Comments on: FCoE: Framing deep dive http://blog.nigelpoulton.com/fcoe-framing-deep-dive/#utm_source=feed&utm_medium=feed&utm_campaign=feed with nigel poulton Tue, 20 Dec 2011 19:37:54 +0000 http://wordpress.org/?v=2.8.6 hourly 1 By: Nigel Poulton http://blog.nigelpoulton.com/fcoe-framing-deep-dive/comment-page-1/#comment-553 Nigel Poulton Mon, 12 Oct 2009 18:59:59 +0000 http://blogs.rupturedmonkey.com/?p=544#comment-553 Hi Stephen,<br /><br />No takers on the job market yet....... might have to take a non FCoE related role :-(<br /><br />As for FC-CMP.  I have to hold my hand up and say I'm not sure......  But then Silvano Gai is the "geezer" when it comes to FCoE and is on the bleeding edge. Not only am I several steps behind him - I am several light years behind.  If I find out though I will let you know more and probspost on it. Hi Stephen,

No takers on the job market yet……. might have to take a non FCoE related role :-(

As for FC-CMP.  I have to hold my hand up and say I’m not sure……  But then Silvano Gai is the "geezer" when it comes to FCoE and is on the bleeding edge. Not only am I several steps behind him – I am several light years behind.  If I find out though I will let you know more and probspost on it.

]]> By: stephen2615 http://blog.nigelpoulton.com/fcoe-framing-deep-dive/comment-page-1/#comment-552 stephen2615 Sun, 11 Oct 2009 07:19:54 +0000 http://blogs.rupturedmonkey.com/?p=544#comment-552 opps. FC control and management plane no plans. opps. FC control and management plane no plans.

]]> By: stephen2615 http://blog.nigelpoulton.com/fcoe-framing-deep-dive/comment-page-1/#comment-551 stephen2615 Sun, 11 Oct 2009 07:18:48 +0000 http://blogs.rupturedmonkey.com/?p=544#comment-551 Any takers on the job market Nige?  You might have to work with some MDS's..   We have someone keen at work to do something.  Pity about no hardware or budget and to make it worse, almost 95% of our thousands of servers are blades. I was reading some of the files on the t11 site for FCoE and they do make for good err umm reading.  Cisco has a document written by Silvano Gai called: FCoE end -to-end: adding details (T11/09-539v0) <a href="http://www.t11.org/ftp/t11/pub/fc/bb-6/09-539v0.pdf" rel="nofollow">http://www.t11.org/ftp/t11/pub/fc/bb-6/09-539v0.pdf</a> and it mentions something called a FC-CMP or FC Control and Management Plans.  This appears to be servers.   A Google search offers nothing.  I tend to not be bothered with the techical details of whats in a frame but when I see a server doing something in the middle of this cloud, I have to ask.  What is this FC-CMP?   Any takers on the job market Nige?  You might have to work with some MDS’s..   We have someone keen at work to do something.  Pity about no hardware or budget and to make it worse, almost 95% of our thousands of servers are blades.
I was reading some of the files on the t11 site for FCoE and they do make for good err umm reading.  Cisco has a document written by Silvano Gai called:
FCoE end -to-end: adding details (T11/09-539v0)
http://www.t11.org/ftp/t11/pub/fc/bb-6/09-539v0.pdf
and it mentions something called a FC-CMP or FC Control and Management Plans.  This appears to be servers.   A Google search offers nothing.  I tend to not be bothered with the techical details of whats in a frame but when I see a server doing something in the middle of this cloud, I have to ask.  What is this FC-CMP?
 

]]> By: mike http://blog.nigelpoulton.com/fcoe-framing-deep-dive/comment-page-1/#comment-550 mike Sat, 10 Oct 2009 00:14:46 +0000 http://blogs.rupturedmonkey.com/?p=544#comment-550 Nigel, thanks for doing these. They are very helpful and I hope you do continue them.Brad, I guess I don't understand why the dmac would change in FCoE. Why wouldn't the DMAC remain the target HBA? Hrm. I have to go read some more.  :)Thanks! Nigel, thanks for doing these. They are very helpful and I hope you do continue them.Brad, I guess I don’t understand why the dmac would change in FCoE. Why wouldn’t the DMAC remain the target HBA? Hrm. I have to go read some more.  :) Thanks!

]]> By: Nigel Poulton http://blog.nigelpoulton.com/fcoe-framing-deep-dive/comment-page-1/#comment-549 Nigel Poulton Fri, 09 Oct 2009 21:55:58 +0000 http://blogs.rupturedmonkey.com/?p=544#comment-549 Hi Mike,<br /><br />Good spot, I was referring to to hops between FCFs. I will edit the post and clarify that point.   However, what Brad says is true (obviously as he is a Data Centre networking legend).<br /><br />I may at some point put up a post talking in more detail about the encapsulation process including FCFs and FCoE_LEPs etc....  But as this is kind of a tutorial series of posts it would have been too much to include that detail in this post.  IF I make them too long then nobody will read them ;-)<br /><br />Thanks for joining the dicsussion and poiting out the ambiguity.<br /><br />@Brad thanks for keepings things straight<br /><br />Nigel Hi Mike,

Good spot, I was referring to to hops between FCFs. I will edit the post and clarify that point.   However, what Brad says is true (obviously as he is a Data Centre networking legend).

I may at some point put up a post talking in more detail about the encapsulation process including FCFs and FCoE_LEPs etc….  But as this is kind of a tutorial series of posts it would have been too much to include that detail in this post.  IF I make them too long then nobody will read them ;-)

Thanks for joining the dicsussion and poiting out the ambiguity.

@Brad thanks for keepings things straight

Nigel

]]> By: Brad Hedlund http://blog.nigelpoulton.com/fcoe-framing-deep-dive/comment-page-1/#comment-548 Brad Hedlund Fri, 09 Oct 2009 21:13:30 +0000 http://blogs.rupturedmonkey.com/?p=544#comment-548 Mike, If there is an intermediate Ethernet switch between the CNA and FCoE FCF, correct, the intermediate switch does not change any encapsulation, just forwards. However if the FCoE FCF sends the frame to another upstream FCoE FCF, the encapsulation will change (new DMAC of the new next hop FCoE FCF). If you haphazardly design your Ethernet the shortest path may not be guaranteed, no different than a bad FC design. Designing Ethernet and FCoE with common sense best practices does not need TRILL. Cheers, Brad Mike,
If there is an intermediate Ethernet switch between the CNA and FCoE FCF, correct, the intermediate switch does not change any encapsulation, just forwards. However if the FCoE FCF sends the frame to another upstream FCoE FCF, the encapsulation will change (new DMAC of the new next hop FCoE FCF).

If you haphazardly design your Ethernet the shortest path may not be guaranteed, no different than a bad FC design. Designing Ethernet and FCoE with common sense best practices does not need TRILL.

Cheers,
Brad

]]> By: mike http://blog.nigelpoulton.com/fcoe-framing-deep-dive/comment-page-1/#comment-547 mike Fri, 09 Oct 2009 20:49:13 +0000 http://blogs.rupturedmonkey.com/?p=544#comment-547 <em>To put this in perspective, here is as good a place as any to point out that the encapsulation and extraction process occurs at every hop on the FCoE network (we won’t talk about multi-hop FCoE here).  In other words, every time an FCoE frame traverses and FCoE switch, it is re-encapsulated ready to be moved on to its next hop. </em>huh?<br /><br />The outer encapsulation of FCoE is just an Ethernet. Why would it need to be decapsulated and encapsulated at each hop. That would be silly. The MAC address doesn't change within the broadcast domain. The reason why multi-hop FCoE is not widely supported (Cisco does it using thier own methods) is because you can't yet guarantee shortest path through the switched network. <br /><br />Isn't that what TRILL is for? To put this in perspective, here is as good a place as any to point out that the encapsulation and extraction process occurs at every hop on the FCoE network (we won’t talk about multi-hop FCoE here).  In other words, every time an FCoE frame traverses and FCoE switch, it is re-encapsulated ready to be moved on to its next hop. huh?

The outer encapsulation of FCoE is just an Ethernet. Why would it need to be decapsulated and encapsulated at each hop. That would be silly. The MAC address doesn’t change within the broadcast domain. The reason why multi-hop FCoE is not widely supported (Cisco does it using thier own methods) is because you can’t yet guarantee shortest path through the switched network.

Isn’t that what TRILL is for?

]]> By: Brad Hedlund http://blog.nigelpoulton.com/fcoe-framing-deep-dive/comment-page-1/#comment-546 Brad Hedlund Fri, 09 Oct 2009 20:14:52 +0000 http://blogs.rupturedmonkey.com/?p=544#comment-546 Another brilliant post Nigel! Keep'em coming! Cheers, Brad Another brilliant post Nigel! Keep’em coming!

Cheers,
Brad

]]>