Question in regards to VC Active and\or passive configuration

[chuckk281]

There was a recent discussion over the benefits of "Active/Active" NIC teaming and "Active/Passive" configuration.

Eirik commented:

**********************************************************************************************************

Me and my collegues have been looking at the configuration of active – active and active – passive configurations of the VC-modules.

In the VC – cookbook the example is basically two drawings of the same cake, using different icing.

What is the real pros\cons of using one over the other, and why should a customer implement one or the other ?

**********************************************************************************************************

Cullen pitched in:

**********************************************************************************************************

This is personal opinion only, so take it for what it’s worth.

“Active-Passive” provides you with the same total upstream bandwidth in both a normal and failure mode.  You get the benefit of redundancy outside the enclosure without doing NIC teaming on the servers; however servers will lose some connections if one of the Virtual Connect modules is down.  With NIC teaming you can lose a Virtual Connect module without losing connections assuming you configure teams properly.

“Active-Active” provides you with the most available bandwidth in normal operation but only 50% of that in a failure mode.  You must use NIC teaming to take advantage of this level of redundancy – if all uplinks from one Virtual Connect module are down, you lose all connections to that network.  This is closer to how you’d operate with non-bladed servers.

*************************************************************************************************************

What do you think? Let us know what you do.