An ASM cluster is a collection of interconnected nodes, each with an ASM instance, operating as a unified cluster with using Oracle Clusterware. An ASM cluster presents a shared pool of storage to one or more Oracle Databases that are also operating on the nodes. The databases can also be clustered using Oracle Real Application Clusters, but that is not a requirement. The disks that provide the shared storage pool must be accessible from all of the nodes in the ASM cluster.
ASMLIB is an application programming interface (API) developed by Oracle to simplify the operating system–to-database interface and to exploit the capabilities of vendors' storage arrays on Linux-based systems.
A Cluster File System, or CFS, is a file system that is distributed across multiple computer systems. Generally, the computer systems share access to a disk connected through a Storage Area Network (SAN). The CFS component on each individual computer system coordinates access to the disks using a global communication mechanism.
A data extent is the raw storage used to hold the contents of a file. Each data extent consists of one or more allocation units (AU) on a specific disk.
An ASM disk group is a collection of disks that ASM manages as a unit. Within a disk group, ASM exposes a file system interface for Oracle Database files. The content of files that are stored in a disk group are evenly distributed, or striped, to eliminate hot spots and to provide uniform performance across the disks. ASM files may also be optionally mirrored within a disk group. The performance of disks in a disk group is comparable to the performance of raw devices.
Direct Attached Storage, or DAS, consists of storage devices that attach directly to host without an intervening network. DAS generally costs less than SAN attached storage, but also offers less flexibility and functionality. Examples of DAS include SATA disks, common to most desktop systems, and SCSI disks that are found on many servers.
An extent map is a list of extent pointers that point to all the data extents of a file. This is the raw storage for the file. Each extent pointer gives the disk and allocation unit of the data extent. For reliability each extent pointer also includes a check byte to ensure it has not been damaged. This is particularly important when using an in-memory copy of the extent map to direct file I/O to a particular disk location.
Fibre channel is a set of standards that define the interface between computer systems and peripherals. The fibre channel interface is a serial bus interface originally designed to supersede the SCSI standard. However, both the fibre channel and SCSI interfaces have evolved independently of each other and have benefited from the existence of the other. Fibre is spelled with re rather than an er to indicate a distinction from terms such as fiber optics. Both copper and fiber optics are commonly used as a media for fibre channel.
A file system is a software component providing structured access to disks. File systems present objects, such as files, to application programs. Access to files is generally specified with a standard API defining operating system calls such as Open/Close and Read/Write that the application program uses for accessing files. File systems are usually provided as a component of an operating system, but may be provided as an independent software component.
A file is an object presented to an application program by a file system. A file is subdivided into blocks by the file system. A file system typically places what appears to the application program as consecutive blocks, into nonconsecutive locations on disks. The mapping of the file blocks to the locations on disks is kept in what is known as an extent map.
A host bus adapter (HBA), or host adapter, connects a host system to other network and storage devices, such as, eSATA, and SCSI devices.
In SCSI terminology, a Logical Unit Number, or LUN, is a number representing a device address presented to a computer system. In storage technology, the term LUN is often used to denote a disk presented to a computer system by a storage array.
In storage systems, mirroring is a means for protecting the integrity of data by storing copies of data on multiple disks. If a disk fails, then a secondary copy of the data is available on a second or third disk. Generally, mirroring can be deployed in one of two ways. In the most common case, as with a RAID storage array, a fixed amount of physical storage space on a single disk is uniformly copied on multiple disks. Through software on the storage array, the physically mirrored space on multiple devices is presented to the host as a single disk. Another approach to mirroring is to copy fixed segments of a file onto multiple disks. ASM uses the second approach for maximum flexibility.
Network Attached Storage (NAS)
Network Attached Storage, or NAS, comprises a class of systems providing file services to host computers. A device driver in the operating system accesses data using protocols such as NFS or Common Internet File System (CIFS). From the application program's point of view, Network Attached Storage appears as a file system local to the host computer on which the application program resides.
A network file system, or NFS, is an industry standard providing Network Attached Storage on many platforms. NFS extends the local file system framework to provide device sharing to users on an IP network. The user view of an NFS is that a remote file system on a host appears as if it were locally mounted.
Oracle Cluster File System (OCFS)
Oracle offers two cluster file systems, OCFS for Windows and OCFS2 for Linux. While OCFS for Windows is a proprietary file system, the source for OCFS2 for Linux is available to all under GNUs' General Public License (GPL). The two file systems are not compatible.
Operating systems typically provide a means for splitting a disk into sections called partitions. To an application, a partition appears as an independent disk, even though that is not the case. The operating system provides a command for managing the size and locations of partitions on a disk. This command writes a partition map onto the disk in an area that is not visible to the application program.
Redundant Array of Inexpensive Disks (RAID)
Redundant Array of Inexpensive Disks, or RAID, is a means for combining multiple disks through software and presenting them to a host as a collection of apparently distinct disks. Mirroring and striping are two means used for combining the disks as a collection. RAID can be implemented as a software component of a computer system or as software operating inside of a storage array. RAID operates on the physical presentation of storage as opposed to ASM mirroring and striping at the logical file level.
Storage Area Network, or SAN, is a specialized network designed for attaching storage devices, such as disk arrays and tape drives, to computer systems. The most common network infrastructure used in SANs is fibre channel. SANs provide great flexibility by allowing a common interconnect infrastructure in which any storage device can be connected to any computer system. Zoning and LUN masking are commonly used for access control for SANs. Zoning is a mechanism where the SAN infrastructure determines the access rights of storage devices to computer systems. LUN masking is a mechanism where the storage device determines the access rights of what computer systems is permitted to connect to the device.
Small Computer System Interface (SCSI)
Small Computer System Interface, or SCSI, is a standard defining the interface between computer systems and peripherals, most commonly storage devices. The SCSI standard defines both the logical and physical aspects of the interface between computer systems and peripherals.
In storage systems, striping is a means for spreading data across multiple disks as opposed to storing the data on a single disk. Striping is usually done to improve performance. Generally, striping can be deployed in one of two ways. In the most common case, as with a RAID storage array, a fixed amount of physical storage space that could have been stored on a single disk is uniformly striped across multiple disks. Through software on the storage array, the physically striped space on multiple devices is presented to the host as a single disk. Another approach to striping is to stripe fixed segments of a file across multiple disks connected to the host. ASM uses the second approach for maximum performance.
In the storage realm, the meaning of volume has many related definitions. Volumes are often thought of as the virtual entity represented as aLogical Unit Number (LUN). Volumes often are presented as an aggregation of pieces from several disks. The mapping of the collection of the pieces of the disks into a volume is managed by a software component called a Volume Manager.