|
Hi Lou, see comments below. This is a key point and one that wasn't
clear until we split off the compatibility draft from the framework and
had some questions on the list. Regards Greg Lou Berger wrote: --> This document extends the GMPLS control plane to include "asymmetric" switches. No matter how they are implemented.Authors, (WG) I have a pretty basic question on the WSON framework document: What is the scope of the optical networks covered by this framework? I think the document should be very clear on this. Keep in mind that a framework (or even protocol) may allow for, i.e., not preclude, functions and aspects of an architecture or solution without fully defining them. To elaborate on my question: A) It's clear that some portions of optical networks are in scope, such as: "wavelength division multiplexed links, tunable laser transmitters, reconfigurable optical add/drop multiplexers (ROADM) and wavelength converters". But it is unclear if other portions are in or out; in particular regenerators, OEO switches and, to some degree, OXCs. All are mentioned in the document, but the document is a bit vague on whether they are part of the framework or not. So what's included and what's excluded? It also includes models for shared wavelength converters and detailed link-wavelength usage. Such information is useful across many WSONs. What it doesn't model or cover is signal constraints imposed by particular technology implementations such as an OEO wavelength converter. It also would not be aware of regenerators since it doesn't understand differences among optical signals or impairments. We've moved such text out to the compatibility and impairment drafts. --> Yes. But no checking of signal compatibility is provided.My personal opinion is that the framework should allow for all aspects of today's optical networks, but need not define support for each (e.g., impairment based RWA). B) Are multiple signal types included or not? ---> Hmm, we say in the abstract:The document seems to say that they are not, but then there is section 3.2.2. It seems to me that it really doesn't take much to allow for support of multiple signal types in this document. Given that multiple signal types are a reality in today's optical networks, I personally think they belong in the framework. "nor does it address potential incompatibilities between some types of optical signals and some types of network elements and links" This is a fairly general statement in line with the scope that we intend. But confusingly we say in section 3.2.2: "As pointed out in the introduction, for the purposes of this document we assume that all optical signals used in a WSON are compatible with all links, network elements, and receivers in that WSON." This is too restrictive since it says all signals should be compatible with all links, elements, receivers all over the network. We should probably say something like: As pointed out in the introduction, for the purposes of this document we assume that all optical signals used in a WSON are compatible with the links, network elements, and receivers in that WSON that they traverse or terminate on. A "signal" may change its characteristics as it traverses a path, e.g. modulation format, or FEC, particularly at a regeneration point. For this draft we assume that the compatibility of signals and elements has been pre-planned and is not a control plane function to worry about. We split this off as the signal compatibility aspects grew and based on the chairs (Deborah's) comments in Stockholm. See the latest compatibility draft for more complete analysis of the compatibility issues with regenerators, OEOs and the like. I have a bunch of more minor comments on the document, but will hold these until the above is closed. Much thanks, Lou -- =================================================== Dr Greg Bernstein, Grotto Networking (510) 573-2237 |