Lecture 6: Data Storage Technology

Endian

refers to a computer architectures byte order, or the way a computer stores the bytes of multiple-byte data element.

SDRAM Latency

refers to the delays incurred when a CPU tris to access data in SDRAM; in with latency is often in RAM memory clock speed.

SO-DIMM

small outline dual inline memory module--meant for laptops. Essentially this is a type of computer memory which is, in fact approximately 2/3 the size of desktop comp equivalent. Used in systems with limited space.

Solid State Drive (SSD)

solid state drive uses one of various types of high-speed flash memory instead of magnetic media. Mimic behavior of magnetic disk drive but uses flash RAM or other NVM devices as the storage medium and read/write mechanism. No moving parts and electronic circuitry is built with semiconductors.

RAM Types

static RAM (SRAM) and dynamic RAM (DRAM)

Volatile

storage device or device loses its data when power is removed. Primary storage devices are volatile. RAM vanishes as soon as the system is turned off.

Non-Volatile

storage device or medium loses its data when power is removed. Secondary storage devices are generally non volatile.

Little endian (intel)

store the least significant byte (LSB) at the lowest memory address.

Big endian (motorola)

stores the most significant byte (MSB) at the lowest memory address.

Cache Memory

take advantage of Spacial Locality. Take advantage of temporal locality. Quick info is stored in the cache.

Detect Disk Errors in RAID

the data recorded on disks is stored in sectors. Each sector has a frame check sequence that can detect one or more errors in a sector. Ability to detect errors means that RAID array can rapidly respond to the failure of an individual drive unit.

Mirrored Array

two or more disks contain the exact same blocks of data

Striped Array

A file segment is divided into stripes (blocks) and stored on different disks.

USB Flash Drive

A small, portable device that can fit on a key chain, used to carry data. Can hold up to several GB or storage Capacity. Have less storage capacity than external hard drive. Do not contain any moving parts.

Summary

A typical computer system has primary and secondary storage devices. The critical performance characteristics of primary storage devices are their access speed and the number of bits that can be accessed in a single read or write operation. Programs generally are created as through they occupied contiguous primary storage locations starting at the first location.

Hybrid Drive (SSHD)

Also called a Solid State Hybrid Drive (SSHD), these are a compromise between a magnetic HDD and an SSD. They are faster than an HDD but less expensive than an SSD. They are a magnetic HDD with an onboard SSD serving as cache. The SSHD drive automatically caches data that is frequently accessed.

RAID 1

Also called mirroring, this RAID array type provides fault tolerance because all the data is written identically to the two drives in the mirrored set. Data security is increased b/c you can remove one of the drives without losing data. Level 1 RAID system increases data security and improves access time. The write time is longer of the two parallel writes. Most accesses are reads rather than writes. it is possible to cache writes and allow the disks to do writing when it is free. 1 RAIDS are expensive.

Parallel Access

An access method that can access multiple storage locations simultaneously; can also be achieved by subdividing data items and storing the component pieces on multiple storage devices.

DDR2 SDRAM Modules

DDR2-400 (PC2-3200): 400 Million, DDR2-533 (PC2-4200): 533 Million, DDR2-667 (PC2-5300): 667 Million, DDR2-800 (PC2-6400): 800 Million, DDR2-1066 (PC2-8500): 1066 Million

DDR SDRAM1 Modules

DDR266(PC-2100): 266 Million, DDR333(PC-2700): 333 Million, DDR400(PC-3200) 400 Million

DDR3 SDRAM

DDR3 increases prefetch buffer size. The primary benefit of DDR3 is the ability to transfer twice the data rate of DDR2, thus enabling higher bus rates and higher peak rates than earlier memory technologies. Data bus can run @ four times the speed of internal memory clock. achieves a total of 8 data transfers /memory clock cycle. Provides reduction in power consumption compared to DDR2. DDR3 not compatable with earlier types of SDRAM memory.

DDR 3 SDRAM Modules

DDR3-800 (PC3-6400): 800 Million, DDR3-1066 (PC3-8500): 1066 Million, DDR3-1333 (PC3-10600): 1333 Million, DDR3-1600 (PC3-12800): 1600 Million, DDR3-1866 (PC3-14900): 1866 Million, DDR3- 2133 (PC3-17000) 2133 Million

DDR4 SDRAM

DDR4 keeps the prefetch buffer size the same as for DDR3, but uses bank groups. Two prefetches can be operating in parallel in the two bank groups. With DDR4, up to 4 bank groups can be used.Data rates range from 1600Mbps to 3200Mbps, compared to 800Mbps to 2133Mbps for DDR3. Not compatible with any earlier type sdram memory.

DDR 4 SDRAM Modules

DDR4-1600 (PC4-12800): 1600 Million, DDR4-1866 (PC4-14900): 1866 Million, DDR4-2133 (PC4-17000): 2133 Million, DDR4-2400 (PC4-19200): 2400 Million, DDR4-2666 (PC4-21333): 2666 Million, DDR4-2933 (PC4-23466): 2933 Million, DDR4-3200 (PC4-25600): 3200 Million

RAID 6

Disk striping with parity. RAID-6 uses four or more disks and provides fault tolerance. It can survive the failure of two drives.

Changing BIOS

Most modern PCs have a flash BIOS recorded on a flash RAM (flash ROM) which can be updated if necessary. When device details change, only BIOS needs to be changed.

Increasing Data Rate

SDRAM significant improvement on asynchronous RAM, it still has shortcomings that unnecessarily limit that I/O data rate that can be achieved. Double-data rate DRAM (DDR DRAM) provides several features that dramatically increase data rate. Four Gens: DDR1, DDR2, DDR3, DDR4

Speed (cont)

Secondary storage speed also influences execution speed. Many apps need access to large databases in the secondary storage to support ongoing processing. Program response time in these systems depends on secondary storage.

Serial Access

Stores and retrieves data items in a predetermined, ordered sequence. Data is written to a linear storage medium in a specific order. Info can be accessed only after first getting past the intervening data. Ex. magnetic tape must be rewound or fast forwarded to get to appropriate position.

Primary Storage

The computer's main memory, which consists of the random access memory (RAM), cache memory, and the read-only memory (ROM) that is directly accessible to the CPU. Examples include: register, cache, RAM, ROM.

RAID Levels

The different types of RAID, such as RAID-0, RAID-1, RAID-2, and so on. Various levels provide different functionalities; some emphasize speed and some reliability.

Addressable Memory

The highest numbered storage byte that can be represented in a CPU or computer; usually determined by the number of bits used to represent an address. (physical memory is usually smaller)

Increasing Data Rate (cont)

achieves higher data rate in 3 ways: The data transfer is synchronized to both the rising and falling edges of the clock, rather than just the rising edge. This doubles the data rate; hence the term double date rate. DDR uses higher clock rate on the data bus to increase the transfer rate. A buffering scheme is used.

Physical Memory

actual number of memory bytes that physically are installed in the machine.

Electronic Storage

also called semiconductor storage. Effective for quick and effecient information storage. Can save data quickly without lengthy burn processes required for cd's. Flash memory (usb flash drive, memory card).

Non-Volatile RAM

any RAM device that can hold content without continuous power flow. Unsuitable for primary memory. Slow rewrite time compared to DRAM. 4

RAID 2,3,4

arrays that are striped in different ways.

BIOS for System Booting-up

bootstrap loader is a program that resides in ROM, EPROM or other non-volatile memory. Automatically executed by the processor when turning on the computer. The bootstrap loader has been replaced in computers that have an Extensible Firmware Interface (EFI) and is part of the EFI BIOS. When the computer is turned on, the CPU passes control to the BIOS program. When BIOS boots up (starts up) the computer, it first determines whether all of the attachments (devices) are in place and operational and then it loads the operating system (or key parts of it) into RAM from hard disk or diskette drive.

Hybrid RAID 0+1

combines features of levels 0 and 1, by providing both fast data access and protection against drive failure.

Hard Drive Disk (HDD)

contains one or more spinning metal or glass platters that are coated with a magnetic substance. Typically 3.5 inches wide. Storage ranges from 500gb to 4 TB or larger

DRAMs

contents must be refreshed several times per second. Most common, cheap. Volatile. Although capacitors can store electricity, they cannot do so indefinitely. Information eventually fades unless capacitor charge is refreshed periodically. Only one transistor and a capacitor are required to store one bit information. This allows DRAM to reach very high density and to be used for main memory

Cost and Capacity

cost increases with more speed, volatility, or portability. As access method moves from serial to random to parallel access. Primary storage is expensive (highspeed and combo of parallel/random access). Capacity of secondary storage devices is greater than primary storage devices.

Portability

data can be portable by storing in a removable storage medium or device. cd/dvd, floppy disk. Typically implemented in two ways. Portable devices typically have slower access speed than permanently installed devices and those with removable media.

Storing Electrical Signals

data is repped as electric signals. Digital signals are used to transmit data to and from devices attached to system bus. Storage devices must accept electrical signals as input and output.

Memory Packing

dual inline packages, single inline module, double inline memory module

RAID-1 (cont)

efficiency of two disk RAID 1 system is 50% b/c data is simply duplicated. Provides excellent fault tolerance. If drive fails, can continue working normally. Just remove the failed drive and install the new one, then rebuild the lost data.

RAID 0 Advantage

high throughput. No capacity is lost due to redundancy and storage efficiency is 100%. Since stripes are accessed in parallel, an n drive raid 0 arrays as a single large disk with data rate n times higher than the single-disk rate. offers the best performance in reading and writing, particularly if seperate controllers and caches are used for each disk.

Data Transfer Rate

in bps(bits per second). Amount of data that moves per second.= 1 second/ access time X the block size. The result is a number stated in data units per sedond, for example, 100mbps(megabits), or 10 MBps (megabyts).

Disk Arrays

in large computer environments, it is common to group multiple disks (HDD, SSD, etc) together.

BIOS

initial boot instructions and diagnostics in a program used to get computer started. contains info like # or hard disk drives, disk drive type etc. contains the code required to control keyboard, display screen, disk drives and misc functions.

Storage for BIOS

initially program storage was ROM. then came along PROM (programmable rom) that can be programmed by the customer.

Cloud Storage

internal and external hard drives. Solid State storage are all location dependent: data is written to sectors on specific physical devices. Dropbox, iCloud, Google Drive and one drife all are examples of cloud storage platforms.

Why is ROM Needed?

non-volatile memory. Typically used to store the initial program instructions used to start a computer or device when it is first booted up. ROM would hold a small startup program called BIOS in order to start the OS. Non-volitile primary storage used for this purpose is called ROM.

ROM (Read Only Memory)

non-volitile memory used to hold software that is not expected to change over the life of the system. Most ROM types are also capable of rewriting for updating of held data and programs.

Secondary Storage

not directly accessible by the cpu. Computer usualy uses its input/output channels to access secondary storage and transfers the desired data using intermediate area in primary storage

Random Access

not restricted to any specific order when accessing data. RAM is a random access device. Disk storage (hard sidk, ZIP, floppy, cd/dvd), flash memory, and solid state drives.

RAID 0

not true RAID. no error checking or redundancy, but data is placed across all drives for increased speed. Does not use redundancy to provide adtnl security. Two or more disks are working as a group, sharing data sent out to the disk controller, sending a stripe of data to each disk in the group at the same time. Each stripe is different and forms the overall file. There are massive performance benefits.

DDR2 SDRAM

offers greter bandwidth and density in a smaller package along with reduction in power consumption. Increasing the size of the prefetch buffer between the DDR2 and the data bus so that the data bus can be run at twice the speed of the memory clock. DDR2 achieves 4 data transfers per memory clock cycle. Not compatible with DDR1-SDRAM memory

Primary Storage Charactaristics

performance characteristics are access speed and data transfer unit size. Must closely match CPU speed and word size to avoid wait states. Types: registers, cache, RAM and ROM

Primary Storage vs Secondary Storage speed

primary storage is typically much faster. expressed in nanoseconds (billionths of a second). Secondary storage is expressed in milliseconds (thousandths of a second)

Read Ahead Memory Access

program usually access instructions and data items sequentially, called Spatial Locality. (if a particular memory location is referenced at a particular time, then it is likely that nearby memory locations will be referenced in the near future.)

RAID 0 Disadvantage

no fault tolerance. a loss of any drive in the array renders all data invalid. RAID-0 only makes sense if the data is frequently backed up.

Disk Storage

Falls under 2ndary storage. a read/write head passes over the spinning disk. Reading or writing info as appropriate. zip, floppy drives, dvd etc.

M.2 Connector

Formerly known as the NGFF after mSATA. Current connector standard for mobile SSDs, although it has also been adopted by motherboards as well. Can plug both PCI-Express based and SATA-based SSDs, but is generally PCI expressed based only.

Comparing Various Types of Storage Media

Hard disks provide very fast recording and access to info and can hold a lot of data. B/c they can be written over and over again without encountering the limitations of CDs or DVDs, hard disks are currently the primary storage medium for computers. Optical disks are portable and lightweight. Not affected by magnetic fields. Electric storage media are compact, fast and versatile.

New Nonvolatile Memory

Magnetoresistive RAM (MRAM) stores data in magnetic storage elements, Phase-change memory (PCM)--stores data by changing the sate of the material used. Can change back & forth between amorphous and crystalline states.

Access Storage

Primary & Secondary Storage

RAID Levels 2 to 6

RAID levels 2 to 6 all distribute data in the form of stripes like level 0, but they do not fully replicate the data like level 1. In other words, levels 2 to 6 fall between the extremes of level 0 (no redundancy) and level 1 (redundancy by replication).

Synchronous Read Operations

RAM has a synchronous interface. It waits for a system clock signal before transmitting data (reading or writing) and therefore.runs in sync with the computer's system bus (the connection between the CPU and other components on the motherboard). Ex SDRAM-only one word of data is read or written per system clock cycle (single data rate).

RAM Capacity

RAM is the temporary storage used by apps run by the CPU. Amount of RAM used by an app varies according to whether the app is idle or acitve. More memory in a device, more programs can be run @ the same time and larger data files can be stored.

DDR SDRAM (double data rate SDRAM)

Transfers data twice, instead of once, for each clock cycle Faster than SDRAM. I/O bus uses the same clock rate as the memory chip. Transfers data on both rising (0->1) and falling (1->0) edge of data bus clock pulse. Reads or writes two words of data per system clock cycle. Typical clock rates are 133, 166 and 200 MHz.

RAID 5

Uses block-level and parity data striping. Uses one parity block on each disk. Uses most simple XOR

RAM Access Time

Within the instruction fetch or execute cycle, the slowest steps are those that require memory access. memory is slow compared to CPU processing speeds. CPU must wait for the read/write operation to complete. A wait state is a CPU cycle spent waiting for a storage access or I/O operations to finish.

RAID (redundant array of independent disks)

a group of two or more integrated hard disks to give either/both fault tolerance or a performance boost. Redundant implies degree of fault tolerance. Today most PCs have motherboards with RAID support.

Striping

key concept in RAID tech is striping. Disk space is divided into units called stripes (blocks) that may be as small as a single sector or as large as several megabytes. Strips are spread out across several disks in parallel. The way in which data is allocated to the stripes, and the stripes allocated to the individual drives determines the level of the RAID system.

Physical Memory Organization

main memory can be regarded as a sequence of contiguous, or adjacent memory cells. Each memory location has a unique address.

Speed (cont1)

measured by access time and data transfer rate. Access time is the average time required to execute one complete read or write operation in nanoseconds.

RAM Charactaristics

microchip implementation, ability to read and write with equal speed. Random access to stored bytes, words or larger data units. Saving information to RAM is referred to as "writing to RAM," and retrieving RAM contents is called "reading the RAM".

SRAM (Static Random Access Memory)

more expensive than DRAM. Faster since only uses transistors. Significantly less power usage. Used where bandwidth or low power, or both are principle considerations. Still volatile. SRAM is less dense, therefore not used for high capacity, low-cost applications such as main memory. Small capacity, so used more for registers and cache memory for high-speed access used.

Endian-ness

most UNIX machines are big endian, whereass most PCs are little endian machines. Most newer RISC architectures are also big endian.

Speed

most important characteristic differentiating primary and secondary storage. CPU continuously needs data to keep it busy. Primary storage extends limited capacity of CPU registers. Register location enables zero wait states for access. Executable code is copied from secondary to primary storage. CPU moves data instructions between registers and primary storage.

Use RAID

multiple disk drives are operated in parallel and the bits of a single file are divided between the individual drives. Can improve performance of disk systems or to improve reliability. Important in applications where data security is vital. If info is dist. between disks of an array, the failure of one disk does not cause the system to crash.

Solid-State Drives (SSD) cont.

use microchips and data is retained in non-volitile memory chips. Like HDD's SSD's are typically less susceptible to physical shock, perform much faster than a spinning hard drive, and have lower power consumption.