Cisco Blogs

Will The Real FCoE Standards Please Stand Up

July 30, 2010 - 7 Comments

Earlier this week Cisco made an announcement with partners NetApp and VMware discussing end-to-end FCoE. Not surprisingly, in the storage space this was picked up by quite a few press outlets and retweeted across the Twittersphere. Unfortunately, many people – the press included – started throwing warnings up about how FCoE standards weren’t done, trying to throw a heavy douse of cold water on a very solid announcement.


As a result, in this, my inaugural post as a Cisco blogger, it wasn’t very difficult to try to come up with a topic that was both important and topical. Obviously, there is a huge misunderstanding about the standards process, particularly when FCoE is concerned. There’s more FUD thrown around than you-know-what in a monkey cage.


The problem: most people would rather gouge their eyes out with a grapefruit spoon than think about FCoE standards.


While the ostrich approach may seem like the path of least resistance, the downside is that in this case, what you don’t know can hurt you. Plus, you get sand up your nose.


FUD to the Left of Me, FUD to the Right of Me


As I mentioned, this week has been a field-day for the FUD-mongers. SearchStorage in particular came out with a double-whammy on July 29 with two articles that shed very little light (or truth) to the nature of where FCoE standards are in the process.


For instance, one article states that the “standards that enable FCoE have not been ratified.” (Bizarrely, the article immediately links to another article from a sister site discussing a book by Cisco’s Silvano Gai and Claudio DeSanti that “deals with the recently approved FCoE standard!)


On the very same day, another article states that the standards have been approved, but that “standards that enable FCoE between multiple FCFs remain under development” (in other words, multi-hop FCoE standards still need work). 


Neither of these claims, obviously, are correct. Then again, you probably already figured that out…


Breakin’ It Down


There are three things you never want to see being made: Sausage, the Law, and FCoE Standards. Pretty much all of it will make you want to lose your appetite.


Fortunately, I’ve been following the standards closely… so you don’t have to. (You’re welcome; I’m a giver.)


Now, going into details about FCoE standards would either send you into a coma or into an infuriated rage complete with pitchforks and torches outside of my office building (and trust me, my colleagues wouldn’t hesitate to throw me off the balcony just to see what happens), so let me give you some of the key highlights.


First, the FCoE standards are ratified. That was called FC-BB-5 and done in 2009, and published in May of 2010.


Second, FC-BB-5 also sets the standards for multi-hop FCoE. So, we’re not waiting for a different standard for that, either. This means that, despite SearchStorage claims to the contrary, the standards have been laid out and put to rest the issues “that enable FCoE between multiple FCFs.” (If you’re so inclined, you can check out sections and of the final draft of FC-BB-5.)


Third, while not directly connected to FCoE, the standards for Converged I/O (more commonly referred to as the Data Center Bridging (DCB) standards) have progressed far enough so that you can actually use FCoE in a real-world environment. Not every standard within the DCB family is relevant to running FCoE, and those that are still in the works are ones we don’t have to worry about to solve FCoE problems.


Fourth, it’s true that these standards bodies are working on new versions of standards. The mongers of FUD would have you believe that it’s important to wait until these new standards are set in stone. But these new versions are for new features and capabilities, not for changing established ones. Think of it like an Operating System: if you were to perpetually wait until the next version just for features that aren’t available right this very minute you’d never buy a computer!


What’s Next?


Without question there is more. So, so much more. You’ve probably gone through some of these with a “yeah, but what about…” question on your lips. 


Good! You’ve been suckered in! I mean, erm, perhaps I’ve managed to peak a little curiosity.


If that’s the case, I’ll be periodically bringing updates to address these questions over time. I’ll also be providing webinars to provide a little more in-depth understanding to the process and answer some of the specific, technical questions as well. I also periodically update with newsflashes on Twitter


I know I’ve glossed over a lot of details, and haven’t even hit some of the most egregious knee-slappers out there. In the interest of time, space, and your own sanity, I’ve kept the scope of this post limited. 


Nevertheless, we’re all interested in trying to figure out what’s going to happen next. It requires, however, being able to accurately assess where we are at the moment, and sometimes you just need a refreshing dose of accuracy. 


When it comes to understanding where we are with FCoE, as long as people keep using standards as an excuse to misinform and manipulate you, it makes sense to immunize ourselves with a basic knowledge of what’s really going on. 


There, that wasn’t so painful, was it?



In an effort to keep conversations fresh, Cisco Blogs closes comments after 60 days. Please visit the Cisco Blogs hub page for the latest content.


  1. Ronald – Thanks for joining the conversation. You raise an extremely valuable point, and one that deserves its own post (watch this space in the near future for more detail). Part of the problem (as I see it) is that people don’t actually know when standards are done,”” vs. still being in a draft phase. The standards bodies themselves don’t make this information easily accessible to the average layman (not that they aren’t transparent, just that most people don’t understand what certain terms mean), so how can they be expected to know when a standard has moved from draft to approval phases?To answer the question, we need to take the question in reverse: What problems need solving and where are the standards in development with respect to solving that problem?For FCoE, you have a need for solving the following problems:Placing Fibre Channel frames onto other media (e.g., Ethernet), andMaking the medium lossless.The first issue has a standard developed to address that problem. It’s T11’s FC-BB-5 and the technical draft was completed in October 2008, Submitted for publication in June 2009, and published May 2010. Since October 2008 there weren’t any technical changes (i.e., it was no longer in “”draft”” form).The second issue has *two* standards to address that problem. The first (though no one uses it because you can’t run other protocols with it) is PAUSE (802.3-2008), the second is PFC – 802.1Qbb. PFC – Priority Flow Control – is a document that is part of the DCB working group, and it has been submitted for publication in July, 2010. For what it’s worth, I have been told by the chair of that committee that PFC has not been technically changed since 2002 (it has been held up for reasons other than technological disputes).So, there is nothing related to FCoE in terms of standards that are in draft form that prevents you from implementing it in your data center – they aren’t going to change from underneath you. What you (and other customers) need to do, then, is look at how FCoE fits into your overall strategy from the perspective of value. But if all you’re waiting for is the standards to be “”done,”” your wait is over.”

  2. You said yourself Data Center Bridging (DCB) standards have progressed far enough so that you can actually use FCoE in a real-world environment.”” But just because you can use something in the real world doesn’t mean the standard has been ratified and set in stone.So it appears to me that Cisco is shipping products based, in part, on DCB standards that are still in draft form. Until all standards are fully ratified it doesn’t make too much sense to me to use it in an enterprise environment. There is a risk (however small you might claim) that changes in the final ratified standard could make current implementations incompatible. As an enterprise customer I will not deploy products based, even in part, on draft standards.”

  3. J.Long time bud… I think what Sid was talking about is the loseless factor in DCB, which keeps networks from having to send retransmits. While there are currently no other technologies other than FCoE and maybe NFS that can take advantage of this lossless network connectivity, it is exciting to see TLVs coming for iSCSI. I look forward to reading more of your blogs, glad to see you are with Cisco now.

  4. Illia,You are correct that there is a lot of misinformation and misunderstanding. I will continue to try to cut through the confusion by writing here, but Stu’s link to Wikibon is also a great place to look. Sid,Thanks for contributing. I’d like to make one small correction in your claim: while FCoE standards for mapping FC onto different types of networks is accurate, that fact does not in and of itself make those networks faster. It does, however, make the networks more flexible in the types of traffic that can get sent over a single medium, however.J

  5. I read wiki about FCoE and get worried about my brain.. There are such missunderstanding for me. Is there any new about FCoE in future? Is there any work regarding that?

  6. Hi,FCoE Standards for mapping of Fibre Channel over selected full duplex networks.That’s make network Faster.

  7. J,I agree that there are still plenty of misunderstandings about how the various standards for the FCoE ecosystem get put together. While Wikipedia has some of the pieces, it doesn’t tell the whole story. Wikibon has a wiki page on FCoE standards that was first created in 2009. I updated the page to reflect the latest standards updates. I would encourage you and others in the industry to edit the document as need so that we reach a consensus on the foundational technologies that make up this space.Thanks,Stu