Storage Area Network SAN
( kw : Storage Area Network )
A Storage Area Network is a high-speed, specialized network that allows storage devices to access the Internet. SANs typically consist of hosts, switches and storage elements. They can also be interconnected with a variety topologies and protocols. Storage Area Networks may span multiple sites.
A Storage Area Network presents storage devices directly to a host so that they appear to be local attached. Different types of virtualization make this possible.
Storage Area Networks are frequently used to:
- Increase application availability (e.g. multiple data paths).
- Enhance the performance of your application (e.g. off-load storage functions or segregate networks, etc. ),
- Storage efficiency and consolidation can be improved (e.g. consolidate storage resources, tiered storage), to increase storage utilization. ), and increase security and data protection.
Storage Area Networks play an important role in an organisation’s Business Continuous Management activities (e.g. by spanning multiple site). Switched fabric technology is the basis of SANs. Fibre Channel (FC), Ethernet and InfiniBand are some examples. Data can be moved between different SAN technologies using gateways.
Fibre Channel is often used in enterprise environments. Fibre Channel can be used to transport SCSI and NVMe as well as FICON and other protocols.
Ethernet is often used by small and medium-sized businesses. Ethernet infrastructure can be used by SANs to connect storage and IP protocols into the same network. Ethernet can be used to transport SCSI and FCoE as well as NVMe and other protocols.
InfiniBand can be used in high-performance computing environments. InfiniBand can be used to transport SRP and NVMe as well as other protocols.
A SAN, whether virtual or traditional, offers many compelling benefits that are essential for enterprise-class workloads.
- High performance. A typical SAN uses a separate fabric for storage tasks. For best performance, the fabric is typically FC. However, iSCSI or converged networks can also be used.
- Scalability. The SAN is capable of supporting large scale deployments that include thousands of SAN hosts and storage devices, or even entire storage systems. As needed, additional storage and hosts can be added to the SAN in order to meet specific organizational requirements.
- High availability. Traditional SANs are based on the concept of a network fabric that interconnects everything. A full-featured SAN deployment does not have a single point of failure between a host device and a storage device. Communication across the fabric can always find another path to storage availability.
- Advanced management capabilities. A storage area network (SAN) will provide a variety of enterprise-class storage features such as data encryption, data replication, and self-healing technologies that maximize storage capacity, security, and data resilience. Almost all features are centrally managed and can be applied to all storage resources on the Storage Area Network .
Storage Area Network SAN Disadvantages
However, Storage Area Networks offer a few drawbacks. IT leaders should consider the potential downsides before installing or upgrading a new SAN.
- Complexity. While there are more convergence options available for SANs such as FCoE or unified options today, traditional Storage Area Network have the additional complexity of a second network. This second network includes expensive, dedicated HBAs (host servers), switches, cabling, and cabling within a redundant fabric. Storage processor ports are also included at the arrays. These networks need to be planned and managed with care. However, the complexity of such networks is becoming more difficult for IT organizations with smaller staff and lower budgets.
- Scale. A Storage Area Network is only effective in large and complex environments that have many servers and substantial storage. Although it is possible to set up a SAN on a smaller scale, the complexity and cost are prohibitive. Smaller deployments may be able to achieve satisfactory results with an iSCSI SAN. This is a converged SAN that runs over a single network, such as the FCoE.
- Management. The complexity concept is primarily focused on hardware. This makes Storage Area Network management a significant challenge. For busy organizations, configurations such as LUN mapping and zoning can be difficult. Although it can take time to set up RAID or other self-healing technology, as well as the corresponding logging/reporting — and security — this is necessary to ensure compliance, DR and BC.
SAN vs. NAS
Network-attached storage is an alternative method of storing and accessing data. It relies on file-based protocols, such SMB and NFS. This is in contrast to block-based protocols like FC and iSCSI that are used in SANs. There are some differences in a SAN and NAS. NAS uses a network to connect storage and servers, while SAN relies on a file server that is located between storage and servers.
There are architectural differences between NAS storage and SAN. Each enterprise storage type has its advantages and disadvantages.
Both systems can store data but the system that is chosen will be determined by the data to be stored. A SAN is preferred for block-based storage. This works well for structured data, such as data for enterprise-class relational databases applications. A NAS, with its file-based approach, is more suited for unstructured data, such as email attachments, documents, images and other common files.
A NAS, like a SAN consolidates storage and can support data protection tasks such as backup and data archiving. A NAS is a network that connects all devices and has a lower cost and complexity than SANs. However, SANs excel in performance and scalability. They can deliver the highest performance for even the most complex enterprise applications.
SAN and NAS do not have to be mutually exclusive. A SAN and NAS can co-exist within the same data center if both file-based and block-based storage is needed. To improve performance, streamline management, combat shadow IT, and limit storage capacity, both NAS and SAN deployments can be upgraded. Sometimes, different storage systems can be replaced by a unified system or the SAN network may be simplified with an iSCSI Storage Area Network .
Major suppliers and products
There are many vendors that offer products and services to help support enterprise SAN deployments. SAN architects should consider hosts (servers), the network (fabric), components, and storage subsystems when planning a deployment.
Hosts. All hosts can access the Storage Area Network . However, every host server must have a suitable network interface in order to access the fabric. Multi-port FC HBAs can be installed on enterprise-class servers. This is a common technique for technology refresh projects. An HBA can be installed on servers that don’t have one. An HBA can be added to a server’s motherboard as an aftermarket upgrade, but it will need a PCIe slot. Before purchasing or proceeding with upgrades, IT staff must inspect each target server to ensure that the appropriate upgrade slot is available. These upgrades will also require that the server be turned off. IT staff must plan for server downtime and prepare for disruptive upgrades.
HBA Cards are often manufactured using core communications chips from tech leaders such as ATTO, Broadcom and Brocade. Many technology vendors and channels sell the actual HBAs.
Network. The SAN fabric is made of both optical and copper cabling, as well as other networking components such as switches. Similar to HBAs and other technology vendors, the appropriate cabling can be purchased through common procurement channels. Based on the technologies used by major chip- and technology manufacturer, both director and edge switches are possible.