CCIE or Null!

My journey to CCIE!

Archive for August 2014

VN-Tag you are becoming more maintstream.

leave a comment »

VN-Tag, typically was a technology only seen in the Data Center (When using Nexus 2000 series FEXs)  however this has started changing. If you check out Cisco’s new 6800’s series Catalyst switch you’ll see they are now pushing a new ‘instant access’ model. This new model allows us to deploy ‘dumb’ switches that are centrally managed by the main 6800 Chassis essentially making those switches act as ‘external linecards’. Now in the Data Center this is nothing new, sounds just like the Nexus 2000 Series Fabric Extenders (FEX’s) right? Well, they are a little more similar then you think deep down. VN-Tag is the technology used by Nexus 2k FEX’s with their upstream parent switch, that same VN-Tag is also being utilized in the Cisco 6800 Series with its downstream external switches.

The VN-Tag technology itself simply adds an additional header to the packet as it traverses between the ‘Instant Access’ (Or FEX) and the it’s parent switch, where all the switching occurs. It is important to call-out the ‘Instant Access’ switches do NOT do any local switches (once again just like Nexus 2k FEX) all the information is forwarded up the 6500 or 6800 chassis where the packet will get switched or routed accordingly.

Let’s take a quick look at the VN-Tag information itself:


Surprisingly it’s nothing more than an additional 6-bytes. The fields are as follows:

  • EtherType [16-bits] = 0x8926
  • Destination Bit [1-bit] – Indicates which direction the frame is flowing.
  • Pointer bit [1-bit] –
  • Destination VIF – [14-bits] – Identifies the destination port.
  • Looped – Identifies the source vNIC, ment to identify multicast frames to ensure it is not forwarded back to where it originated.
  • Reserved [2-bits] – For future use.
  • Source VIF [12-bits] – vif_id of the downstream port.
  • Version [2-bits] – Set to 0

Expanding the VN-Tag to products outside of the Nexus line is definitely a great move, and it definitely has great use-cases in many large enterprises that want a more consolidated approach for management, after all this easy way to cut down on your Layer-2 spanning-tree topology. You can also cut down on configuration/implementation time, since this technology allows you to pre-provision the ports on an Instant Access before it is actually connected to the chassis.


Additional Info:

Cisco Instant Access White Paper

IEEE  802.1Q Whitepaper


P.S. – This was a blog post I had start middle/late last year when the 6800 & Instant Access were first announced. Finally got around to finishing it better late than never right?

Written by Stephen J. Occhiogrosso

August 27, 2014 at 11:34 AM

Posted in Cisco

Tagged with ,

FC vs FCoE

leave a comment »

FC vs FCoE infrastructure, usually a common debate when designing the network infrastructure of a new Data Center or new part of a Data Center, after all the advantages of running a converged storage & IP network are hard to turn down. Many of us are probably already aware of the why FCoE is always good option vs the traditional dedicated FC design, but I wanted to point out one interesting  fact that make FCoE more efficient and that fact stems down the Ethernet transport.

When FC frames are sent over a Fibre Channel network they are placed onto the physical medium in 8b/10b encoding which has an overhead of 25% to ensure the signal is sent successfully. (and intact with no corruption, essentially for every 8-bits, 2-bits are used to ensure the signal is still intact think of this an CRC or FCS for the electrical signals)

When we consider Ethernet, it is sent over the wire with a 64b/66b encoding, about a 3% overhead as the information hits the physical medium. So while FCoE may encapsulate the already large FC Frame with additional header & trailers FCoE from physical perspective is much more efficient means transmission. (So for every 64-bits sent to the wire, 2-bits are used for this ‘error checking’ and much better ratio than 8b/10b)

The 8b/10b encoding is used for 1, 2, 4, & 8 Gb Fibre Channel, the new 10+Gb Fibre Channel technologies are relying on 64b/66b encoded, which may tip the scales. However the converged infrastructure & cabling make FCoE the better option for most environments from a cost management perspective.

Random tidbit:

This is also an aspect the differentiates 100Mbps from 1Gbps from 10Gbps

  • 100Mb Ethernet –  4b/5b encoding
  • 1Gb Ethernet – 8b/10b encoding
  • 10Gb Ethernet – 64b/66b encoding


Written by Stephen J. Occhiogrosso

August 17, 2014 at 11:41 AM

VMworld sessions up for viewing. It’s free to!

leave a comment »


How’s that for a topic out of left field, never before seen on my blog. VMWare! Well, I got wind that WMworld sessions are available to watch for free so I just wanted to spread the word. All you need to do is create an account, you can find the sessions here. I’ll say It’s nice see vendors continue this trend, Cisco is already doing that with their CiscoLive365 site.

Written by Stephen J. Occhiogrosso

August 10, 2014 at 10:11 AM

%d bloggers like this: