Hardware - Concepts
CPU
A Central Processing Unit (CPU) is the computer’s main processor. The CPU executes instructions comprising a computer program. The choice for a specific CPU depends on your product’s computing requirements.
HDD
A Hard disk drive (HDD) uses rotating magnetic disks to store data. HDDs provide larger storage capacity but slower read/write speed compared to SSDs. They are perfect for storing large volumes of data.
NVMe
NVM Express (NVMe) disks are SSDs that use the PCI Express bus for communication with the host.
SSD
A Solid state disk (SSD) uses flash memory to store your data. Since there are no rotating parts, they provide faster read/write speed than HDDs.
RAID
RAID is a data storage technology used for data redundancy, performance improvement, or both by combining multiple physical hard drives into one or more logical drives.
The different RAID schemes are referenced as levels and are named by the word RAID itself, followed by a number (For example, RAID 0, RAID 1 or RAID 5). Depending on the RAID level, data is distributed and stored across the disks in one or several ways.
RAID 0
RAID 0 uses data striping. This increases the storage capacity of the virtual disk to the sum of all available disks in the RAID. The write and read performances of the machine are increased in RAID 0 to concurrent read and write operations. There is zero fault tolerance in RAID 0, as the contents of each file are distributed among all disks in the set, the failure of any single disk in the RAID array causes the entire RAID 0 volume to break down. The minimum number of disks in a RAID 0 set is two.
RAID 1
RAID 1 uses disk mirroring, meaning data is written identically to any of the two disks in the set. File requests are broadcasted to any drive in the array and can be served by the drive that accesses the data at first, improving read performance. Write performance can be slower than using a single drive, as any write request has to be sent to all drives in the set, limiting write performances to the speed of the slowest disk. A fault tolerance is provided using this RAID level, as data can be restored from the second disk in case one disk fails. The minimum number of disks available in a RAID 1 set is two, and the number can be increased in odd pairs (For example, 2, 4, 6, …).
RAID 5
RAID 5 uses block-level data striping with distributed parity. This means parity information is distributed among all available drives, resulting in a fault tolerance where all drives but one need to be present to operate. If a single drive fails, subsequent reads can be calculated from the distributed parity available on the other drives, so that no data is lost. This RAID level requires a minimum of three disks in the set, and the total available space of the virtual disk is the sum of all drives minus one.