----- Original Message ----- >
>
> >TCP/IP vs. ISO OSI
> <random anecdote>
> This is true, but I think the market example is
> TCP/IP vs SNA/APPN
TCP/IP vs SNA/APPN/HPR
Yep. I was a SNA bigot once. SNA was optimized to squeeze the most out of
available bandwidth and to handle HUGH numbers of concurrent clients
(100,000's back in the early 80's) but those optimizations did not fit well
with usage patterns on the web (lots of short lived connections). There is
some cool technology in HPR, ARB Adaptive Rate Based flow control, for
isochronous byte transfer and avoiding rather than reacting to network
congestion. Haven't touched it in 7 years or so now :-)
>
> OSI never really got any real deployment, is was just
> a proposal when both were nascent. SNA was well
> deployed prior to TCP/IP and APPN was hailed
> (at least by the SNA bigots I knew) to be the
> thing that fended off the TCP/IP push for good.
> (in business networks)
> Ultimately, while SNA still exists for legacy
> systems, very few new SNA deployments are happening,
> instead being TCP/IP ones. The part I find most
> amusing is that the most common method of expanding
> SNA networks is via IP routers which do DLSw.
I was an advocate of APPN/HPR backbones with TCP/IP on the periphery. When the
market went the other way, I finally came around that TCP/IP was the way to
go. And I have MUCH more respect for open standards these days too :-)
Bill
This archive was generated by hypermail 2b29 : Fri Apr 27 2001 - 23:18:47 PDT