IT is an industry notorious for jargon, and when it comes to data centers and their related performance, hardware, storage and capacity, there’s certainly plenty of acronyms and new words to be learnt. Here, the Procurri team explains all you need to know on data storage in layman’s terms.
Please note this list is in alphabetical order and not that of priority. Every business functions differently and every data center configuration is unique, so there’s really no one-size-fits-all for what you should know!
Access Methods
The method used to access data on any storage device. If an access method is described as direct, random access is permitted to any data point. If an access method is described as indexed, it uses an index to locate specific information. If an access method is described as sequential, the data is read in order.
Access Time
The amount of time it takes to retrieve data from the device on which it is stored.
Bit Rot
A colloquialism for the slow process of data deterioration over time. When individual bits of data flip from their intended state and fail, they create gradual accumulated errors within storage. This can lead to complete loss of data over time.
Capacity Units
The measurement of the available size of data storage on a device. This begins with bytes, then kilobytes (KB, 1000 bytes), megabytes (MB, 1000000 bytes), gigabytes (GB, 1000000000 bytes) and terabytes (TB, 1000000000000 bytes).
Compression
The compression of files to reduce their size. This is achieved through efficient encoding to save on taking up unnecessary storage space.
Cloud Storage
The storage of data via Cloud computing services to access via the internet.
Cloud Storage Redundancy
The practice of storing data across Cloud data servers that are geographically dispersed, for the process of ‘redundancy’.
Data Lifecycle Management (DLM)
A management process of policies to define how a business manages data through its lifecycle; including guidance on when to migrate it, delete it, or archive it.
Data Migration
The process of migrating or moving data from one data storage system to another one.
Data Sprawl
The uncontrolled spread of data across an entity or organization as it is distributed across different servers, storage devices, cloud platforms and sites – but without the necessary governance or security oversight to keep it secure. This can create an influx of unstructured data that is not secured as it should be.
Data Types
The type of storage being used, dependent on data access requirements. This includes:
- Block Storage, which divides data into blocks for ease of access
- File Storage, which stores data as individual named files with extensions
- Object Storage, which stores data as objects with metadata descriptions.
Deduplication
The practice of eliminating duplicated blocks of data by replacing them with a direction to the first iteration of the same dataset.
Disk Thrashing
A phenomenon of the gradual degradation of computer system performances when the hard drive is overworked. Its continual work to read and write data makes the system’s virtual memory manager overly active, and it switches data between physical RAM and disk storage more often than necessary. This excess activity slows down the performance of the overall storage system.
Fault Tolerance
A system that guarantees no service interruption. Fault tolerant systems are fairly complex in order to be able to provide such a guarantee, as they must run active copies of data at all times to avoid any failure or downtime.
Latency
The time delay between the requesting and the receiving of data from a storage device.
LUN
An acronym for ‘Logical Unit Number’. LUNs manage block storage to present different views of a disk to a server.
Magnetic Digital Data Storage
The oldest and most conventional form of digital data storage. These devices include magnetic tape to record and play back information. This includes removable floppy disks and common hard drive disks.
Mirroring
The process of creating an exact duplicate of data on a different storage device to the original for the practice of ‘redundancy’.
NVMe
An acronym for ‘Non-Volatile Memory Express. NVMe is a protocol designed for high-performance SSDs to streamline a command set and utilize low-latency and high-speed flash memory. The use of NVMe can help maximize SSD potential and deliver enhanced performance.
Optical Digital Data Storage
Digital data storage disks that hold data through the stamped pits onto plastic. A long-term data storage solution, this includes CDs, CD+Rs, DVDs, DVD-Rs, DVD+Rs, and Blu-Ray disks. Data stored on optical digital data storage solutions isn’t required to be frequently edited or changed.
OS
An acronym for ‘Operating System’. OS is system software that manages hardware and software resources.
Overprovisioning
A pre-emptive approach to allocate more resources to a system or network than is actually required. This is intended to prevent any performance issues if there are spikes in demand for data.
Partitions
A container created with a file system to allow the writing to disk. There may be one partition for a whole storage device’s space, or multiple to divide up the capacity. Once a partition is created, it is then formatted with a file system to allow files to be written to it.
RAID
An acronym for ‘Redundant Array of Independent Disks’. This is a system that combines multiple hard drives to provide redundancy services.
RAM
An acronym for ‘Random Access Memory’. A memory type that stores data being actively used by a computer or device on a temporary basis. RAM is known as a ‘volatile’ memory type.
Redundancy
The practice of storing duplicate data in multiple locations so that a back-up is available if the original fails. This is a business continuity method to prevent data loss for anything critical in case of hardware failure, system crashes, or other disruptions.
ROM
An acronym for ‘Read-Only Memory’. A memory type that stores data that cannot be easily modified. ROM is known as a ‘non-volatile’ memory type.
RPO, RTA, RTO
Acronyms for ‘Recovery Point Objective’, ‘Recovery Time Actual’ and ‘Recovery Time Objective’.
RPO is the time period of data loss suffered during an incident; from the time of the incident occurring back to the last known good status point for data recovery. RTA is the actual time period experienced to complete data recovery and make a storage copy available for access. RTO is the time period the organization defines as their SLA to restore all affected service after a disruptive incident.
Secondary Storage
The additional memory devices are deployed to meet the need for extra storage where a device’s primary memory is not sufficient. Secondary Storage devices can be used for long-term storage and backup of data; usually only accessed as and when needed. The use of Secondary Storage devices improves data management by extending capacity.
Software-Defined Storage
Technology used in data storage management to separate the functions responsible for provisioning capacity, protecting data and control data placement from the physical hardware on which it is stores. This allows for the easy replacement, upgrade and expansion of storage.
Solid State Digital Data Storage (SSDs)
The eventual replacement for magnetic digital data storage devices. These include flash drives, memory cards, and the internal storage facilities in smartphones. SSDs use flash memory and semiconductor technology to store data. SSDs are an easy way to transfer and share data with others as they’re easily transportable.
Storage Media
The physical (hardware) device on which data is stored. This may be a hard disk drive (HDD), magnetic tape, optical disc or solid-state drive.
Storage Tiers
An indication of how often data is accessed from its storage. This colloquialism includes ‘hot data’, which is accessed often on high-performance fast storage; ‘warm data’, which is accessed less frequently and is stored on a device that is balanced between performance and cost (as no higher requirement is needed); and ‘cold data’, that is rarely accessed and so held on inexpensive storage.
Thick and Thin Provisioning
Two approaches to storage capacity management.
Thick Provisioning refers to allocating a fixed amount of disk space to a dataset from its initiation, dedicating and reserving it for that volume, independent of the actual amount of data stored. Thin Providing refers to a more dynamic approach that allocates disk space only as it is needed for data storage, with no upfront reservation. This creates a virtual storage pool from which storage is drawn upon as required.
Vendor Lock-In
The deliberate practice of a vendor (or OEM) ensuring that only they can be employed to manage a storage facility. This may be through the use of employ custom-designed controllers, which don’t allow for integration with other vendors’ hardware; through the use of non-standardized data formats to prevent migration; or through the use of unique connection ports and proprietary cabling solutions to avoid storage array connections with other hardware.
Volume
A single accessible storage area within a filesystem; usually a partition of a drive or a hard drive. A volume may not be a physical disk but could be accessed via the logical interface of an OS.
Still unsure of which approaches may work best for you? Get in touch with the Procurri team today and learn more on the configuration complexities we can manage for your business!