Understanding DAS, NAS, SAN and RAID: A Complete Storage Fundamentals Guide

1. Storage Overview and Methods

1.1 Introduction

Storage is the practice of saving data on media in a safe, efficient manner, serving both as the physical medium for data residency and the method that guarantees data integrity.

1.2 Common storage methods: DAS, NAS, SAN

DAS

DAS (Direct Access Storage) connects storage devices directly to a computer via SCSI or Fibre Channel. It resembles typical PC storage architecture, with the storage device part of the server’s internal bus.

DAS is suitable for small networks, geographically dispersed sites where SAN/NAS interconnection is difficult, and special application servers that require direct‑attached disks.

Limitations

DAS relies on the host operating system for I/O, consuming CPU and I/O resources for backup and recovery; as storage capacity grows, it can become a performance bottleneck.

NAS

NAS (Network Attached Storage) attaches storage devices to a standard network (e.g., Ethernet) and provides a dedicated storage server that separates storage from application servers, reducing bandwidth usage and total cost of ownership.

Advantages

True plug‑and‑play.

Simple deployment.

Flexible device placement.

Easy management and low cost.

Disadvantages

Lower performance compared with DAS/SAN.

Reliability not as high as direct‑attached solutions.

SAN

SAN (Storage Area Network) originated in 1991 with IBM’s ESCON technology and uses Fibre Channel to provide high‑speed, networked storage. It separates transport protocols from physical media, allowing multiple protocols over a single connection.

3. Comparison of DAS, NAS, SAN

3.1 Connection comparison

DAS uses direct connections, NAS uses TCP/IP networks, and SAN uses Fibre Channel, each offering different performance, scalability, and cost characteristics.

3.2 Application scenarios

DAS suits small‑to‑medium enterprises with modest data volumes and high speed requirements.

NAS is ideal for file servers and unstructured data, offering flexible, low‑cost deployment.

SAN fits large applications or databases where performance and scalability justify higher cost.

2. Disk Arrays and RAID Technology

2.1 Disk Arrays

Disk arrays combine many inexpensive disks into a large logical volume, using techniques such as parity to provide redundancy and improve performance.

2.2 RAID Overview

RAID (Redundant Array of Independent Disks) groups multiple disks to achieve higher speed, data protection, or both. Common levels include 0, 1, 5, 10, etc.

RAID 0

Stripes data across disks for maximum throughput but offers no redundancy.

RAID 1

Mirrors data on pairs of disks, providing full redundancy at the cost of 50% storage efficiency.

RAID 10

Combines RAID 0 striping with RAID 1 mirroring, delivering high performance and fault tolerance.

RAID 5

Distributes parity across all disks, allowing one disk failure while maintaining good read/write performance.

Advantages and disadvantages of each level

RAID 0: fastest, no fault tolerance.

RAID 1: full fault tolerance, high cost.

RAID 5: good balance, requires at least three disks.

RAID 10: high performance and fault tolerance, but 50% capacity loss.

Hardware RAID vs Software RAID

Hardware RAID uses a dedicated controller to present a virtual disk to the OS, hiding individual member disks. Software RAID implements the functionality in the OS layer, making member disks visible but presenting a virtual RAID volume to users.

4. Common Storage Brands

EMC, IBM, NetApp, HP, HDS, Dell, Huawei, etc.

5. Common Storage Media

Hard disks

Tapes

Optical discs

Portable storage devices