Sep
05

This past week Mark Snow and I completed the first of three of our CCIE Data Center Live Online Bootcamps – Nexus Switching. This class focused on the core Layer 2 Switching and Layer 3 Routing features of Nexus NX-OS on the 7000, 5500, and 2000 platforms, and the Data Center specific applications of the platforms with technologies such as vPC, FabricPath, and OTV, just to name a few. The videos from class are now in post-processing, and will be available both for download and in streaming format, both of which have cross platform support (Desktop, iPhone/iPad, Android, Windows phone, etc.) Also as usual our videos are DRM free, so once you purchase them they are yours to do with as you please.

All Access Pass subscribers will get access to the videos in streaming format for no additional fee, and can also purchase the download version at a discounted rate. The download version can be purchased standalone here for people who are not AAP subscribers. Our next two classes, CCIE Data Center Storage and CCIE Data Center Unified Computing are coming up at the end of September and October respectively. All AAP subscribers can attend the live online classes for free, while anyone who want to purchase the download in advance also gets access to attend these classes.

Below are some excerpts from the class relating to the new FabricPath technology. FabricPath is a new alternative to running Spanning-Tree Protocol in the Layer 2 DC Core, and is a pre-standard version of the TRansparent Interconnection of Lots of Links (TRILL) feature. The videos below cover the underlying theory of FabricPath, it’s basic configuration, it’s more advanced configurations and verifications, and its integration with Virtual Port Channels (vPCs) with the vPC+ feature.

Enjoy!

CCIE Data Center :: Nexus :: FabricPath Overview

CCIE Data Center :: Nexus :: FabricPath Initial Configuration

CCIE Data Center :: Nexus :: FabricPath Configuration & Verification

CCIE Data Center :: Nexus :: FabricPath & vPC+

About Brian McGahan, CCIE #8593, CCDE #2013::13:

Brian McGahan was one of the youngest engineers in the world to obtain the CCIE, having achieved his first CCIE in Routing & Switching at the age of 20 in 2002. Brian has been teaching and developing CCIE training courses for over 8 years, and has assisted thousands of engineers in obtaining their CCIE certification. When not teaching or developing new products Brian consults with large ISPs and enterprise customers in the midwest region of the United States.

Find all posts by Brian McGahan, CCIE #8593, CCDE #2013::13 | Visit Website


You can leave a response, or trackback from your own site.

23 Responses to “FabricPath – CCIE DC Nexus Switching Class Preview”

 
  1. GS says:

    Few quick questions:

    > When DC rack rentals will be available
    > When DC workbooks will be released for different levels (Initial, Medium, Advance, Mock labs & TS)
    > Will INE be coming with Deep Dive classes for DC Topics since that I personally feel would be needed
    > Any reason for choosing same price for Download Version Vs Live Class. Since usually Download version is cheaper.

    Thanks!

  2. Joshua Walton, CCIE #19763 says:

    Well done, Brian & Mark. Keep up the good work. Hopefully I pass my R&S in April and then I will be focusing on this track.

  3. mhon says:

    Thanks a lot INE!
    Is the price of downloadable version same to the online class?

    • Yes, the price is the same. When you buy the live class you get the download, and vice versa. If you have an All Access Pass subscription you get access to the live class and the streaming recordings, but not the download version.

  4. GS says:

    When the material will be added to AAP ?

  5. Alexander Lim says:

    Just finished watching the first two FP VoD. It’s really cool.
    d(^ ^)b Keep the good work INE!!!

  6. Amit says:

    Hi Brian,

    Great work on the videos.

    Just a question regarding the usage of Multi-destination Trees (by default, 2 trees are created max.) for Broadcast/Multicast/Unknown Unicast (BUM) traffic – there seems to be a better provision for handling BUM traffic in TRILL than FP where the Switches announce what VLAN traffic they want to receive. In FP, the BUM traffic is just flooded to every Switch via the Root switch.

    I have heard that TRILL implementation in NX-OS would be available by just a software upgrade, do you think the FP frame format and handling of BUM traffic will also change to conform with TRILL standard?

    And an added question – TRILL also seems to be better placed to handle smth like FCoE due to its hop-by-hop forwarding, while FP cannot be used for FCoE. Do you think that will happen as well?

    Thanks.
    Amit.

    • I doubt that the FabricPath format will change to conform to TRILL, since FabricPath is already implemented in the ASIC. The main difference to remember is that TRILL is Ethernet. FabricPath is not Ethernet, it is FabricPath. I have a feeling that in the log run FabricPath will be aged out by TRILL, as in the end the standards based protocols usually win out. The difference though is that FabricPath is here now, and is being implemented. Who knows how long it will take TRILL to actually be implemented and field tested.

      • Amit says:

        I was wrong about inefficient BUM traffic forwarding in FP. Here’s an excerpt from an article on Cisco’s website-

        “While conceptually every Cisco FabricPath VLAN is carried on every core port, in practice, the presence or absence of a particular VLAN ID in the VLAN database of a particular switch can influence whether traffic for a particular VLAN traverses a particular Cisco FabricPath switch. In fact, each Cisco FabricPath switch explicitly advertises its interest in each VLAN configured on the switch. The advertising of VLAN interest mimics the mechanism used to constrain IP multicast traffic, as described in Section ‎5.5, “Multicast Routing.” When a switch advertises its interest in a VLAN, it floods a GM-LSP through the Cisco FabricPath domain, advertising its interest to other switches. Other Cisco FabricPath switches build a forwarding state for multidestination frames on a per-VLAN basis, constraining multidestination traffic to only those core ports that need it, helping to conserve link bandwidth in the fabric.
        If a VLAN does not exist on a given Cisco FabricPath switch, that switch does not advertise interest, and neighboring switches will not forward multidestination frames for that VLAN on core ports connected to that switch. Use care in using this behavior to control where VLAN traffic propagates, because VLAN traffic will not transit a particular Cisco FabricPath switch if either the VLAN does not exist or the VLAN is not configured for Cisco FabricPath mode.”

        • Surya says:

          “If a VLAN does not exist on a given Cisco FabricPath switch, that switch does not advertise interest, and neighboring switches will not forward multidestination frames for that VLAN on core ports connected to that switch. Use care in using this behavior to control where VLAN traffic propagates, because VLAN traffic will not transit a particular Cisco FabricPath switch if either the VLAN does not exist or the VLAN is not configured for Cisco FabricPath mode.”

          As far as I know (and according to a Cisco consulting Engineer) this part is wrong; that’s why Cisco created multi-topologies aware FP; multicast and broadcast trees are shared among all FP switches; if at some point a VLAN is missing on a switch; then you may face a blackhole, because the tree topology is unique regardless the way Vlans span the network.

  7. Phil G says:

    As usual – great info and looks a good purchase. One thing Brian mentioned in the FP VoDs was that F2 linecards are L2 only – this is not true they can do L3 routing also, but just aren’t as feature rich in this space as M series linecards.

  8. Akash says:

    Will LISP,& MPlS be part of the CCIE-DC scope?

 

Leave a Reply

Categories

CCIE Bloggers