Comparison of storage technologies (FC SAN, IP SAN, DAS, NAS)

Concept of SAN: SAN (Storage Area Network) is a high-speed, dedicated network for storage operations, usually independent of a computer local area network (LAN). SAN connects hosts and storage devices together, and can provide a dedicated communication channel for any host and any storage device on it. SAN introduces channel technology and network technology into the storage environment, providing a new type of network storage solution that can simultaneously meet the requirements of throughput rate, availability, reliability, scalability and manageability

FC-SAN

Usually, a SAN consists of a disk array (RAID) connected to a Fibre Channel (often called FC-SAN to distinguish it from an IP SAN). “SAN can also be defined as data storage-centric, it uses a scalable network topology, through the direct connection of optical channels with high transmission rates, to provide multiple selectable data exchange between any node within the SAN, and centralize data storage management in a relatively independent storage area network. SAN will eventually achieve in a variety of operating systems, maximum data sharing and data optimization management, and seamless expansion of the system.

1.1 Components of FC-SAN

In FC-SAN, there are some special hardware and software. The hardware includes an FC card, FC HUB, FC switch, storage system, etc. The software is mainly the driver of FC control cards for various operating systems and storage management software.

• FC card: mainly used for the connection between host and FC devices.
• FC HUB: Internal running arbitration ring topology with nodes connected to the HUB sharing 100MB/S bandwidth (or higher).
• FC switch: internally running Fabric topology with 100MB/S bandwidth (or higher) per port exclusively.
• FC storage devices: Using FC connections, there can be one or more fibre interfaces. FC storage devices usually use hard disks with fibre, but there are also Fibre to SCSI (Fibre to ATA) solutions that use SCSI (or ATA) hard disks, which are cheaper in the overall configuration.
• Storage network management software: The main function of storage management software is to automatically discover the network topology and mapping, and to automatically discover and configure when adding or subtracting in the storage network.
• High-performance Fibre Channel switches and Fibre Channel network protocols are the keys to FC-SAN. The network topology with Fibre Channel switches as the backbone is called “SAN Fabric”. The Fibre Channel protocol is another essential feature of FC-SAN, which uses the Fibre Channel protocol loaded with SCSI protocol to achieve reliable block-level data transfer.

1.2 Applications of FC-SAN

Because FC-SAN is optimized for transferring large blocks of data between servers and storage devices, it is ideal for applications such as the following:

• Mission-critical database applications where predictable response time, availability and scalability are essential elements.
• Centralized storage backup, where performance, data consistency and reliability can ensure the security of critical enterprise data.
• High availability and failover environments ensure lower costs and higher application levels.
• Scalable storage virtualization that separates storage from direct host connections and ensures dynamic storage partitioning.
• Improved disaster tolerance features provide Fibre Channel high performance and extended distances between host servers and their attached devices.

1.3 Key benefits of FC-SAN

Facing the rapidly growing demand for data storage, enterprises and service providers are gradually choosing FC-SAN as their network infrastructure because of the excellent scalability of SAN. In fact, SANs offer more significant advantages than traditional storage architectures. For example, traditional server-attached storage is often difficult to update or manage centrally. Each server must be shut down to add and configure new storage. In contrast, FC-SAN can add storage without having to go down and disconnect from the server, and FC-SAN can also manage data centrally, reducing the total cost of ownership.

Using Fibre Channel technology, FC-SAN can efficiently transfer blocks of data. By supporting the transfer of massive blocks of data between storage and servers, SANs provide an efficient way to back up data. As a result, the network bandwidth traditionally used for data backup can be saved for other applications.

Open, industry-standard Fibre Channel technology also makes FC-SAN very flexible. FC-SAN overcomes the traditional cable limitations associated with SCSI connectivity and greatly expands the distance between servers and storage, thereby increasing the possibilities for more connectivity. The improved scalability also simplifies server deployment and upgrades, protecting investments in legacy hardware equipment.

In addition, FC-SAN allows for better control of the storage network environment and suits the needs of those transaction-based systems in terms of performance and availability. SAN leverages highly reliable and high-performance Fibre Channel protocols to meet this need.

Another strength of FC-SAN is the ability to transport blocks of data to enterprise-class data-intensive applications. During data transfer, FC-SAN has less processing overhead at the communication nodes (especially servers) because the data is divided into smaller chunks as it is transferred. As a result, Fibre Channel FC-SAN is very efficient at transporting large blocks of data, making the Fibre Channel protocol well-suited for storage-intensive environments.

IP-SAN

In simple terms, IP-SAN (IP storage) communication channel is the use of IP channels, rather than Fibre Channel, the technology to connect servers and storage devices, in addition to the standard that has been adopted iSCSI, there are FCIP, iFCP and other standards being developed. And iSCSI is the fastest growing and has become a strong representative of IP storage.

Like Fibre Channel, IP storage is switchable, but unlike Fibre Channel, IP networks are mature and do not have interoperability issues, which is the biggest headache with Fibre Channel SANs. IP has been widely recognized by the IT industry and there are a very large number of network management software and service offerings available.

IP storage standards: IP storage in addition to the standard that has been adopted by iSCSI, there are iFCP, FCIP and other standards being developed.

Advantages of IP storage:

• Leveraging the ubiquitous IP network protects existing investments to a certain extent.
• IP storage transcends the limits of geographic distance. as far as IP can extend, storage can extend, making it an almost epochal revolution, well suited for remote backup of existing critical data.
• IP network technology is mature. IP storage reduces the complexity of configuration, maintenance, and management.

2.1 iSCSI Standards

2.1.1 The concept of iSCSI

iSCSI (Internet Small Computer System Interface) is a standard for data block transfers over internet protocol networks, particularly Ethernet. Simply put, iSCSI enables the SCSI protocol to run on IP networks, enabling routing on, for example, high-speed Gigabit Ethernet, enabling the connection of SCSI and TCP/IP protocols.

iSCSI is a technology standard based on the IP protocol that allows users to build storage area networks (SANs) over TCP/IP networks. Before the emergence of iSCSI technology, the only technology to build a storage area network was the use of Fibre Channel, but its architecture required high construction costs that were far beyond the reach of the average enterprise. iSCSI technology emerged as a network environment for users of LANs, which requires little investment and allows for easy and fast interactive transmission and management of information and data. Compared to the previous network access storage, iSCSI emerged to solve the problems of openness, capacity, transmission speed, compatibility, security, etc., and its superior performance has been released since the day the market’s attention and favour.

2.1.2 Technical advantages of iSCSI:

• iSCSI is based on traditional Ethernet and internet, while significantly reducing the total cost of ownership.
• The bandwidth of IP networks is developing quite rapidly, 1Gbps Ethernet has already occupied the market in large quantities, and 10Gbps Ethernet is ready to be launched.
• The bandwidth of IP networks is developing quite rapidly, 1Gbps Ethernet has already occupied the market in large quantities, and 10Gbps Ethernet is ready to be launched.
• Simple to manage and deploy, no need to invest in training, you can easily have professional iSCSI talent.
• iSCSI is a technical standard based on the IP protocol, which implements the connection of SCSI and TCP/IP protocols, and requires little investment for easy and fast interactive transmission and management of information and data.
• Completely solve the problem of remote data replication and disaster recovery. In terms of security, iSCSI has built-in support for IPSEC mechanisms, and the implementation of the relevant instructions at the chip level to ensure security.

2.2 FCIP standards

FCIP is the standard protocol for Fiber Channels over IP. Within the same SAN, TCP/IP packets are then encapsulated with FC commands and data, thus transmitting FC commands and data over the IP network.

FCIP is an Internet Protocol (IP)-based storage networking technology that uses IP networks to implement Fibre Channel Protocol data transfer between SAN devices via data channels, combining true global data mirroring with the flexibility of Fibre Channel SANs and the low cost of IP networks to reduce the cost of remote operations, thereby taking both cost savings and data protection to a new level.

FCIP is proposed as a standard method of connecting Fibre Channel SAN “islands” over existing IP networks and can be used to overcome Fibre Channel’s current distance limitations by connecting SAN islands across distances greater than those supported by Fibre Channel. It retries if the IP network has a high error rate.

FCIP solutions provide various more flexible ways for users to effectively manage business continuity systems, enabling real-time remote replication of data, and can provide users with disaster-tolerant, single-point-of-failure-free SAN solutions based on Fibre Channel controllers, enabling users to use IP networking technology on existing IT infrastructures and use regional SANs as a broader national or even global infrastructure as a data recovery site.

Because FCIP data recovery applications can run on top of existing network infrastructure, users do not need to allocate dedicated fibre optic cables for data volumes in the Fibre Channel when planning for business continuity. By leveraging FCIP solutions, enterprise users can now extend the reach of their SAN beyond the data centre and optimize their infrastructure investments with a variety of low-cost, high-performing remote storage applications.

Because FCIP data recovery applications can run on top of existing network infrastructure, users do not need to allocate dedicated fibre optic cables for data volumes in the Fibre Channel when planning for business continuity. By leveraging FCIP solutions, enterprise users can now extend the reach of their SAN beyond the data centre and optimize their infrastructure investments with a variety of low-cost, high-performing remote storage applications.

2.3 iFCP standards

iFCP (Internet Fibre Channel Protocol) is a standard for running Fibre Channel communications over TCP/IP networks. iFCP has a gateway function that connects Fibre Channel RAID arrays, switches, and servers to IP storage networks without requiring additional infrastructure investment.

iFCP works by encapsulating FC data in IP packets and mapping IP addresses to separate FC devices. Since each type of FC device has its own unique identification in the IP network, it is able to send and receive storage data separately from devices located in other nodes of the IP network. FC signals are terminated at the iFCP gateway and storage communication is carried out in the IP network after signal conversion, thus iFCP breaks the distance (about 10 km) limitation of traditional FC networks.

2.4 Comparison of iSCSI, FCIP and iFCP standards

• iSCSI： Used to establish connections and manage connections between IP-based storage devices, encapsulating SCSI data for transmission over existing IP networks.
• FCIP： Used to connect geographically dispersed FC SANs. only for applications that require the interconnection of two or more FC switches using IP.
• iFCP： Use IP infrastructure to interconnect FC devices or FC SANs. The protocol is dedicated to all FC switching architectures and is not limited to solving distance problems.

NAS

NAS (Network Attached Storage) is network-attached storage. In a NAS storage architecture, the storage system is no longer attached to a server or client through an I/O bus, but is directly connected to the network through a network interface and accessed by users over the network.

A NAS is actually a storage device with a thin server that acts as a dedicated file server. This dedicated storage server removes most of the computational functions that were not applicable to a general-purpose server and provides only file system functions. Compared to traditional server-centric storage systems, data is no longer forwarded through the server memory, but directly between the client and the storage device, with the server playing only a control and management role.

3.1 Key features of NAS

• NAS uses traditional Ethernet protocols, and when performing file sharing, NFS and CIFS are utilized to communicate with NT and Unix systems. Since both NFS and CIFS are OS-based file-sharing protocols, NAS performance is characterized by performing small file-level shared access.
• NAS devices are storage devices that connect directly to the Ethernet and provide file services to clients with standard network file systems such as NFS, SMB/CIFS over TCP/IP interfaces. NAS devices provide file-level services to clients. But internally it still communicates with its storage device at the data block level. The file system is in this NAS storage.

3.2 Key strengths of NAS

• NAS is for users who need to transfer file data to multiple clients over a network. NAS devices work well in environments where data must be transferred over long distances.
• NAS appliances are very easy to deploy. It allows NAS hosts, clients and other devices to be widely distributed throughout the enterprise network environment. NAS can provide reliable file-level data consolidation because file locking is handled by the device itself.
• NAS is used in efficient file-sharing tasks, such as NFS in UNIX and CIFS in Windows NT, where network-based file-level locking provides advanced concurrent access protection.

DAS

DAS (Direct Attached Storage), Direct Attached Storage refers to the use of storage devices connected directly to a server through a bus (SCSI, PCI, IDE, etc.) interface. DAS is attractive to small businesses because of its low acquisition cost and simple configuration.

4.1 Problems with DAS

• Servers themselves tend to be system bottlenecks.
• Server failure and data inaccessibility.
• For systems where multiple servers exist, the devices are scattered and inconvenient to manage. Also when multiple servers use DAS, storage space cannot be dynamically allocated among servers, which may result in considerable waste of resources.
• Complex data backup operation.

Differences among FC-SAN, IP-SAN, and NAS DAS:

If you have some more ideas or queries on Storage products, Please feel free to let us know, Thanks