Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

RE: [EFM] EFM Requirements > Talking about voice transport...


See comments mixed in below...
At 08:37 AM 8/24/01 +0100, Bob Barrett wrote:

> Talking about voice transport...

Voice in p2p:

If we mess with the PHY (unlikely) then carrying voice becomes easy (can
have its own channel effectively, outside of the 1GE).

If we don't mess with the PHY (probable) then carrying voice requires some
QoS mechanism, either within the EFM spec.,

Exceedingly unlikely, especially with the proposed fiber solutions. There MAY be a need for QoS elsewhere in the network but with the bandwidth available in the links proposed vs offered traffic there is very little chance that there will be any interference with voice. Over-provisioning has been beating out QoS as the preferred solution method for Ethernet for quite some time now.

or above it. If above it, then
the upper layer would need to control _all_ packets, not just the time
critical packets, else we are back in the non-deterministic realm. If it's
above EFM then it's an application / implementation specific issue.

Roy - I am not talking about voice over the internet here. We are talking
about voice over well controlled private packet networks (owned and
controlled by service providers), so they are reasonably quantifiable :-).

Bob B.