back to article Network switches look different in the cloud

Cloud computing takes more than just a philosophical shift. It requires new skills, processes and architectures. In particular, traffic patterns in cloud networks can be quite different from those of the familiar enterprise network and the scale of operation can be significantly higher. That, according to experts in the field …


This topic is closed for new posts.
  1. kosh

    Layer 2 is a horrible solution.

    Article is pandering to switch vendor marketing without any critical thinking.

    Article is downright wrong.

    Flat layer 2 networks have horrible scaling, security, managability and reliability characteristics, all of which get solved by layer 3 segmentation. Not by reinventing the layer 3 wheel at layer 2 (c.f TRILL and other half-arsed horrors)

  2. Anonymous Coward

    Agility matters more than bandwidth

    Most of the marketing blurb here is about inter-rack bandwidth, which isn't that much of an issue except when you are moving VMs. If there are problems with the web app on rack1, the database on rack2 and another instance of the webapp on rack3, that's a problem of the placement tooling and client APIs, not "cloud" per se. If your API let you say "3 servers with 15 mpbs between them" the resource manager can put them close.

    Where VM-hosting clouds do have problems is dealing with change. Your routing tables are no longer fixed, you have many more VMs, each with a mac address, than you have ethernet ports. You need to update route tables as machines come and go, and if you have redundant switches, you need to keep those updates in sync.

    These are the problems that cloud infrastrastructure providers have, and if the switch vendors don't acknowledge that, then it shows they don't understand customer needs.

This topic is closed for new posts.