To: "Robin Tasker" Cc: P8021 From: Peter Wang/HQ/3Com Date: 10 Apr 96 13:57:18 EDT Subject: Re: ["Robin Tasker": Priority in 802.1p] Robin, A well-engineered solution will involve the use of multiple mechanisms at different protocol layers. I don't know that there's been any precedence among the standards group to coordinate all of the related standards such that vendors can implement a well-engineered and yet standard-based solution. In reality, the higher layer protocols can take advantages of simple mechanisms provided at the link layer. Also, in today's environment, applications don't usually have direct access to class-of-service (or priority levels) being provided at the link layer. Similar to the fact that Microsoft will have to provide default QoS parameter set for backwards compatibility with station which don't speak ATM or RSVP, there will be default mapping from QoS, at the WinSock level, to CoS at the NIC driver level. Looking at this from a different perspective, a truely well-engineered solution that will address the QoS needs on a congested backbone will likely be too constraining. Just look at how complex ATM has gotten. If we're to look for a comprehensive solution at the link layer, it may be more useful if we should just take up ATM's work. In any case, I'd still claim the no well-engineered solution can address all of the issues. The applications and the users must take some responsibilities in the overall scheme of things. And they are. That's why a number of the multimedia applications, e.g. PictureTel's desktop videoconferencing system, includes a connection server which limits the number of simultaneous users of such product. Philosophical points aside, the simple mechanism that's being provided by the current .1p is useful if you consider the following: - application (or user) request QoS at WinSock level; - RSVP sets up the QoS with the routers and or the high function switches; - QoS is mapped to CoS at the link level; - switches at the edge of the bridged LAN provides the CoS. Backbone switches in the center of the network will likely implement some form of RSVP, i.e. more control than that offered by .1p alone. Also, the way .1p specifies CoS doesn't prevent any vendor from implementing more sophiscated management scheme within a switch in addition. Hence, it makes sense to keep the link layer CoS mechanism simple and let the higher layer build on it. Peter [attached: P802.1-96/040 (see d96/d96n040.txt)]