Andrew, thanks for your comment. You’re right in that we do need to ensure that one type of traffic doesn’t clobber another type, but we already have QoS mechanisms in place to do that (NetIOC in vSphere, QoS in Nexus 1000V, QoS in UCS)—so why aren’t we using those technologies instead of reverting to separate uplinks? Logical link multiplexing, on its own, doesn’t address bandwidth management concerns. And you mention separation for sake of security but then turn around and talk about Palo vNICs—those aren’t separate, and the vNICs all share the same physical link. That implies that VLANs are an acceptable form of separation (in most instances), which then brings me back to the question: why are we placing so much emphasis on logical link multiplexing?
↧