DCB and TRILL have nothing in common

The emerging Ethernet bridging technologies have been hyped to an extent where the lines between them completely blurred, resulting in statements like “we need DCB and TRILL for FCoE”. Actually, none of that is true, but let’s focus on DCB and TRILL first.

TRILL and 802.1aq are Shortest Path Bridging (SPB) technologies. They replace the venerable Spanning Tree Protocol (STP) and all its variants and modify the frame forwarding rules. Replacing STP with TRILL is like replacing truly-stupid RIP (with maximum-paths set to one) with OSPF or IS-IS: you get faster convergence and better network utilization. The forwarding paradigm in network core is changed, so you could say a better analogy would be replacing RIP with IS-IS+MPLS.

DCB is a Quality-of-Service solution. Introducing DCB in your bridged network is like introducing IP QoS in your routed network.

SPB and DCB are completely independent, like OSPF and QoS are. You can use one or the other as needed in your network: DCB to give you standardized QoS behavior and lossless transport, SPB to give you faster rerouting and multipath forwarding.

You’ll get an overview of DCB, TRILL, 802.1aq and numerous other Data Center technologies in my Data Center 3.0 for Networking Engineers webinar (buy a recording or yearly subscription).

2 comments:

  1. hello Ivan , I think we can mention here about the draft http://tools.ietf.org/html/draft-eastlake-trill-rbridge-dcb-00 , section 5.4 , which proposes the integration with the Congestion Notification Message (CNM) into the trill c-plane , which you probably already analysed in more recent posts.

    Cheers ,
    michal
  2. Thanks for the comment and the link. In my opinion this is another (quite needed) kludge to fix a layering problem introduced by routing L2 frames across L2 domains.

    When will these people realize most of the problems they're solving (after creating them in the first place) were solved decades ago on L3 :-P
Add comment
Sidebar