There have indeed been many times when people have asked about how things work, and have used layers 1 through 3 for solving issues. Both the OSI model and the TCP/IP model work very well for that - they have those layers pretty much in common.
But after that? In my longish tenure on this list I don't remember (and this could be early senility, I suppose) any differentiation of layers above that, aside from a few tongue-in-cheek references to layer 8. Pleased to meet you, glad to know your name, Mr. D.A. :) You switch models when the circumstances require it. I can't speak to OOP, as I'm not a programmer, but you can, in theory, reduce all of chemistry to quantum mechanics - it just doesn't make sense to do so for most industrial applications. Ditto for rocketships to the moon. It seems reasonable to me that you should use the model that allows you to solve the problem with the least amount of effort. Kurt On Sun, May 30, 2010 at 23:41, Ken Schaefer <[email protected]> wrote: > There are many times that people have asked questions *on this list* that > could have been solved with a rudimentary knowledge of how networking > protocols work, and what are pre-requisites for a particular protocol or > application to work. > > For IIS, there are always many questions on whether a problem is with > TLS/SSL, HTTP, or even something that's broken within an application. > Understanding that there are layers (or pre-requisites) enables one to devise > tests or gather evidence (e.g. packet captures) to determine at what point > the problem is occurring, then postulate a theory on what's broken. Without > that understanding, all you can go is attempt random "fixes" or changes in an > attempt to solve the problem. > > I think it's quite clear from this conversation that you don't see any value > in this framework. Others do (I find OSI quite helpful, as well as > normalisation, and OOP, and many other frameworks). The net is that no one is > going to change anyone else's mind, so I think we should leave it at that. > > Devil's advocate: :-) > As for Newtonian laws being 99% accurate - it's still a flawed model. So, > where exactly is your boundary? Models that are fundamentally wrong, but make > good approximations for human interaction are OK, but models that don't have > real-world implementation (but have no inherent flaws) are somehow "not > scientific" and "flawed"? Interesting argument :-) > > Cheers > Ken > > -----Original Message----- > From: Kurt Buff [mailto:[email protected]] > Sent: Monday, 31 May 2010 4:23 PM > To: NT System Admin Issues > Subject: Re: Applicability of the OSI model (was: Big Changes) > > Sigh. > > Newtonian physics works to several 9's of accuracy, and is good enough for > almost everything that humans encounter. That's a whole different beast than > the OSI stack, where, unless I'm thoroughly confused, the only thing that's > even close to widely used that somewhat follows that model is X.400. > > Tell me - when was the last time in your memory where you thought something > like "Oh, this is operating at layer 5 instead of layer 6 or layer 4"? > > Kurt > > On Sun, May 30, 2010 at 23:14, Ken Schaefer <[email protected]> wrote: >> Well, if that's your only argument, then no one should be using Newtonian >> physics, because it's "not scientific", it's a flawed model, and it fools us >> into thinking something that isn't so... But last time I checked, Newtown's >> laws of motion were good enough to put man on the moon. And it still gets >> taught in science classes all around the world. >> >> Cheers >> Ken >> >> -----Original Message----- >> From: Kurt Buff [mailto:[email protected]] >> Sent: Monday, 31 May 2010 1:46 PM >> To: NT System Admin Issues >> Subject: Re: Applicability of the OSI model (was: Big Changes) >> >> On Sun, May 30, 2010 at 13:06, Ben Scott <[email protected]> wrote: >>> On Sat, May 29, 2010 at 4:05 PM, Kurt Buff <[email protected]> wrote: >>>> The 4-layer TCP/IP model makes a better model, because it's closer >>>> to what's actually used. >>> >>> Way to not respond to anything I or anyone else wrote. >>> >>> It appears your entire argument is "OSI isn't TCP/IP". I could >>> reiterate my entire post, and also touch on what some other people >>> wrote, but if seems like you're not listening and won't respond, so I >>> don't see the point. >> >> Well then, to expand upon my thoughts: >> >> Using a model that bears little relationship to reality is a faux pas, and >> likely to lead you to bad conclusions. There are also dangers involved in >> adding layers to a conceptual model of networking, as described in RFC3439 >> (https://tools.ietf.org/html/rfc3439). >> >> It ain't scientific. >> >> I believe it's better to acknowledge that everything above layer 3 is a >> bunch of different protocols, some of which stand alone and some of which >> are encapsulated in other protocols, than to use a flawed model and fool >> ourselves into thinking something that isn't so. >> >> Kurt >> >> >> ~ Finally, powerful endpoint security that ISN'T a resource hog! ~ ~ >> <http://www.sunbeltsoftware.com/Business/VIPRE-Enterprise/> ~ > > ~ Finally, powerful endpoint security that ISN'T a resource hog! ~ ~ > <http://www.sunbeltsoftware.com/Business/VIPRE-Enterprise/> ~ > > > ~ Finally, powerful endpoint security that ISN'T a resource hog! ~ > ~ <http://www.sunbeltsoftware.com/Business/VIPRE-Enterprise/> ~ ~ Finally, powerful endpoint security that ISN'T a resource hog! ~ ~ <http://www.sunbeltsoftware.com/Business/VIPRE-Enterprise/> ~
