Article View: comp.protocols.ppp
Article #36Re: SPARC 1+ performance vs PPP
From: Bob Sutterfield
Date: Thu, 16 Apr 1992 16:21
Date: Thu, 16 Apr 1992 16:21
68 lines
3561 bytes
3561 bytes
In article <RICH.92Apr15205332@mallet.med.ge.com> rich@med.ge.com (Richard B. Ralston) writes:
We just got a "demo" copy of Morning Star PPP to try out....
(Morning Star is pretty expensive, but the documentation is almost
worth it!...)
Besides the documentation, one of the things you pay for when you buy
our software is access to our customer support group, which one might
hope would give higher quality, quicker, and more specifically
knowledgeable response than asking on the net.
Demo/Evaluation/Pre-Sale users have the same privilege. Use it!
However, while I have the PPP connection up, performance on my
SPARC really gets bad!
We need more specifics...
The documentation was nice enough to say that Telebit in PEP mode
can be slow and erratic (would 2400 be better? I'd be surprised if
it was.),
It might be, depending upon your application. You might see better
responsiveness at 2400 (especially with MNP4/5 enabled, and the DTE
still at 19200), but FTP throughput will probably be better with PEP.
We only recommend running PPP or SLIP over PEP in those situations
when PEP is the only carrier that will hold the line, such as between
La Paz and Manila. In most of the western world, use V.32 or V.32bis.
but I didn't expect my SPARC to stop responding while using PPP.
This is not normal behavior. On most workstations, you can't even
tell that the PPP daemon is running. Please give us a call at +1 800
558 7827 or write to support@morningstar.com and we'll help you
diagnose your installation.
1) will things go better with the SPARC 2?
No. Since a pppd should normally consume <1% of your CPU when running
full-tilt, and since the system call and interrupt loads are
negligible when transferring traffic at normal async speeds (38400 or
below), the additional CPU juice available in the SPARCstation 2 won't
solve your problem.
2) would the "Snaplink" improve things?
Our SnapLink SCSI-attached serial interface is for running synchronous
PPP, typically over a 64K or T1 leased line to a router. It's not
what you need, since it doesn't address your problem.
3) Is there something I can change to help my SPARC performance and
either not hurt or improve PPP performance? (Other than
pitching the Telebits which we're stuck with.)
Your problem isn't related to your use of Telebit PEP modems, it's in
your installation. Yes, there is almost certainly something you can
change, but I can't tell what just yet.
4) Would running SLIP mode help? (Even though it's "non-standard"
our dial up site will run whatever we do.)
Your problem almost certainly isn't related to the encapsulation
protocol you're running over the wire, so changing to SLIP shouldn't
change anything. Once your problem is solved, you won't be able to
tell any difference in performance between SLIP and PPP, since our
daemon does "VJ" TCP header compression in both.
5) other suggestions?
Call us!
Technical Details
Message-ID:
<BOB.92Apr16122128@volitans.MorningStar.Com>Path:
arc1-ca.pugleaf.net!sparky!uunet!zaphod.mps.ohio-state.edu!mstar!mstar!bobReferences:
<RICH.92Apr15205332@mallet.med.ge.com>Show Full Headers
Complete Article Headers
Newsgroups: comp.protocols.ppp Path: arc1-ca.pugleaf.net!sparky!uunet!zaphod.mps.ohio-state.edu!mstar!mstar!bob From: bob@MorningStar.Com (Bob Sutterfield) Subject: Re: SPARC 1+ performance vs PPP In-Reply-To: rich@med.ge.com's message of 15 Apr 92 20: 53:32 Message-ID: <BOB.92Apr16122128@volitans.MorningStar.Com> Sender: news@MorningStar.Com Nntp-Posting-Host: volitans.morningstar.com Organization: Morning Star Technologies References: <RICH.92Apr15205332@mallet.med.ge.com> Date: Thu, 16 Apr 1992 16:21:35 GMT Lines: 68