Jeff Kell writes:

> Wirt Atmar wrote:
>
>  > Outside of the HP3000 world, telnet has become the universally standard
>  > mechanism for terminal-like devices to communicate with their hosts --
and
> as
>  > more and more diverse clients are connected to the HP3000, telnet was
> bound
>  > to rise in popularity very quickly, to the point that even though the
> ratio
>  > was 10,000:1 three years ago, it seemed reasonable at the time to presume
>  > that NS/VT's days were numbered.
>
>  If QCTerm and CSY want to keep with the times, you should look beyond
>  classic telnet into ssh, unless VPN connectivity surpasses it.

Trying to predict what is and what will not be a standard five to ten years
out is always difficult. In that regard, I have substantial doubts about SSH.
It's just too complicated to become an easy standard. Regardless of what the
most enthusiastic advocates of any technology may say, simplicity remains key
to widespread acceptance, especially when other essentially comparable
methodologies already exist.

While I had to read Jeff's sentence twice, I ultimately came to agree with
it. If I were to guess, hard security for host-terminal connections is most
likely going to be accomplished by putting some sort of wrapper around the
entire conversation, essentially converting every link into a VPN. That's the
really easy way to upgrade security. It leaves all existing communications
protocols (HTTP, FTP, telnet, etc.) untouched, completely compatible with all
"legacy" uses, without having anyone on either end, host or terminal, being
required to change anything.

Deducing methods like the introduction of NTSC color television that do not
require the obsolescence of the previous standards is always, by far, the
most desired goal.

Nonetheless, in the interim, all of the security mechanisms inherent to HTTP
are being also brought over to telnet and FTP too by the IETF. It's just that
they're not implemented by anyone in any practicable numbers yet. I would
tend to believe that this will be the next step in the evolution of  secure
telnet (or telnet-like) communications -- if some easily added VPN construct
isn't designed first.

Wirt Atmar