Is it finally time to consider Fibre Channel over Ethernet (FCoE) for your storage network?
When Fibre Channel over Ethernet (FCoE) arrived on the storage networking scene several years ago, many expected a rapid increase in FCoE adoption, which offers high-speed Fibre Channel traffic and input/output (I/O) consolidation. This did not happen for several reasons, including:
- Lack of a significant performance advantage due to the availability of 8GB (and now 16GB) Fibre Channel;
- Vendor feature/function inconsistencies leading to interoperability concerns;
- Organizational disruption and confusion;
- Availability of other "converged" networking options, like Internet Small Computer System Interface (iSCSI) and Network File System (NFS); and
- Limited cost advantages.
If an organization is already heavily committed to Fibre Channel for storage and is undergoing a data center redesign or a major technology refresh, then FCoE should seriously be considered.
However, users eventually overcame these initial hurdles, leading to an increase in FCoE adoption. This growth has been aided by the inclusion of necessary support elements into vendor network, server and storage offerings. Data Center Bridging (DCB), which includes lossless Ethernet capability (among other useful functions), is now included in most network devices; server vendors are incorporating Converged Network Adaptors (CNAs) in blade servers; and storage arrays include native FCoE support.
Currently, the two major benefits of FCoE adoption are:
- The ability to converge to a single type of network, which greatly simplifies cabling and may, under the right conditions, reduce management and device costs.
- The ability to leverage expanded bandwidth of 40 GB Ethernet.
Therefore, there really are two main decision factors:
- Converged vs. non-converged network -- this is a question of timing (i.e., when, not if).
- Once converged, whether to go with FCoE or an alternative.
The possibility of cost savings from FCoE adoption depends on the specific scenario. There are other options available for convergence, including iSCSI and file-based protocols, such as SMB and NFS. If an organization is already heavily committed to Fibre Channel for storage and is undergoing a data center redesign or a major technology refresh, then FCoE should seriously be considered, as it protects its investment and leverages existing skills and capabilities. If the company is growing only incrementally, such a transition may be less cost-effective.
While there is a potential for 32GB Fibre Channel in a few years, the future path for Ethernet is more clearly established with 100 GB planned to follow 40 GB. So, Ethernet will have a growing bandwidth advantage for the foreseeable future. Organizations leveraging Fibre Channel may want to transition to FCoE now, in order to be better positioned to take advantage of increased bandwidth as it becomes available. Of course, the bandwidth advantage of Ethernet also benefits the other protocols mentioned above; so ultimately, it is critical to determine which protocol options are most appropriate for the organization.
This was first published in June 2013