back to article Ethernet standards for hyper-scale cloud networking

What if the largest Ethernet networks we see today are just precursors, initial steps on the path to what's been called hyper-scale cloud networking? "Hyper" is the term used generally for something almost unfathomably and exceptionally large. We might say that a regional group of airports is a small air transport network, a …

COMMENTS

This topic is closed for new posts.
  1. pompurin

    Thanks I found this to be a well written and informative article. Surprised to see ISL being mentioned, it doesn't sound right. I thought Cisco had killed that off years ago?

    There are ways to load balance spanning-tree if you so wish. You can run multiple instances of spanning tree that span over different links. However if your traffic levels hit 50% and above you risk overloading your network in the event of a failover.

    1. kevinanchi
      Happy

      "It's possible, we suppose, but unlikely that new Ethernet standards will evolve to cope with this, ether by specifying latency at a switch level or at a network level. "

      We have that today too, it's called routing protocols, what do you think this hyper scale network with tens of thousands of switches would be a big flat layer 2 network? what a disaster that would be!

  2. Nate Amsden

    STP obsolete for over a decade

    STP has been obsolete for over a decade, multiple protocols from multiple vendors are have been good alternatives to STP long before TRILL was a twinkle in anyone's eye. It's kind of funny and perhaps sad that the majority of folks don't seem to have realized this.

    also the article doesn't mention the use of 802.3ad as a stopgap for faster ethernet, You can combine at least 8 links together to form a logical connection, with MLAG I think you can at least double that number if I recall right. This is of course widely deployed now and has been for years/decades(except MLAG which is kinda new)

    "It's possible, we suppose, but unlikely that new Ethernet standards will evolve to cope with this, ether by specifying latency at a switch level or at a network level. "

    We have that today too, it's called routing protocols, what do you think this hyper scale network with tens of thousands of switches would be a big flat layer 2 network? what a disaster that would be!

    Be mindful of the costs of 100Gbit ethernet too. If your running a cross pacific undersea cable then costs for the optics probably aren't a big deal, but for most other sorts of deployments (intra data center at least) 100Gbit ethernet just isn't worth it today from a cost perspective.

    1. A Non e-mouse Silver badge

      "You can combine at least 8 links together to form a logical connection, with MLAG"

      The problem with link aggregation protocols, is that they typically distribute connections between the links. Each connection can only achieve the maximum bandwidth of one link. (This also assumes that the protocol is smart enough to split two high bandwidth connections across two different links...)

  3. Graham Wilson
    Thumb Up

    Extremely High Speed Ethernet.

    I can only guess what the future holds for extremely high speed Ethernet or network generally but one thing's pretty certain which is that Ethernet has been very enduring.

    Given that it was invented almost 40 years ago (ca1973/4) by Robert Metcalfe and team at Xerox's PARC, it seems to have scaled remarkably well from its humble 10MHz beginnings.

    1. Anonymous Coward
      Anonymous Coward

      Backup...

      I've been a storage designer for nigh on 12 years, it's only in the last year or so that I've been seriously thinking that Ethernet may well be the way forward, rather than SAN attached hosts.

      There will for the time-being at least, still be a need for fibrechannel at the server side, but that's only because tape drives talk FC. If tape drives start to talk Ethernet, this could be the death-knell for FC. That said I won't right it off, there could be a lazarous-like recovery, when 8GB FC goes to 16GB that will be great for server-side tapes and disks, but 10GB at the client will still keep Ethernet in that tier.

      1. fdsa

        Re: Backup...

        I don't think there is an immediate need for all targets to have fcoe ports. The switches support both fc and ethernet ports. So use the switch to convert for now. Fcoe is just encapsulation of FC frames so it isn't much work at all for the switches to convert between the 2 protocols.

  4. Kirbini
    Pirate

    RE: STP, RSTP, TRILL et al...

    Doesn't OpenFlow networking make all the loop prevention protocols obsolete? I was under the impression that software defined networking could create dynamic, loop-free forwarding paths across any network topology thereby eliminating the need for loop control.

This topic is closed for new posts.

Other stories you might like