Sales up for 16 Gbps Fibre Channel; is it going the way of the dodo?

Sales for 16 Gbps Fibre Channel swell to their highest volume since 2008, but questions remain about the technology's future.

It was the best of times; it may soon be the worst of times. Fibre Channel technology, which has been available since 1988, is seeing a significant boost in sales because of a recent doubling in speed from 8 Gbps to 16 Gbps. But in the long term, its future may not be so bright.

Fibre Channel (FC) emerged as a way to move data from large central servers to adjacent storage systems. Approved as an American National Standards Institute standard in 1994, the standard's top speed has been regularly upgraded, starting at 1 Gbps and now reaching 16 Gbps.

"Every technology must have a roadmap to survive, and [16 Gbps] FC is just the next chapter in Fibre Channel's long history," said Sam Barnett, directing analyst for data center and cloud at Infonetics Research.

The 16 Gbps specification was completed in the fall of 2011. In addition to higher speed, the new standard features include the following: lowered use of power; backward compatibility with previous versions of the standard; 60% to double the improvements in storage area network I/O density; 60% to twice as many virtual machine I/O channels per port; and the use of fewer PCIe slots than the previous version of the standard.

Various adapter, switch, storage and server suppliers have been developing compliant products. In December 2011, the Fibre Channel Industry Association (FCIA), a nonprofit international organization of manufacturers, systems integrators and vendors, held a Plugfest to test different devices; Amphenol Corp., Broadcom Corp., Brocade, Cisco, Data Center Systems Inc., Emulex Corp., EMC Corp., Hewlett-Packard Co., JDS Uniphase Corp., NetApp Inc., QLogic Corp. SANBlaze Technology Inc. and Teledyne LeCroy took part in the test held at the University of New Hampshire Interoperability Lab.

A couple of factors pushed development of the faster storage system networking option.

"The move to cloud computing is moving companies away from dispersed systems to centralized data centers," said Mark Jones, director of tech marketing at Emulex and vice chairman of the FCIA. As information becomes more concentrated, enterprises move large volumes of data over short distances rather than small volumes of data over larger distances and therefore need more bandwidth in the data center.

In addition, server and storage technologies have become faster and thus may strain existing connections. The upcoming launch of servers powered by Intel Corp.'s new Romley platform should increase the volume of data being pushed by servers. Larger network pipes will be needed to support the higher I/O rates.

When it comes to Fibre Channel products, there are a limited number of suppliers in each market segment. Brocade has been the vendor most aggressively pushing 16 Gbps Fibre Channel switches. The company has held that market segment virtually to itself, and according to initial reports 16 Gbps now represents about 30% of its new director class switch sales.

Cisco has been a bit of market straggler, saying there is not a lot of customer demand for these products, according to Barnett. However, Cisco has been much more proactive in promoting movement to Fiber Channel over Ethernet (FCoE) products, since they fit more cleanly with its legacy systems than Fibre Channel does.

More competition is evident in other market segments. Vendors such as Broadcom and Emulex have begun delivering 16 Gbps Fibre Channel server adapters for HP, Dell and IBM servers that support virtualization (initially just VMware software) and I/O management functions. Storage suppliers such as EMC, NetApp and QLogic have begun adding Fibre Channel support to their devices.

The first wave of 16 Gbps Fibre Channel products began shipping at the end of 2011. Initial customer interest has been high and provided the market with a much-needed boost. In fact, Dell'Oro Group reported that revenue for Fibre Channel products reached its highest level in four years, at $1.3 billion during the first six months of this year.

"Many corporations see [16 Gb] Fibre Channel as a way to future-proof their storage networks," said Jones.

However, as with any new technology, sometimes issues arise involving interoperability and management functions. So far, such problems have seemed miniscule.

"We used the same interfaces with [16 Gb] Fibre Channel that we relied on in earlier versions of the standard," said Jones.

Additionally, when new products first start to ship they often come with premium price tags, and that has been the case with 16 Gbps Fibre Channel solutions. The per-port connection costs started off about 50% greater than comparable 8 Gbps connections. The pricing trajectory has been moving downward, but a few years will pass before the two will be in sync. In sum, as pricing drops and technical advances push more data over storage connections, keen interest in faster-speed Fibre Channel seems likely during the next few years.

So what's next? The FCIA has been developing a standard for 32 Gbps Fibre Channel products. Draft versions of the specifications should be ready later this year, and products are expected to start arriving in 2014. The group has even outlined a long-term roadmap with 512 Gbps products being delivered in 2024.

But how much longer will Fibre Channel remain viable? The data center has seen significant movement away from autonomous systems and toward more integrated networks and data center devices. A version of FCoE was developed in 2007, and it has been steadily gaining popularity in the marketplace. Ethernet dominates the server market and has made significant progress in the wide area network arena. Moving storage solutions to an Ethernet network provides greater cohesion than continually upgrading to faster versions of Fibre Channel.

Currently, Infonetics pegs FCoE at about 15% of the total data center Ethernet switch market and thinks that will continue to increase. In sum, Fibre Channel remains a viable storage networking solution in the short term, but its long-term prognosis is not as certain.

ABOUT THE AUTHOR: Paul Korzeniowski is a freelance writer who specializes in data center issues. He is based in Sudbury, Mass., and can be reached at paulkorzen@aol.com.

This was last published in October 2012

Dig Deeper on SDN and other network strategies

PRO+

Content

Find more PRO+ content and other member only offers, here.

Join the conversation

17 comments

Send me notifications when other members comment.

By submitting you agree to receive email from TechTarget and its partners. If you reside outside of the United States, you consent to having your personal data transferred to and processed in the United States. Privacy

Please create a username to comment.

Does Fibre Channel have a future?
Cancel
I dont see other alternatives coming up strongly at the moment.
Cancel
Bright future – I wont move from Fibre channel anytime soon.
Cancel
FC still is the best protocol for high performance storage service
Cancel
Thats what I am selling all the time
Cancel
Even though we use both FCoE and FC, the connection from the Nexus switch to our storage (NetApp) is FC. I see us going to straight FCoE but we still have to run plain FC to the storage system. I'm not real sure that 32 or 256 Gbps if FC is needed.
Cancel
FCoE is not all that vendors make it out to be. Virtualization rations and over subscription brings this home. A number of servers doing heavy I/O over a single ethernet connection even at 10GB is still CSMA/CD. The CSMA/CD protocol functions somewhat like a dinner party in a dark room. Everyone around the table must listen for a period of quiet before speaking (Carrier Sense). Once a space occurs everyone has an equal chance to say something (Multiple Access). If two people start talking at the same instant they detect that fact, and quit speaking (Collision Detection.)
Cancel
Can't beat its reliability. Worked on Mcdata and now Brocade for years and haven't a single issue. Using Cisco switches with NFS, well that is another story.
Cancel
Contrary to urban myth and FUD, I don't see Fibre Channel going away any time soon. A "compelling argument" for moving data to storage over FCoE has simply failed to materialized.
Cancel
At our current era, The Data is moving so fast and it requires more speed to have it access or to move a files, so that the information will be readily available for the user to access.
Cancel
In many (and perhaps most) instances, Fibre Channel remains the best possible solution for Data Center applications.
Cancel
As virtulization & cloud computing demands more Bandwidth Fibre channel is the only present option available till some other medium is found.
Cancel
On the long run it is clear that storage will also be a cloud-service. The products that support this are being developed. Look at the UDS product of Huawei, designed for CERN. These products will not use FC, but standard ethernet.
leon lips
Cancel
Ethernet is rapidly catching up to FC, with cost optimization in mind, not needing to add HBAs to the entire infrastructure.
10GbE and now we see a few devices also with 40GbE links while FC is still planning for 32Gbps
Cancel
Its is a mature and stable technology
Cancel
Convergency - efficiency - openess - non vendor locking , will gradually phase out FC.
Cancel
I certainly hope FC is not going away as we are replacing our existing iscsi network with FC due to growth and an increase in VDI workstations. 16g FC is the answer for us now. It is not as inexpensive as iscsi, but not completely out of reach. Competition between iscsi and FC has been a benefit for us.
Cancel

-ADS BY GOOGLE

SearchWindowsServer

SearchEnterpriseLinux

SearchServerVirtualization

SearchCloudComputing

Close