"I have seen the future, and it's 1975!"
I'm very much in favor of AICS Research, Advanced Telnet,
although the main technique that makes it work so well,
is that it works in good ol' half-duplex. Which is not
new at all, but hearkens back to Teletype methods, when
every character ran at 66, 75, or 110 BPS. Running
full-duplex cut throughput in half, because every
character was echoed back.
Now that large amounts of data (and graphical data) is
also returned to the user in the form of internet packets.
IMO half-duplex removes half the burden of those packets for
the textual portions, hence some of the reason for the
"apparent" improvement in speed for AICS Research QCTerm.
> -----Original Message-----
> From: Neil Harvey [mailto:[log in to unmask]]
> Sent: Monday, January 03, 2000 2:36 AM
> To: [log in to unmask]
> Subject: Re: NS/VT vs TELNET
>
>
> Don't forget AICS Research's "Advanced Telnet" protocol,
> which is somewhere
> between telnet (each char echoed by host) and block mode
> (chars sent when
> you press enter), I think.
>
> Over sloooww internet connections it works astonishingly well.
>
> However, it is in development, so may have to wait.
>
> Regards
>
> Neil
>
> -----Original Message-----
> From: Gilles Schipper [mailto:[log in to unmask]]
> Sent: 02 January 2000 20:52
> To: [log in to unmask]
> Subject: Re: NS/VT vs TELNET
>
>
> There are 3 ways for a session to connect to the HP3000 via a network
> connection:
>
> 1. hardware telnet 9via DTC telnet card or telnet access box).
>
> I would think this method would be the most efficient, since
> most of the
> character processng overhead is done by the hardware and
> frres up the host
> cpu for more user-oriented processing tasks.
>
> The main disadvantage is the restriction on the max. number
> of users per
> card (20 or 40 or 80 - dpending upon which model dtc and/or
> telnet access
> box).
>
> Also, of course, you require HP Openview DTC manager software
> on a PC to
> configure and manage hardware telnet - which raises a cost
> issue, as well
> as an operator management issue.
>
>
> 2. host software, via NS/VT
>
> Probably the second choice in terms of overall efficiency -
> since NS/VT
> host processing is more efficient that telnet host processing
> (see choice
> number 3).
>
> From an operational management viewpoint, this option is a no-brainer,
> since it requires only the initial configuration of the host network
> configuration and it's plug and play after that.
>
> The disadvantage is that you require a non-standard layer of
> software for
> the client - available with various 3rd-party terminal
> emulator packages
> such as WRQ Reflection for HP with NS/VT, and MS92 from Minisoft.
>
> 3. host software, via software telnet
>
> Probably the least overall efficient, but the most standard
> and available
> from a client perspective.
>
> The care and feeding of a background job (jinetd) is the
> major operational
> consideration. Nothing significant to worry about here,
> although tweaking
> that job is important to ensure minimal glitches in its operation.
>
> I should say here that I believe that software telnet is
> improving each new
> iteration - although I still think it is still more
> problematic than ns/vt
> or hardware telnet. Others more knowledgable in that area will surely
> correct me if I'm wrong.
>
>
>
>
> At 09:04 AM 2000-01-02 -0800, Paul H Christidis wrote:
> >In recent weeks some of our 'out of state' users have been
> configuring
> their
> >Reflection 7.X to use VT-MGR instead of TELNET (We have 2
> DTCs with telnet
> >access cards).
> >
> >What are the pros and cons of each of the above protocols?
> Are there any
> points
> >of concern from the system operation prospective?
> >
> >Regards
> >Paul H. Christidis
> >
> >
> --------------------------------------------------------------
> -------------
> Gilles Schipper
> GSA Inc.
> HP3000 & HP9000 System Administration Specialists
> 300 John Street, Box 87651 Thornhill, ON Canada L3T 7R4
> Voice: 905.889.3000 Fax: 905.889.3001
> Internet: [log in to unmask]
> --------------------------------------------------------------
> -------------
>
>
|
