Overview of Storage Arrays, DAS, NAS, SAN, RAID, Snapshots, and Replication

Overview: A storage array combines multiple hard disks into a single logical unit, storing data in segments across different disks, which greatly reduces access time and improves space utilization. Common array architectures (e.g., Huawei) are divided into integrated controller and separated controller designs.

Storage systems are the foundational platform for IT technology. Initially, storage devices were installed in the same chassis as application servers and were dedicated to a single server.

As the number of servers grew, disks proliferated across many machines, making it difficult to monitor and replace them. This led to the concept of detaching disks from servers and managing them centrally, which is the focus of this article.

DAS (Direct‑Attached Storage) connects storage devices directly to servers via SCSI cables or Fibre Channel. A SCSI loop can attach up to 16 devices, while Fibre Channel can support up to 126 devices in a loop.

NAS (Network‑Attached Storage) provides file‑level sharing services using its own file system and protocols such as NFS or CIFS. It typically runs a lightweight Unix/Linux OS or a specialized Windows kernel to optimize file management and access.

SAN (Storage Area Network) connects storage devices and application servers via a dedicated high‑speed network, enabling fast block‑level data transfer between servers and back‑end storage.

SAN and NAS are often seen as competing technologies, but they complement each other: SAN excels at high‑performance block‑level transfers, while NAS provides convenient file‑level access and sharing.

SCSI (Small Computer System Interface) is a long‑standing interface standard for connecting disks to computers; versions SCSI‑1, SCSI‑2, and SCSI‑3 are widely used.

iSCSI extends SCSI over TCP/IP, allowing block‑level storage traffic to run over standard Ethernet networks, which simplifies deployment and reduces cost.

Fibre Channel (FC) provides high‑speed (2‑32 Gb) connections between servers and storage devices, commonly used for SAN deployments.

RAID (Redundant Array of Inexpensive Disks) combines multiple disks into logical units to improve performance and fault tolerance; various RAID levels (0, 1, 2, 3, 4, 5, 6, 01, 10, etc.) offer different trade‑offs.

Snapshots create point‑in‑time, read‑only copies of data, enabling online backup, rapid recovery, and testing without affecting the original dataset.

Data replication copies data from one disk to another using software (e.g., EMC SRDF), providing an immediately usable duplicate that does not require a separate recovery process.

Tiered storage moves infrequently accessed data to lower‑cost storage tiers, freeing high‑performance resources for hot data and improving overall cost efficiency.

Disclaimer: The original author is credited; any copyright issues should be reported to the publisher.