[e2e] e2e principle..where??....

John Day day at std.com
Mon Jun 4 21:08:06 PDT 2001


At 16:59 +0100 6/4/01, Panos GEVROS wrote:
>John Day typed :
>
>  |the question is where?  The primary argument at the time was: the
>  |PTTs arguing that you didn't need transport protocols that hop-by-hop
>  |was reliable.  While the networking crowd argued that hop-by-hop was
>  |not reliable and never could be and regardless no host was going to
>  |trust the network anyway, so if the hosts were going to do error
>  |recovery, then the network didn't have to do as much and could be
>  |simpler and cheaper.  The PTTs didn't and don't like this because it
>  |makes the network a commodity and a commodity business is hard work.
>  |(Notice how some large router vendors have learned this and talk
>  |about the importance of putting more intelligence in the network.
>
>"trust" is the key word i'm borrowing from your message -
>
>the conflict between elegant design vs. viable bussiness model above

I disagree.  If done right they are one and the same.  This "elegant 
design vs viable business model" argument is the refuge of the 
mediocre who don't know how to do both.

>has its roots (like many other things) in mutual distrust : the endpoints do
>not trust the network and the network does not trust the endpoints
>
>moreover the assumption that end-points cannot be trusted so that the network
>must be fortified (for resource management etc. which dominated research in
>the 90's) did not have the expected results,

That's because it fell right back into the telephony mindset, the 
beads on a string model, rather than the layered end to end model. 
If I had to guess it seems that well over half of the profs teaching 
networking, never really understood the network model and are mostly 
re-treaded data comm types that don't really believe it can be done. 
Rather than creatively assimilate what we had learned and create a 
new synthesis, we just got a tired re-hash of old stuff.

>
>i m not sure whether the end-to-end argument addresses issues of end-point
>"trust" and this is where  everything starts : the providers have to do things
>*inside* their networks

Trust in the sense you use it has nothing to do with the problem.

>
>IF the providers could *control* the endpoints  transport functionality
>included (or ensure *conformance*, or *enforce* certain behaviours to the
>endpoints) many problems would magically go away
>.. then they couldnt care less about how dumb their network is, and sure they
>wouldn't have any reason why to put more intelligence into it .. they would
>only have to provision (now easier) and their revenue would come from what
>they allow on the end-points
>
>that should be the only way to avoid the conflict above and have the best of
>both worlds : both the network is kept simple and the computation is pushed at
>the end-points.. and the providers do bussiness with it.

Nice idea.

>
>otherwise the market forces will enforce PTT-like model (because they
>understand it and they know how to use it do business) and i think it is going
>to happen fast.

It sure looks that way and why shouldn't they.  Even the people who 
you would expect to be doing the innovative next steps are pushing 
telephony-like solutions.

>
>so imho there is only one way for the pure internet, e2e, model to survive,
>the end-points (their behaviour,etc) must be trusted ,
>i'd put that on top of the research agenda.

Based on their behavior to date, it ain't going to happen.

>
>i'd love to hear from people that think the same or are aware of research
>efforts in this direction

Take care,
John




More information about the end2end-interest mailing list