NIC fault tolerance

Home Forums Networking General Networking NIC fault tolerance

Viewing 1 post (of 1 total)
  • Author
    Posts
  • Avatar
    Blood
    Member
    #164244

    Hi

    As mentioned near the end of this thread, the network engineer made a recommendation that we install a second cable to introduce redundancy/fault tolerance.

    The servers that will use both these cables comprise:
    Windows Storage Server 2008
    Windows Server 2008 standard
    Windows Server 2012
    Windows Server 2012 R2

    Presently, these servers are connected to an 8 port desktop switch which is connected to an RJ45 connection. The CAT5e cable from the RJ45 socket then terminates at a patch panel.

    A second cable will be installed next to the present cable and will connect to the same patch panel.

    My question is what is best practice regarding configuration of the network cards. I have read that only Server 2012 onwards supports teaming, and that the use of teaming on earlier versions of Windows can introduce more problems that it solves.

    Firstly, definitions. As I understand it ‘teaming’ groups network cards together so that they appear as one interface. This is also referred to as ‘Link Aggregation’ or ‘Load Balancing and Failover’.

    Because we will have two physical cables from the server space to the patch panel, would it better to assign separate IP addresses to each network card and then rely on Round Robin to ensure that both cards are equally utilised? This described in section 2.3 of this blog post. This would seem to be a simpler solution.

    The priority is to provide fault tolerance, not extra bandwidth.

    Anyone have any thoughts about this?

Viewing 1 post (of 1 total)

You must be logged in to reply to this topic.