What is a Storage Area Network (SAN) and how does it work?

A Storage Area Network (SAN) is a spe­cial­ised, high-per­form­ance network that connects servers with shared storage devices, allowing block-based access. It is primarily used in data centres to ef­fi­ciently and centrally manage large amounts of data and storage capacity.

What is a Storage Area Network (SAN)?

A Storage Area Network (SAN) is a storage network designed to connect disk systems to server systems. In a SAN, the entire storage capacity—provided by storage media like HDDs, SSDs, disk arrays (disk sub­sys­tems), or tape libraries—is con­sol­id­ated into a virtual storage unit and centrally managed. Access to the SAN storage is achieved through ap­pro­pri­ately con­figured servers. The storage network is operated parallel to a Local Area Network (LAN) and provides the entire mass storage to this computer network. SAN can thus be viewed as a secondary network dedicated solely to data transport to mass storage.

SAN servers act as the con­nec­tion point. These don’t need to be located at the same place as the in­di­vidu­al disk drives. Ad­di­tion­ally, the SAN operates in­de­pend­ently of the operating system of the computers accessing the storage. This makes a Storage Area Network ideal for cross-platform data man­age­ment.

SAN systems were developed with the un­der­stand­ing that dedicated disk drives in server systems often lead to man­age­ment issues. With shared virtual storage capacity, Storage Area Networks are sig­ni­fic­antly more effective and flexible in handling large volumes of data. Unlike simple network storage solutions like Network Attached Storage (NAS), SAN offers the advantage that the LAN is not burdened by data access to the mass storage due to the ad­di­tion­al network.

System ar­chi­tec­ture of a Storage Area Network

Setting up a Storage Area Network is con­sidered complex and expensive, as it requires a complete system ar­chi­tec­ture to be purchased with sometimes very costly hardware. Apart from cabling, three basic com­pon­ents are needed: SAN servers, Fibre Channel switches, and storage media.

• Fibre Channel fabric: In IT, a ‘fabric’ refers to a network of in­ter­con­nec­ted cables and switches that ensures high re­dund­ancy through multiple cross-links. This ar­chi­tec­ture delivers not only im­press­ive data through­put but also out­stand­ing bandwidth and re­li­ab­il­ity. At the heart of a Storage Area Network (SAN) lies such a fabric, made up of Fibre Channel switches and fibre-optic cables. All devices within the SAN connect to the Fibre Channel switch, which dy­nam­ic­ally manages the real-time data paths between senders and receivers. Fibre Channel was developed as a standard interface for storage networks and supports con­sist­ent, high-speed data transfers — with rates of up to 16 Gbit/s.
• Storage elements: In Storage Area Networks, disk arrays are generally used for data storage. These are devices con­tain­ing multiple hard disk drives (HDDs) or solid-state drives (SSDs) for mass storage. To ensure high avail­ab­il­ity of stored data, modern disk arrays have a con­trol­ler re­spons­ible for redundant data storage and load balancing during data transfer. This increases data security and enhances transfer rates. The im­ple­ment­a­tion of a redundant storage process usually involves RAID-systems (Redundant Array of In­de­pend­ent Disks), which combine in­di­vidu­al physical disk drives of a disk array into a logical drive. Al­tern­at­ively, tape libraries or in­di­vidu­al disk drives can also be in­teg­rated into a SAN.
• SAN servers: To ensure smooth in­ter­ac­tion with the virtual storage unit, each SAN requires spe­cific­ally con­figured servers that manage data access, acting as a link between the storage network and the devices connected in the LAN. These SAN servers are connected to the Fibre Channel switch via special hardware in­ter­faces known as Host Bus Adapters (HBA).

How exactly does a Storage Area Network work?

A SAN is based on a network structure spe­cific­ally designed for storage access, separate from con­ven­tion­al data networks like a LAN. The found­a­tion of the Storage Area Network includes Fibre Channel or iSCSI con­nec­tions, through which servers or hosts com­mu­nic­ate with storage resources.

The pre­vi­ously mentioned storage elements provide block-based storage to the SAN. Com­mu­nic­a­tion is conducted through SAN switches, which operate similarly to tra­di­tion­al network switches but are designed for high data rates and minimal latency.

Through what are known as LUNs (Logical Unit Numbers), storage is allocated to in­di­vidu­al servers. These servers access the LUNs through special Host Bus Adapters (HBA) or iSCSI ini­ti­at­ors, as if they were locally attached drives. The Storage Area Network ensures that these storage accesses occur in parallel, reliably, and without in­ter­fer­ence. The man­age­ment and al­loc­a­tion of storage resources are handled by cent­ral­ised man­age­ment software, which controls access rights and avail­ab­il­ity. This central or­gan­isa­tion allows storage resources to be flexibly expanded, re­as­signed, or mirrored even while in operation.

How to set up a SAN System step by step

A Storage Area Network is designed for high avail­ab­il­ity. In IT, this refers to the ability of a computer system to ensure the smooth operation of business-critical ap­plic­a­tions despite the failure of in­di­vidu­al hardware com­pon­ents. Setting up a Storage Area Network involves several steps, as shown below.

Step 1: Assessing re­quire­ments and planning

First, it is de­term­ined what the SAN will be used for. There are various use cases, such as virtual machines, databases, or large data archives. This de­term­ines how much storage space is needed, how fast access needs to be, and how many servers should be connected. Re­quire­ments for fault tolerance are also con­sidered.

Step 2: Selecting and preparing hardware

A SAN, as pre­vi­ously described, consists of several central com­pon­ents. To achieve high avail­ab­il­ity, all central com­pon­ents are du­plic­ated. This means there are at least two switches, two data paths, dual power supplies, and often two RAID con­trol­lers. If one part fails, the other auto­mat­ic­ally takes over. This principle is also known as re­dund­ancy.

Step 3: Setting up the network

In the next step, the physical con­nec­tion between the devices is es­tab­lished. The servers are connected to the SAN switches via their host bus adapters, as are the storage systems. This con­nec­tion is made via Fibre Channel or iSCSI, using fibre optic cables or special network cables. The goal is to create a separate, fast, and stable network.

Step 4: Con­fig­ur­ing the storage system

On the central storage system, the desired storage structure is then set up. This includes setting up RAID systems to increase fault tolerance. RAID 1 or RAID 10 are common variants where data is always doubly stored on different hard drives. This mirroring is part of the so-called redundant storage cycles and ensures that no data is lost even if a hard drive fails.

In addition, so-called LUNs (Logical Unit Numbers) are defined. Through mul­tipath­ing tech­no­lo­gies, servers can sim­ul­tan­eously access the same storage resource via multiple paths, further in­creas­ing fault tolerance.

Step 5: In­teg­rat­ing the server

The servers that need to access the SAN storage are also prepared ac­cord­ingly. They receive the ap­pro­pri­ate drivers for the host bus adapters. Ad­di­tion­ally, a mul­tipath­ing software is installed, which auto­mat­ic­ally detects if a path fails and redirects the data traffic to another path. The assigned LUNs are then in­teg­rated and re­cog­nised by the operating system as ad­di­tion­al drives.

Step 6: Security and access control

To ensure that not every server can access arbitrary storage areas, so-called zones are set up. They define which server is allowed to access which storage resources. For iSCSI con­nec­tions, au­then­tic­a­tion via CHAP (Challenge Handshake Au­then­tic­a­tion Protocol) is also used.

Step 7: Testing and mon­it­or­ing

Before the SAN is put into pro­duct­ive use, all con­nec­tions are tested. This includes checking whether the storage is ac­cess­ible from all des­ig­nated servers and whether it auto­mat­ic­ally switches to redundant systems in case of a failure. After suc­cess­ful test operation, mon­it­or­ing is set up to con­tinu­ously supervise the condition of the hard drives, util­isa­tion, and con­nec­tion path activity.

Use cases of Storage Area Networks

Storage area networks are primarily used in busi­nesses that need to process and store large amounts of data, such as in data centres, with cloud providers, or in media and film pro­duc­tion. A SAN is es­pe­cially suitable for en­vir­on­ments where many servers need to access the same data sim­ul­tan­eously, such as in database ap­plic­a­tions, vir­tu­al­isa­tion solutions, or ERP systems. In research and science, SANs also offer crucial ad­vant­ages due to their high speed and re­li­ab­il­ity.

In the area of data backup, the SAN serves as a powerful and scalable solution for backups and recovery. Thanks to high avail­ab­il­ity and re­dund­ancy, SAN is also fre­quently used for business-critical ap­plic­a­tions where a storage system failure would have severe con­sequences. A SAN is also suitable for companies with multiple locations or dis­trib­uted data centres, as the storage can be managed in­de­pend­ently of the server locations.

Overview of the ad­vant­ages of Storage Area Networks

Operators of a SAN benefit from combining the ca­pa­cit­ies of various physical data storages into a virtual storage found­a­tion that is flexibly and platform-in­de­pend­ently available to any number of servers. This de­coup­ling of storage medium and accessing server sim­pli­fies the man­age­ment and scaling of available storage capacity and reduces the load on the LAN. The fibre-optic-based storage area network supports high data transfer rates due to the standard interface, Fibre Channel.

Mul­tipath­ing and the cross-con­nec­tion-focused Fibre Channel fabric prevent overloads and ensure constant avail­ab­il­ity of the data stored in SAN storage. Multiple paths between the data pool and the user can always be used for access. The redundant dis­tri­bu­tion of data across multiple physical systems guar­an­tees high security of the stored content. Delays in sim­ul­tan­eous data access are minimised. Due to effective storage man­age­ment, SAN storage is often used as a basis for virtual servers in the hosting business, which are offered to end customers as ‘In­fra­struc­ture as a Service’ (IaaS) via the cloud.

Ad­vant­ages of SAN at a glance:

✓ Cent­ral­ised, virtual storage base from multiple physical data carriers

✓ Flexible and platform-in­de­pend­ent access by an unlimited number of servers

✓ De­coup­ling of storage medium and server sim­pli­fies man­age­ment and scaling

✓ Relief for the LAN as the SAN operates as a separate network

✓ High data transfer rates through fibre optic tech­no­logy and Fibre Channel

✓ Mul­tipath­ing allows multiple data paths sim­ul­tan­eously and prevents bot­tle­necks

✓ High avail­ab­il­ity through in­ter­con­nec­tions in the Fibre Channel fabric

✓ Redundant data storage on multiple physical systems increases data security

✓ Min­im­isa­tion of delays during sim­ul­tan­eous accesses

✓ Efficient storage man­age­ment for large volumes of data

✓ Ideal for vir­tu­al­isa­tion and cloud services, e.g., for In­fra­struc­ture as a Service (IaaS)