What Is a Storage Area Network (SAN)? | Pure Storage

A storage area network (SAN) is a dedicated network of storage devices used to provide a pool of shared storage that multiple computers and servers can access. Storing data in centralized shared storage architecture like SANs allows organizations to manage storage from a collective place and apply consistent policies for security, data protection, and disaster recovery.

What is SAN used for?

SANs are used to provide access to a pool of shared storage to multiple computers and servers. Storage appears to computers on the network as direct-attached storage.

A SAN removes the storage responsibility from individual servers and collects it in a central place where it can be accessed, managed, and protected. Connecting storage to servers through a network separate from the traditional LAN optimizes storage traffic performance, as traffic doesn’t have to compete for LAN bandwidth with servers and workloads.

Why is it called SAN?

SAN is an acronym for storage area network. SANs consolidate storage in a single storage area separated from the local area network (LAN). Computers and devices connected to the SAN have access to storage devices like tape libraries and disk arrays available on the SAN servers as though it is local storage.

How does SAN storage work?

The components of SAN include cabling, host bus adapters, and SAN switches attached to storage arrays and servers. SANs use block-based storage and high-speed architecture to connect servers to logical disk units (LUNs), a range of block storage from a pool of shared storage, and appear to the server as a logical disk. 

A SAN comprises three distinct layers: host, fabric, and storage.

Host Layer

The host layer is made up of the servers attached to the SAN, which run enterprise workloads that require access to storage, e.g. databases. 

SAN hosts use host bus adapters (HBAs), separate network adapters dedicated to SAN access, to interface with a server’s operating system. This allows a workload to communicate storage commands and data to the SAN and its storage resources using the operating system.

Fabric Layer

The fabric layer comprises the cabling and network devices that make up the network fabric that interconnects SAN hosts and storage. SAN networking devices can include SAN switches, gateways, routers, and protocol bridges.

The fabric layer offers increased redundancy over a regular network by providing multiple alternate pathways from hosts to storage across the fabric. This means that if there is a disruption on one path, the SAN can use an alternative path for communication.

Storage Layer

The storage layer comprises several storage devices, which are typically hard disk drives (HDDs), but can include SSDs, CDs, DVDs, and tape drives. Storage devices within a SAN can be organized into RAID groups to increase storage capacity and improve reliability.

SAN technologies support multiple protocols that allow the layers, applications, and operating systems to communicate. The most common protocol used is the Fibre Channel Protocol (FCP) which is based on Fibre Channel (FC) technology. Internet Small Computing System Interface (iSCSI), a less expensive alternative to FC, is commonly used by small and medium-sized organizations. 

Is SAN storage a good backup option?

Storage area networks make good backup options for several reasons:

• SANs allow you to merge backups from multiple servers to a single location and give you the flexibility to scale up storage capacity as needed. 

• A SAN solution also supports faster backups. Since servers can back up directly to the SAN, backups being sent to the SAN do not have to travel over the LAN and are not affected by the traffic on the local network. 

• If one server in a SAN cluster goes down, the workload from that server will automatically failover to one of the other servers in the SAN. You can also replicate data to an offsite SAN to support disaster recovery. 

• Because a SAN is built on a network fabric of interconnected storage devices and computers, if a disruption occurs on one network path, an alternate path can be enabled. This reduces the likelihood that a single device failure will make storage inaccessible.

What are the benefits of SAN storage?

SAN solutions offer several advantages for enterprise workloads, including:

Increased Performance

A SAN moves storage processing from servers to a separate network dedicated to storage tasks. Since the SAN is dedicated only to storage, traffic performance is not affected by the traffic on the LAN. A SAN solution also increases performance on the LAN by removing storage traffic and freeing up bandwidth. 

Greater Scalability

SANs can be made up of thousands of SAN host servers and storage devices that can be scaled out by adding new hosts and storage capacity to meet changing business needs. 

Improved Availability

In a SAN, storage remains independent of applications and is accessible through multiple paths. If a communication failure occurs, the network fabric can use an alternative path to maintain storage availability.

Centralized Management

SANs make storage management simpler with centralized features that can be applied to all the storage resources on the SAN. Features can include data encryption, data deduplication, and storage replication.

Improved Data Security

With a SAN, data protection algorithms are implemented in one place. This ensures that security and compliance configurations are consistently applied to all the servers in the SAN.

How to Access a SAN: 3 Typical Ways

There are several protocols you can use to access a SAN. FCP, iSCSI, and FCoE are the three most common. 

Fibre Channel Protocol (FCP) is the most widely used SAN protocol. FCP uses SCSI commands with an underlying Fiber Channel (FC) connection to transmit data between storage, servers, switches, and data centers.

FC is a high-speed networking technology that offers performance up to 16 gigabits per second and supports both optical and copper media at distances up to 10km. The FC protocol is considered a more secure protocol than TCP/IP and less susceptible to malware and human error. 

Internet Small Computer Systems Interface (iSCSI) is a block protocol that allows SCSI packets to be transported over a TCP/IP network. It offers the same block-level storage access as FCP, but uses conventional Ethernet networks. iSCSI works on top of the Transport Control Protocol (TCP) to allow SCSI commands to be sent over the internet, LANs, and WANs (wide-area networks). 

Fibre Channel over Ethernet (FCoE). Instead of TCP/IP and high-speed Ethernet links, FCoE uses Fibre channel links between SAN devices. FCoE works by sending Fibre Channel packets within Ethernet packets across an Ethernet network. This requires special FCoE-enabled switches and adapters which combine the functionality of an FC host bus adapter with an Ethernet network adapter. FCoE is an excellent alternative to the standard Fibre channel when using higher-performance networks.

When should you use SAN storage?

There are two specific use cases where SAN storage is especially indicated:

Database Management  

SANs provide the high I/O processing speeds and low latency necessary for large enterprise databases in high-speed transactional environments. These databases often contain mission-critical data that require high performance and availability. 

With Fibre Channel technology that can support speeds as high as 128 Gbps, SANs can meet the throughput and latency requirements of enterprise databases. SANs also offer parallel processing and support for RAID and hot-swappable drives for enhanced data management.

Virtualization

SANs are suitable for large-scale high-performance virtualization deployments that run a wide range of operating systems and applications. A SAN can rapidly transfer multiple I/O streams between virtual machines and virtualization hosts and provide the scalability and flexibility to support the dynamic nature of a virtual environment. 

A SAN can support a wide range of operating systems, applications, and virtualization technologies and provide the infrastructure reliability necessary in virtualized environments to prevent failures that cause multiple application outages.

How much does SAN storage cost?

The exact cost of SAN will depend on the workload and storage needs of your organization. 

Typically, a SAN requires FC host bus adapters (HBS), fiber cables, FC switches and disks, disk arrays, and FC-connected tape. Smaller SANs with up to 10 servers and half a terabyte of storage can cost around $100,000. More advanced SAN solutions can cost over a million dollars.

Factors that will affect the cost of a SAN deployment include the type of SAN (FC or iSCSI); manufacturer and model; disk type, speed and capacity; controller capacity; functionality and capacity licensing; and maintenance and support requirements.

Leverage Increased Transfers Speeds with Pure NVMe Solutions

Given the benefits described above, more and more modern organizations are deploying SAN solutions using all-flash storage for its high performance, low latency, and lower total costs.

While FCP, iSCSI, and FCoE protocols work well with HDD and tape-based SANs, they can cause performance bottlenecks with all-flash solutions.

NVMe (non-volatile memory express) is a transfer protocol for quickly accessing flash memory storage using a high-speed PCI Express (PCIe) bus. Benefits of NVMe storage include faster transfer speeds, higher data throughput, and massive parallelism with over 64K queues for I/O operations.

Leverage increased transfer speeds with Pure solutions designed from the bottom up to maximize the potential of NVME storage, including:

• DirectFlash® Fabric: Reap the performance benefits of a SAN designed for SSD.
• FlashArray//X: A performance-optimized all-flash array for Tier 0 and Tier 1 block storage applications.