Storage area network (SAN)

Storage area network (SAN)

A storage area network (SAN) is a specialized, high-speed network that provides block-level access to consolidated storage devices. To operating systems and applications, SAN storage appears as locally attached disks, even though the physical media reside on remote storage arrays. A SAN separates the concerns of file systems and raw storage, enabling multiple servers to share centralized storage resources while retaining local control of the file system layer.

How a SAN is structured

• Hosts: servers that mount remote storage volumes and run local file systems.
• Storage arrays: disk or flash hardware that expose logical volumes (LUNs) to hosts.
• SAN switching fabric: dedicated switches that route block I/O between hosts and storage.
• Management layer: software for provisioning, zoning, replication, and monitoring.

Common protocols and features

• Fibre Channel (FC): a long-established, low-latency protocol often used in enterprise SANs.
• iSCSI: encapsulates SCSI over IP networks to use existing Ethernet infrastructure.
• FCoE (Fibre Channel over Ethernet): carries FC frames over Ethernet while preserving FC features.
• Multipathing, zoning, and LUN masking: techniques to ensure redundancy, security, and controlled access to volumes.

File system vs. storage: distinction that defines SAN and NAS

The file system provides the namespace and implements file access semantics for applications; storage provides blocks where data actually reside. In direct-attached storage (DAS), file system and storage live in the same server. In network-attached storage (NAS), the file system is served by a remote appliance and clients access files over a network file protocol. In a SAN, the file system stays on the host while only the raw block storage is remote. For a concise comparison see SAN presentation to hosts and for hardware background see disk and array technologies.

Typical use cases and benefits

SANs are common in environments that need high throughput, low latency, predictable performance, and central administration: large databases, virtualization clusters, high-performance computing, and enterprise backup or replication. Centralized storage simplifies provisioning, snapshot-based backups, site replication, and storage tiering. SANs can scale capacity and performance independently of server resources.

Trade-offs and operational considerations

Deploying and running a SAN involves higher capital and operational cost than simple NAS or DAS. Designing the switching fabric, handling zoning, implementing multipath I/O, and ensuring firmware and driver compatibility require expertise. Nevertheless, for workloads that demand block-level semantics, multipath resilience, and consistent predictable I/O, SANs remain a core enterprise building block.

Short history and notable facts

SAN concepts emerged as businesses sought to decouple storage from servers in the late 1980s and 1990s, maturing with Fibre Channel standards and later with iSCSI and converged Ethernet solutions. Modern SANs often integrate flash, software-defined storage features, and cloud replication to bridge on-premises and remote environments.