Its still not about SAN connectivity alternatives
Updated: Aug 6, 2020
Its been long said that rumors of the death of Fibre Channel are greatly exaggerated. A stolen line I used way to often. While the statement is humorous, it is reflective of a simple truth - FC is a hard technology to kill off.
Now having said that, I don't want to say that FC is immortal, but I am also not willing to forecast its death - at least not yet. There has been considerable evolution in the storage market that have been potential negatives for SAN; scale-out storage architectures versus scale-up, Software Defined Storage (SDS), hyper-converged infrastructures. Additionally, resurgence of Direct Attach Storage (DAS), increased capacity of Network Attached Storage (NAS), and the migration of workloads into the Cloud have created headwinds for FC. On the other hand there have also been some changes that are positive such as, all-flash arrays, and in NVMe (non-volatile memory express), and NVMe over fabrics.
However, when looking specifically at block-based storage, FC remains the lion share of the market. Alternative connectivity types just haven't been able to take much share from FC. While FCoE, iSCSI, and Infiniband have found some success in niche markets, they have not enjoyed widespread adoption. The long standing benefits of FC remain attractive after all these years. Things like speed, security, stability, and the fact that FC networks are separate from the rest the network still warms the hearts of storage admins.
For the time being, existing customers will continue to choose FC for mission-critical enterprise workloads. Trends like flash arrays and the potential of NVMe over fabrics will provide oxygen for this mature market. Also, product refreshes and increased speeds will continue drive sales into the installed base.
Next time I will dig a little deeper into NVMe and its expected impact on the SAN market.