General Knowledge on Storage Systems

Basic terms in relation to storage systems are described here, mainly for the terms used in this Web site.

* These terms are listed in alphabetical order.

Term Description
Access The term "access" in this case means to read, write, or update information (data) on some recording media such as disks or tapes.
AFP AFP stands for "Apple Filing Protocol," which designates a protocol used by AppleShare that enables file sharing in an AppleTalk network environment for Macintosh computers.
Archiving Broadly speaking, archiving means to store important information for a long period of time in some recording media such as magnetic tapes.
Array Controller An array controller included in a disk array system is a section of the disk array to control RAID operation being instructed by host systems. It is also called a director (DIR).
Average Response Time This is the average time between the instant when disk array devices or magnetic tape devices receive an instruction from a host and the instant that completion status is reported to the host.
Backup Just in case original data are crashed, damaged, or erased by accidents, data copies are saved (backed up) in a separate device other than the recording device containing the original data. Removable media such as magnetic tapes are often used for storing backup data copies.
Bridges Bridges are used in a network to relay data that flow inside cables. For example, an FC-SCSI bridge unit relays between media of different physical layers within a network, i.e., between fibre channels and electrical SCSI interfaces.
Cache Memory Cache memories are used in NEC Storage systems for temporarily retaining data in order to achieve high speed read/write operations. Furthermore, as an optional feature, a logical disk can be set up inside this cache memory so the data inside this cache-resident disk can be accessed at a cache hit speed (Available for S3000/S4000 Series systems).
CIFS CIFS stands for "Common Internet File System," which is a network protocol used by Windows NT or Windows 95/98 clients for issuing file access requests to Windows servers.
Crossbar Switch Crossbar switches were used in early telephone switching system. With a mechanical crossbar switch, many conductor bars each attached with a number of electrodes are arranged in a crossed configuration (meshed vertical and horizontal bars) for line switching. A modern crossbar switch inherits this concept to implement it electronically using multiply layered electronic selectors. Both parallel and serial crossbar switches are available. However, recent high speed demands generally require use of serial crossbar switches.
Cut Through Method This is in relation to the use of crossbar type switches. In the "store and forward" method, input data is received for a complete frame, its destination address is detected, and then it is transmitted to the destination. On the other hand, the "cut through" method detects the destination to start transmission while input data reception is still ongoing. While this method can reduce switching speeds, use of high speed components is required possibly resulting in higher cost of the equipment.
Data Compression A data compression algorithm is used for reducing total amount of data, e.g. total file volume.
Data Transfer Rate Data transfer rate is the rate of data exchange between a (disk or tape) drive device and a controller. This rate is normally expressed in megabytes per second.
Disk enclosure A cabinet to install disk units.
fabric The term "fabric" is indicative of the network configuration of a SAN (Storage Area Network) in that the network is structured in a meshed form. Thus it implies the network itself or aggregative switches. Also, the term "Fabric Topology" is used as opposed to "Loop Topology."
FC-AL FC-AL stands for "Fibre Channel - Arbitrated Loop." This loop type fibre channel network can connect up to 126 nodes in a topology.
FC Switch This term is for "Fibre Channel Switch." Fibre channel switches are sometimes called "Fabric." However, the term "Fabric" typically refers to a state where a number of switches are connected together or to the fabric topology as opposed to the loop topology.
Fibre channel A so called "fibre channel" is a serial I/O bus capable of transferring data at a maximum rate of 200 MB/s.
Formatting Formatting is to make disk drives available for use by writing necessary data onto the media.
Gbps This represents data transmission capability for a serial interface channel meaning the number of bits that can be transferred in a second. It stands for "Gigabits per second."
HBA HBA stands for "Host Bus Adapter." This term implies of an interface controller used for connection between a server and peripheral equipment. An HBA for a SCSI interface is called a "SCSI Adapter" while an HBA for fibre channels is called a "Fibre channel adapter."
IOPS IOPS indicates number of instructions that can be executed in a second. It stands for "I/O per Second."
IO Density Amount of I/O instructions given from a host within a specified time period.
Logical Disk Recording areas are virtualized by RAID or other schemes so a host can view the disk image as a logical disk. On the other hand, each of the disk drives comprising a RAID system is called a "physical disk."
LTO LTO (Linear Tape-Open) represents specification for a specific tape format to carry high speed and large volume of data. This specification was developed based on LTO technology, which was co-developed by IBM, Hewlett-Packard, and Seagate Technology.
LVD This stands for "Low Voltage Differential." This is indicative of the SCSI interface, in which parallel data are converted to differential signals to use low voltage drive outputs.
Microprocessor A microprocessor is a central processing unit (CPU) to control equipment or a system. In disk array systems, microprocessors are implemented in array controllers or directors to perform appropriate control functions.
MTBF MTBF stands for "Mean Time Between Failure." MTBF is used as a measure of system (equipment) stability. It indicates the average time period from a failure until the next failure occurs. This measure corresponds to the average time period of continuous system operation. The concept is that the bigger this value, the higher reliability.
MTTR MTTR stands for "Mean Time To Repair." MTTR is used as a measure of system (equipment) maintainability. It indicates the average time period required for restoring a failed system.
NAS NAS stands for "Network Attached Storage." NAS is a storage device connected to a network for providing file services to hosts. A NAS system is comprised of an engine to provide file services and data storage devices.
NFS NFS stands for "Network File System." NFS is used mainly for UNIX-based distributed file systems and as a UNIX-based network protocol.
NIS NIS stands for "Network Information Service," which is a distributed data management system first implemented for Sun Microsystems' UNIX machines. NIS uses a hierarchical domain structure for controlling host names of computers within a network.
Operation Management Cost "Operation Management Cost" implies any efforts and costs necessary for managing and maintaining a system. Operation management costs.for a storage system include the manpower of system administration staff personnel and other expenses required for adding storage capacity to cope with storage capacity demands, monitoring for adequate capacity usage, monitoring operational status of the storage system against any occurrence of failures, and data backup operation to appropriately protect data.
RAID RAID is a technique to improve the availability of disk systems by implementing redundant disk arrangements. RAID stands for "Redundant Array of Independent Disks." When one of the member disks constituting a RAID system fails, the data of the failed disk is restored using redundant data for continuing disk access services. Depending on how redundant data are retained, several RAID methods are available for use. RAID-1, RAID-3, RAID-5, and RAID-10 are the typical RAID applications.
Redundancy For example, a redundant system is equipped with spare components that can be used in place of a failed part in order to continue its service functions.
Replicated Volumes Replicated volumes are copied volumes on a logical disk. While RAID-1 is used for mirroring physical disks to provide fixed copies, replicated volume in logical disks can be freely controlled for initiation or separation of replication in accordance with operational requirements. These functions are supported in NEC Storage products to generate replication volumes by "DynamicDataReplication" or "RemoteDataReplication" functions.
Restoration When original data are crashed, damaged, or erased by accidents, restoration is made to recover the original data from backup data.
rpm rpm (revolutions per minute), is a measure to indicate how many times the media is rotated in a minute. A higher rotation rate can result in higher disk performance. This is because a higher rotation rate reduces the waiting time due to disk rotation before a head can access the target sector after its position has been determined.
SAN SAN stands for "Storage Area Network." In this storage network system, storage devices are separated from the storage server and a dedicated network is used to connect between them.In a narrow sense, SAN is sometimes regarded as the network itself. Storage devices are connected to storage servers via fibre channels or Ethernet channels.
Scalability Scalability implies availability of system enhancements. It indicates the extent to which the system can grow from a minimum composition. For example, scalability is indicated by the maximum available number of disk drive units or host ports that can be included in a system.
Server A server system is the main processing equipment that provides services within a computer network. Servers for specific services are designated by attaching their purpose of use in front - such as to call it a "backup server" or "file server."
SPOF SPOF means "Single Point Of Failure." SPOF is indicative of a system component or a path for which failure results in a total system failure.
Storage Capacity Storage capacity means the maximum amount of data that can be recorded in a recording system that use disks or tapes. Data are stored in bytes. Megabytes (MB), gigabytes (GB), and terabytes (TB) are typically used to express storage capacity.
Tape Recording System A tape recording system uses magnetic tapes for recording data. Different types of magnetic tapes are available such as audio cassette tapes, video cassette tapes, or tapes for computer use. Unlike magnetic disk systems, tape recording systems are best suited for sequential access operations. However, they are unsuitable for random access applications.
TCP/IP TCP/IP is a generic term for a group of communications protocols used for the Internet, in which TCP, IP, UDP, and ICMP protocols are included.
Throughput Throughput is a figure to indicate data processing capacity. In general, it is given by the amount of data that can be transferred within a unit time such as MB/s or GB/s.
TOE TOE stands for "TCP/IP Offload Engine." Control software stacks for Internet-based communications protocols are implemented in a hardware device to work as a TOE. Higher processing speed can be achieved
Ultra2 SCSI This is one of the data transfer modes with the SCSI interface having a data transfer rate of 80 MB/s with 16-bit data width. Sometimes it is called "Ultra2 Wide SCSI."
VHDCI VHDCI stands for "68 pin Very High Density Cable Interconnect," representing one of the SCSI interface connector types.
WWN WWN stands for "World Wide Name."
Zoning Zoning is one of the access control functions to enclose irrelevant data flow within FC switches or to make zones for port connections that are intended not to show to others. WWN (soft) zoning and HW (or port) zoning can be applied. The former is software based; it is convenient for use but not robust in security. The latter is understood vice versa.