard created have to run faster than 66 MHz, DRAM designers the one-transistor, small-capacitor DRAM in 1966. needed to overcome the significant latency issues

42 Computer 0018-9162/02/$17.00 © 2002 IEEE that still existed. They did so by implementing a The bus, or data path, consists of four 8- synchronous interface, which also offered other bit paths, allowing four separate data IBM and 3M advantages. requests at the same time. Because the mem- introduced tape The current standard, Synchronous DRAM, is ory modules are intrinsic parts of the data synchronized to the system clock. RAM accesses path, none of the three memory sockets can 50 years ago in occur in burst mode, which means that memory remain empty. Therefore, a socket without 12-inch reels accesses occur in bunches, with the first of the memory must hold a continuity module, a that stored bunch requiring more time to allow for setup. special board that carries the clock and data 1.4 Mbytes of data. SDRAM is tied to the system clock and is rated by signals. Unlike earlier memory modules, megahertz instead of nanoseconds. It must be rated which had to be the same size, to run at least as fast as the system bus—preferably allows each socket to hold a different-sized a little faster. module. Current designs. There are many different memory Other features that distinguish Rambus from designs today. Synchronous Link DRAM uses pack- SDRAM technology include a lower voltage re- ets for address, data, and control signals to operate quirement (2.5 V as opposed to 3.3 V), a much on a faster bus than standard SDRAM—up to at higher clock frequency (200 MHz to start, projected least 200 MHz. SLDRAM operates the output sig- to move rapidly to 800 MHz and 1.6 GHz), and nal at twice the . This puts the output greater difficulty in manufacturing. operation as high as 400 MHz, with some engineers The manufacturing problem revolves around the claiming it will reach 800 MHz in the near future. tight tolerances that Rambus requires. Timing and SDRAM—also known as data transmission signals must travel at precise rates. SDRAMII—does precisely what the name implies. This requires all traces to be exactly the same length We can represent a clock cycle by a square wave. with the same impedance. Such precision is difficult SDRAM uses only one of the wave’s edges, to mass manufacture, and it accounts, at least partly, but DDR SDRAM references both to effectively for the long delays in Rambus product releases. double the data transmission rate. Unlike 168-pin The immediate competition in memory technol- SDRAM, DDR SDRAM uses a 184-pin plug. ogy will focus on Rambus versus SDRAM. Ram- Although DDR SDRAM does not require changes bus must overcome manufacturing and cost battles in the basic motherboard technology, it is not back- in time to compete. Otherwise, ongoing innova- ward compatible on motherboards designed for tions in SDRAM technology ensure its continuing SDRAM. market lead. Compared with DDR SDRAM, SLDRAM seems to offer a much better alternative. Its actual clock Flash memory speed is lower, reducing signal problems. Latency RAM’s volatility makes it suitable only for tem- timings are shorter. So are costs due to the royalty- porary storage. Flash memory, on the other hand, free design and operation on current bus designs. retains its contents when power is lost, making it SLDRAM’s is also much higher than ideal for storage. Flash memory, the essential com- DDR SDRAM at 3.2 Gbps versus 1.6 Gbps. ponent in solid-state hard disks, retains data with- SLDRAM is an adopted by the out a power source. Flash memory has a size IEEE. It uses a narrow 16- or 18-bit bus with packet advantage (much smaller than a floppy, CD, or disk addressing that effectively increases the bandwidth. drive), making it the preferred storage option for The technology focuses not on the memory itself, small portable devices such as cell phones and dig- which is largely unchanged from standard DRAM, ital cameras. but on the interface, or protocol, which makes mul- However, flash memory is also relatively slow to tiple simultaneous accesses possible, as long as they write, which limits its applications. If it can over- are all to different physical locations in the RAM. come the write delays, which is likely, flash mem- The current battle for memory market share cen- ory can continue to expand its market. ters on DDR SDRAM and Rambus. RDRAM versus SDRAM. Rambus is a new system Tape design that requires major changes in the bus struc- IBM and 3M introduced tape 50 years ago in 12- ture and the way the RAM carries clock signals. inch reels that stored 1.4 Mbytes of data. Today Also known as RDRAM (Rambus DRAM), high-end tape holds 1 Gbyte of data. Rambus traps the clock signal, rather than broad- Recent price decreases and capacity increases in casting it, as in previous SDRAM designs. disk drives have made them competitive with tape

December 2002 43 RAID Storage

IBM received the first patent for a subsystem in 1978. The company worked with the University of California, Berkeley, to define levels for redundant array of independent disks and released the initial RAID definitions in 1987. Although RAID is not storage in itself, it does ciently over large distances for purposes change the nature of storage by adding tolerance of disk drive failure and allowing for secure and centralized backup, espe- changing the access times to read and write data. cially for disaster recovery purposes. As the “RAID The major levels are RAID-0 through RAID-6. Storage” sidebar indicates, data managers can increase reliability by using RAID—redundant array • RAID-0 (striping) accesses all drives in parallel. It provides high of independent disks. Although RAID does increase data rates, but the data rates come at the expense of reliability. reliability, it does not protect against disasters like Because RAID-0 writes the data across multiple drives, it provides the terrorist attacks of 11 September 2001, when no redundancy. If one drive fails, all the data in the array is lost. only remote backup would allow customers to come • RAID-1 (mirroring) provides high reliability by writing or read- back online within a few days. ing the same data on two or more drives, but the reliability comes IBM demonstrated a 1-terabyte cartridge in 2002 at a price at least twice that of RAID-0. RAID-1 reads faster than that is one prospect for tape’s future. The technol- non-RAID applications, since the first drive to respond to a request ogy, which uses an ultrathin layering technique will provide data, thus reducing latency. from Fuji, will likely be available in products in • RAID-2 requires the use of nonstandard disk drives and is there- three to five years. fore not commercially viable. Double-sided tape is also a prospect, but devel- • RAID-3 (parity) provides extra reliability at the expense of ran- opers must solve technological issues related to tape dom access performance. RAID-3 uses one extra drive to store the thickness and the electrostatic discharge that causes parity, or error-correction, data. If one drive fails, although per- tape to stick. formance will suffer, RAID-3 can recover the data and continue Work is also under way for write-once/read-many processing on the other drives until the failed drive is replaced. tape for in more limited markets. RAID-3 has very high data rates, since writing and reading occur Currently, tape solutions use 50-Gbyte cartridges in parallel. It has higher reliability than RAID-0 (striping) and is (uncompressed). IBM has demonstrated 1-Tbyte car- less expensive than RAID-1 (mirroring). tridges (uncompressed). Although DVDs will remain • RAID-4 (parity) provides extra reliability at the expense of write competition for low-end tape and archival media, random-access performance. RAID-4 is like RAID-3 except that the high-end applications will need tapes with larger it uses blocks instead of bytes for striping and it uses a dedicated capacities and faster access speeds. . Going from byte to striping improves random access performance compared with RAID-3, but the dedicated Hard disk parity disk remains a bottleneck, especially for random write per- Like tape, the hard disk came on the scene almost formance. , format efficiency, and many other 50 years ago, when IBM introduced the RAMAC attributes are the same as for RAID-3 and RAID-5. (random-access method of accounting and control) • RAID-5 (parity) uses blocks of data and parity information striped to replace punch cards. The vacuum-tubed RAMAC across all drives. RAID-5 helps eliminate the performance bottle- took the space of two refrigerators, weighed a ton, neck at the parity drive. and stored 5 million characters (not bytes, but 7-bit • RAID-6 (dual parity) provides very high performance by tolerating characters that totaled less than 5 Mbytes of stor- two disk drive failures, but it does so at the expense of write perfor- age) on 40 large aluminum disks coated with a vari- mance and cost. RAID-6 stripes data on a block level across a set of ation of the paint used on the Golden Gate Bridge. drives, just as RAID-5 does, but the controller calculates a second set To put this in perspective, a 2.5-inch laptop hard of parity that it writes across all the drives. RAID-6 provides extremely disk could store the same amount of data as 12,000 high data fault tolerance, and it can sustain multiple simultaneous RAMACs—and for a lot less money. drive failures. However, RAID-6 requires an additional drive for par- In 1962, removable disk storage entered the ity, adds controller overhead, and has very poor write performance. scene with a storage capacity of 2 Mbytes per removable disk stack. In 1973, IBM introduced the 3340 Winchester drive, featuring two spindles of for long-term storage, but hard disks do not match 30 Mbytes each. tape’s portability and shelf life. CDs and DVDs are The disk drive cost per Mbyte has been in freefall. portable, but CDs lack tape’s capacity. It waits to In 1994, the cost of a Small Computer System be seen if the various standards for DVD will allow Interface drive was about $2.00 per Mbyte. Today, a common DVD library approach to tape replace- the cost of an Advanced Technology Attachment ment—not that tape has ever needed a common drive is around $0.001 per Mbyte, 1/2000th of the format. earlier cost. Wide and local area networks are increasing the Disk drive reliability has also increased dramat- market for tape by allowing data to be shipped effi- ically. In 1994, IBM introduced the first 3.5-inch

44 Computer Table 1. DVD formats available today. Name Format Data capacity Video capacity

DVD-5 SS/SL 4.7 Gbytes 2+ hours DVD-9 SS/DL 8.5 Gbytes 4 hours disk drive with a mean time to failure of 1 million DVD-10 DS/SL 9.4 Gbytes 4.5 hours hours, comparable to today’s drives. DVD-14 DS/multiple layer 13.24 Gbytes 6.5 hours In 1980, thin film heads allowed a transfer rate DVD-18 DS/DL 17 Gbytes 8+ hours of 3 Mbytes per second with an areal density of 12 DVD-RAM SS/SL 2.58 Gbytes n/a Mbytes per square inch. In 1993, IBM introduced DVD-RAM DS/SL 5.16 Gbytes n/a the world’s first 1-inch-high 1-Gbyte disk drive with DS – double sided an areal density of 354 Mbytes per square inch. DL – double layer By 1989, the magnetoresistive head allowed an SL – single layer areal density of 1,000 Mbytes per square inch. In SS – single sided 1995, IBM demonstrated an areal density of 3,000 Mbytes per square inch. Just six years after IBM dis- covered the giant magnetoresistive, or GMR, effect, The software distribution business has aban- the company introduced the spin-valve head, which doned the floppy in favor of CDs. The space con- is five times more sensitive than the best commercial straints for laptop computers will speed the floppy’s heads. GMR technology, however, holds even greater demise, as will the Internet’s support for data trans- promise for the future at 10 Gbits per square inch. fers. Only Apple Computer has been fearless The has a solid place in the ran- enough to sell machines without a floppy drive, but dom access storage area for cost, reliability, and the floppy will soon disappear in favor of USB flash performance, but its portability and shelf life are and CD-RW devices. concerns. Nevertheless, the seemingly endless inno- vations in this area indicate that designers will find CDs and DVDs a cost-effective solution to these concerns. For Optical drives are available in either proprietary example, the microdrive is a 1-inch disk drive that formats or the more common CD and DVD formats. holds 1 Gbyte and could possibly expand the role In 1980, IBM used magneto-optical recording and of disk drives in certain applications. blue-wavelength gas lasers to achieve an areal den- sity of 1,000 Mbytes per square inch. In 1991, it Floppy drives introduced the first 3.5-inch rewritable optical drive. IBM introduced the floppy drive in 1970 as a The CD format took the computer industry by read-only device to hold microcode and diagnos- storm, providing 80-minute audio capacity and tics for large mainframe computer systems. This 700-Mbyte data capacity on inexpensive media first commercial floppy drive used 8-inch diskettes running on inexpensive players. The original read- to record less than 0.10 Mbytes of data. In 1973, only CD player was replaced by the write-once CD- IBM introduced an 8-inch floppy drive with R and then by the rewritable CD-RW. read/write capability and a 0.25-Mbyte capacity. Table 1 lists the basic DVD formats available In 1976, Shugart Associates and Dysan Corp. intro- today. The DVD-5 and DVD-10 formats are the duced a 5.25-inch floppy diskette that initially most common. Although DVD-9 offers extra stored less than 0.10 Mbytes, later increasing it to capacity without having to flip the disc over, it uses 0.16 Mbytes in 1981 (IBM PC), 0.36 Mbytes in a semitransparent layer of data in addition to the 1983 (IBM XT), and 1.2 Mbytes in 1984 (IBM AT). standard reflective layer, which makes it expensive In 1980, Sony introduced a 3.5-inch floppy drive to manufacture. Nevertheless, most DVD players that stored 0.72 Mbytes, and in 1987 the 1.44- and DVD-ROM drives support the format. Mbyte diskette was available. IBM developed a Despite predictions, DVDs have not replaced double-capacity 2.88-Mbyte diskette in 1988, but CDs for software distribution. However, DVD- compatibility problems—even when the computer ROM drives are steadily replacing CD-ROM wrote to it in compatible 1.44-Mbyte mode—pre- drives in computers, and many forecasters expect vented its wide adoption. manufacturers to stop making CD-ROM drives This 20-year-old technology is still with us today, soon. Most computers with DVD-ROM drives can but its 1.44-Mbyte capacity sharply limits its use- also play DVD video disks, CDs, and CD-ROMs. fulness and hastens its replacement by 700-Mbyte DVD-ROM also includes rewritable variations, CDs and CD-Rs. Recently, the Universal Serial Bus such as DVD-R. A super-high-fidelity DVD-audio flash device has also invaded the floppy market. format is available but not widely supported. Although it is more expensive than a floppy or CD, DVD-ROM drives are inexpensive—about $60— a 64-Mbyte flash memory device can plug into a and can read both video DVDs and DVD-R disks, USB port and act like a hard/floppy drive. as well as CD, CD-R, and CD-RW.

December 2002 45 The picture for rewritable DVDs is much cloudier • DVD-RW, primarily for the video market. It is with incompatible formats. Most rewritable drives compatible with many existing DVD-ROM cost between $300 and $600, with media ranging drives and DVD video players but has limited from $4 to $12, but the incompatible formats cause sequential writability and only 1,000 write problems. cycles. DVD-RW, which is primarily supported The rewritable formats are: by Pioneer, can read and write to DVD-R. • DVD+RW, primarily for the video market. It • DVD-RAM, primarily a storage device in which is compatible with most existing DVD-ROM a cartridge protects the media. DVD-RAM sup- drives and DVD video players. Sony, Philips, ports random write and 100,000 write cycles, Hewlett-Packard, Mitsubishi/Verbatim, Ricoh, but it is incompatible with most DVD-ROM and Yamaha support it. drives and DVD video players. Essentially, it is • DVD+R, a cheaper write-once form of like a removable hard disk that can also read DVD+RW. DVD+RW can write to DVD+R. DVD-ROM disks. Panasonic, Toshiba, Hitachi, and other vendors support it. In the rewritable DVD market, DVD-RAM is use- • DVD-R, primarily for DVD production use. ful for data applications, for which cartridge pro- The media is compatible with many existing tection, write-many-times capability, and true DVD-ROM drives and DVD video players. It random writes are essential but compatibility is not. is also compatible with DVD-RW but uses less In the R/W video market, however, the incompati- expensive write-once media. bilities add risk when choosing a storage technology.

BM has developed Millipede, a prototype micro- electromechanical system that provides a ter- I abyte of storage on a single chip the size of a postage stamp. The storage medium is a thin poly- mer film on the chip surface that stores as 10- nanometer-diameter holes. Millipede, which features high-density nonvolatile erasable memory REACH suitable for portable digital systems, targets the flash memory market. Although Millipede is at least two years from fab lines and four years from production, the current HIGHER prototype has successfully erased a single bit more than 100,000 times by heating the polymer and thus causing it to flow back into the hole. The tips of the Millipede device used to read and write the Advancing in the IEEE Computer Society can elevate your standing in the profession. media are sculpted into the silicon and write by heating to 400°C. They read by heating the film to Application to Senior-grade membership 300°C, which does not melt the polymer, and recognizes detecting the cooling effect of dropping into a hole. ✔ ten years or more of professional Millipede is just the latest in a long line of stor- expertise age technology innovations that we can expect to Nomination to Fellow-grade membership continue into the future. I recognizes ✔ exemplary accomplishments in computer engineering John P. Scheible is a senior engineer and scientist at IBM. He received BS degrees in computer sci- GIVE YOUR CAREER A BOOST ence and electrical and computer engineering from Oregon State University. He is a member of the UPGRADE YOUR MEMBERSHIP International Committee for Information Tech- nology Standards T10 and T11 committees on I/O computer.org/join/grades.htm and device-level interfaces, respectively. Contact him at [email protected].

46 Computer