Enterprise Storage Forum content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.
Fibre Channel is a high-speed data transfer technology designed to connect computers to storage devices. It delivers the block data—data divided into equal sized blocks to optimize storage and retrieval speeds—between the servers and shared storage arrays that form the foundation of storage area networks (SANs).
Because many enterprise applications require data to be fed to them rapidly, the speed with which data can be written to and read from storage has become increasingly important. Fibre Channel, or FC, is one of the most important developments among the many protocols and technologies developed to provide high-performance storage for enterprise environments.
Table of Contents
How Does Fibre Channel Work?
FC is based upon the Fibre Channel Protocol (FCP) that ensures a seamless data flow between servers and storage networks. As it is a high-speed form of data transport, it normally runs on optical fiber cables within and between data centers. It can also run on slower copper cabling.
FC evolved as a way to achieve broader, faster access to shared storage within the enterprise and offers higher bandwidth than other data protocols. Over the years, it has expanded from 2 gigabits per sec (Gbps) to as much as 128 Gbps. These steady gains have played a vital role in allowing storage to keep pace with the massive gains made in processing and memory speed required to meet the needs of the modern business.
An FC network is also known as a storage fabric and is composed of several elements:
- Host Bus Adapters (HBAs), or dedicated server storage adapters
- FC switches that can operate at the required speeds
- Storage systems made up of disks and storage controllers that determine what data goes where
The FCP operates as a means of implementing the small computer systems interface (SCSI) protocol across the Fibre Channel fabric, which is composed of five different layers that provide all the functionality built into FC networks and block storage:
- Physical media linking FC ports
- The FCP transmission protocol
- A transport mechanism
- Common services for advanced features
- Application interfaces and mapping rules that execute over FC
Fibre Channel vs. SCSI
SCSI is a standard interface used for computer-to-storage connectivity, but has limits related to local area network (LAN) storage performance and does not excel at supporting multiple host-to-storage device connections. SCSI is best for point-to-point connections or direct attached storage (DAS) interfacing.
Because its throughput rate doesn’t go beyond 640 MBps, SCSI can act as a bottleneck when large amounts of data are involved. There are also limitations related to the number of devices per channel, which make it problematic for enterprise-class storage.
One of the main functions of the Fibre Channel Protocol is to carry SCSI and eliminate many of its server-to-storage limitations. FCP greatly extends connectivity distances and opens the door to such functions as remote backup, archiving, data mirroring, and other advanced storage features.
It becomes possible to make much quicker input/output transfers across longer distances by connecting redundant arrays of independent disks (RAID) to a server using FC. An FC storage area network can also provide storage scalability—instead of a single box, multiple storage arrays can be connected within a single SAN to provide very large amounts of storage that can be managed as one system.
The Importance of Fibre Channel
The FC storage area network ushered in an era of shared storage that facilitates the scaling up of enterprise storage to unprecedented levels. Its importance in the evolution of storage cannot be underestimated—the FC SAN enables the access, management, and sharing of storage among different devices and servers, all across a dedicated network specifically optimized for storage tasks.
As it operates independently of the LAN, the FC SAN eliminates network bottlenecks by keeping storage traffic off the main network, enhancing the performance of both.
Fibre Channel has enabled storage infrastructure at a previously unimagined scale. When first introduced, FC offered a significant bump in bandwidth and geographic range compared to SCSI, bringing capacity- and performance-scaling as well as high availability services, fault isolation, hot-pluggability, and other enterprise storage capabilities.
What is Fibre Channel used for?
Fibre Channel SANs are deployed for many reasons, primarily to provide shared access with high performance to large amounts of storage. Here are some common deployments:
- Low-latency applications. FC is favored for high-speed online transactional processing (OLTP) databases often used in insurance, banking, and online ticketing industries.
- Large virtualized environments. When thousands of virtual machines (VMs) are connected, they need fast-paced shared storage to operate effectively, making FC a good fit.
- Disaster recovery sites. FC provides fast transfer of data from one data center to another with low latency so that it is available in near real time, making it well-suited for mirroring an enterprise data center to a distant disaster recovery data center.
- Fault tolerance. Fault-tolerant access paths with automatic failover are available with FC as well as such features as dynamic reallocation of storage devices, logical unit (LUN) masking, and high-availability clusters.
Fibre Channel vs Ethernet: What are the Differences?
While there are significant differences between Ethernet and FC, in most cases it’s not an either/or choice for enterprises. The two technologies complement one another and are often employed together.
Speed
Ethernet operates over the LAN and competes for bandwidth with other network traffic. Fibre Channel, on the other hand, operates on its network where the bandwidth is only available to storage traffic—FC offers higher, more consistent, and more predictable performance for storage traffic.
Lower Latency
The dedicated nature of the SAN benefits from FC having lower latency than Ethernet. FC traffic also provides a guarantee of lossless packet delivery. Ethernet can’t match that. Instead, it has a best-effort system and can sometimes suffer from dropped packets. As a result, delivery times can vary, latency increases and performance drops. Ethernet doesn’t offer the level of performance that many enterprise applications demand.
Cost
An FC SAN is more expensive than Ethernet using iSCSI. But cost is dependent on configuration—if an organization tries to use Ethernet for a very large amount of storage, the cost of more cables, switches, and cards will soon erase cost advantages and the performance.
Expertise
Ethernet is a far more prevalent technology than FC, and it’s easier to find trained and experienced technicians—the skills required for an FC SAN are highly specialized. Enterprises can expect to pay more for personnel that can operate and maintain them.
Bottom Line
Fibre Channel forms the backbone of the storage universe. When applications need super fast access to storage data, time and again, developers and application managers rely on block data and Fibre Channel to deliver the performance they need.
Read 8 Top Storage-as-a-Service Providers to see the best vendors for outsourcing enterprise storage management.
