[comp.protocols.appletalk] Gatorbox vs. Fastpath

RUBIN@graf.poly.edu (David Rubin) (11/30/89)

A faculty member is interested in getting a device to convert Appletalk
to TCP/IP in order to provide file sharing with UNIX machines (like
NFS).  He has mentioned two devices: Gatorbox and Fastpath.  What
are the differences between the two?  Which is better?  Will these
devices work with EtherTalk as opposed to LocalTalk (we plan to have
the Macs on the same Ethernet with UNIX machines using an Asante
Ethernet card for the Macs)?  Are there any software solutions that do
not require such a device?

Any info would be appreciated.  Please respond via E-mail...

-------
David Rubin                        |     INTERNET: RUBIN@graf.poly.edu
Polytechnic University             |       BITNET: RUBIN@POLYGRAF
Brooklyn, NY                       |

adams@swbatl.sbc.com (745) (06/06/90)

Due to some internal discussion and the preponderance of Gatorboxs around
here I'm wondering if anyone has interesting comparisons of Gatorboxes and
the Fastpath?  The consensus here seems to be that the Gatorbox is horrendously
slow compared to similar boxes, but offers the only Appleshare to NFS
gateway.  True?; False?; Totlally confused? War stories? Pointers?
-- 
uunet!swbatl!adams or adams@swbatl.sbc.com     
Tom Adams: 314-235-7459: Southwestern Bell Telephone Advanced Technology Lab
BOOKS WANTED: pre-1930 radio, electrical & scientific topics

anders@penguin (Anders Wallgren) (06/06/90)

In article <1544@swbatl.sbc.com>, adams@swbatl (745) writes:
>Due to some internal discussion and the preponderance of Gatorboxs around
>here I'm wondering if anyone has interesting comparisons of Gatorboxes and
>the Fastpath?  The consensus here seems to be that the Gatorbox is horrendously
>slow compared to similar boxes, but offers the only Appleshare to NFS
>gateway.  True?; False?; Totlally confused? War stories? Pointers?


I would add a couple of comments to this:

 - The GatorBox is configured with software downloaded from a Mac or a
   Unix TFTP host - I've never had to replace ROM's to upgrade it,
   which we've had to do with our FastPath's.  In general, configuring
   and upgrading the GB's is a joy compared to the FP's.

 - I've never run NFS across a FastPath (has anyone?), so I can't
   really tell you whether it's faster or slower than the GatorBox- it
   seems to me that comparing a FastPath w/Tops to a GatorBox w/out
   NFS but with Tops is the only way to measure relative performance,
   and I haven't seen anyone do that yet.  My GatorBox running NFS is
   fast enough for me, but I don't have any quantitative answers.  I
   wouldn't call it horrendously slow, thought.

 - Early on there were a lot of horror stories about the GatorShare
   software, mostly due to the fact that it was released a bit early
   (probably for revenue reasons...) but the software has matured a
   lot in the last year - I haven't had a single problem with it for a
   long time now - we're really satisfied and have stopped buying
   Fastpaths.

anders

nishri@gpu.utcs.utoronto.ca (Alex Nishri) (06/09/90)

In article <1544@swbatl.sbc.com> adams@swbatl.UUCP (Tom Adams - 235-745) writes:
>Due to some internal discussion and the preponderance of Gatorboxs around
>here I'm wondering if anyone has interesting comparisons of Gatorboxes and
>the Fastpath?  The consensus here seems to be that the Gatorbox is horrendously
>slow compared to similar boxes, but offers the only Appleshare to NFS
>gateway.  True?; False?; Totlally confused? War stories? Pointers?

You might want to check the June 1990 MacUser magazine.  Inside a special
section on networking, including a section on "EtherTalk to LocalTalk Routers",
is just the comparison you are talking about.  In particular, see the graph
on page 47 of the special networking section.

srb813@csc.anu.oz (06/14/90)

In article <1990Jun6.011927.26372@verity.com>, anders@penguin (Anders Wallgren) writes:
> In article <1544@swbatl.sbc.com>, adams@swbatl (745) writes:
>>Due to some internal discussion and the preponderance of Gatorboxs around
>>here I'm wondering if anyone has interesting comparisons of Gatorboxes and
>>the Fastpath?  The consensus here seems to be that the Gatorbox is horrendously
>>slow compared to similar boxes, but offers the only Appleshare to NFS
>>gateway.  True?; False?; Totlally confused? War stories? Pointers?

We use Webster Multigates at the ANU, these are similar to Fastpaths
except that they have _4_ localtalk ports instead of a pathetic 1 (ooh! that's
a bit strong! :-) ).  The 4 localtalks are bridged together as well as being
gated onto the ethernet.

We don't have any Gatorboxes at the ANU, but I know roughly what they do.
The Multigate (and Fastpath) is a much more general-purpose gateway than
the Gatorbox.  For example, DECnet traffic, all IP traffic (telnet/ftp, etc),
TOPS, etc, etc, can all be routed by the Multigate.  I'm don't know whether
the GB can do this (please enlighten me if it can!).

I imagine that since the GB is performing higher-level protocol translations
it may not be as fast as the other boxes.

We use CAP (running on a UNIX host) to then provide AppleShare access to
Macs/PCs.  The UNIX host's disks may be NFS mounted...hence the host is
doing the NFS <-> AppleShare translation for us.

> 
> 
> I would add a couple of comments to this:
> 
>  - The GatorBox is configured with software downloaded from a Mac or a
>    Unix TFTP host - I've never had to replace ROM's to upgrade it,
>    which we've had to do with our FastPath's.  In general, configuring
>    and upgrading the GB's is a joy compared to the FP's.

Multigates also boot from a UNIX bootserver via FTP (or a Mac).  They also
have their configuration in eeprom.  Upgrading the software is a breeze...
but determining the configuration proved to a real pain to set up.  Now that
it's going everything seems fine.

> 
>  - I've never run NFS across a FastPath (has anyone?), so I can't
>    really tell you whether it's faster or slower than the GatorBox- it
>    seems to me that comparing a FastPath w/Tops to a GatorBox w/out
>    NFS but with Tops is the only way to measure relative performance,
>    and I haven't seen anyone do that yet.  My GatorBox running NFS is
>    fast enough for me, but I don't have any quantitative answers.  I
>    wouldn't call it horrendously slow, thought.

I don't have any answers here either...

--- remaining stuff deleted ---

> 
> anders


One question that hasn't been answered is what about Macs directly connected
to the ethernet?  Who does the AppleShare <-> NFS translation for them?
With CAP v5.0 the Multigate actually does the EtherTalk <-> KIP translation,
but I've heard that a later release of CAP will/does talk EtherTalk directly.
Does the Gatorbox do EtherTalk AppleShare <-> NFS translation?

Steven Ball, Department of Computer Science, ANU
E-mail: steve@anucsd.anu.oz.au          Ph. (06) 2495147
Snail-mail: GPO Box 4, CANBERRA CITY ACT 2600, AUSTRALIA
           He's not the messiah, he's a very naughty boy!

swatt@NOC.NET.YALE.EDU (Alan S. Watt) (06/14/90)

I don't subscribe to this list; a piece in the above discussion was forwarded
to me.  If you have responses, please E-mail directly to me.

We have two GatorBoxen on evaluation order because of a feature in their
1.5 software which, as far as I know, nobody else can match.  Using this
feature it is possible to set up two GatorBoxen to act as IP routers
between two ethernets, across a localtalk/PhoneNet link.

Yale has over 200 buildings spread out over New Haven, all of them within
4000 cable feet of a campus telephone switchroom.  This is stretching it
a bit, but we have successfully run PhoneNet over that distance.  Most
buildings are well under the 3000 foot PhoneNet suggested maximum.

The 1.5 GatorShare software makes it possible to connect ethernet-based
IP networks to the campus network at medium speeds over installed
telephone wiring.  If it works as it should, it is a far better
price/performance solution than anything else I'm aware of.  The boxes
should arrive soon and I intend to try them out in a test setup.

Question: is anyone aware of software for the FastPath or other device
	  which provides this capability?

	- Alan S. Watt
	  High Speed Networking, Yale University
	  Computing and Information Systems
	  Box 2112 Yale Station
	  New Haven, CT  06520-2112
	  (203) 432-6600 X394
	  Watt-Alan@Yale.Edu

kre@cs.mu.oz.au (Robert Elz) (06/15/90)

In article <9006141345.AA05983@noc.net.yale.edu>,
swatt@NOC.NET.YALE.EDU (Alan S. Watt) writes:
| 
| If you have responses, please E-mail directly to me.

(doing that as well as this generic message).

| Using this
| feature it is possible to set up two GatorBoxen to act as IP routers
| between two ethernets, across a localtalk/PhoneNet link.
| ...
| Question: is anyone aware of software for the FastPath or other device
| 	  which provides this capability?
| 

Webster's Multigates do that as well.  They have 4 ports, so one box
would connect to 4 of your phone lines.  They support sub-sub-netting
which will result in less address space wastage on the (effectively)
point to point links (if you're hanging multiple gateways on the one
localtalk (phonenet) wire performance isn't going to be wonderful).

kre

disclaimer: I wrote the code for the multigate, so you might not
want to believe a thing I say...

wezel@bio.vu.nl (Jos van Wezel) (06/15/90)

In article <9006141345.AA05983@noc.net.yale.edu> swatt@NOC.NET.YALE.EDU (Alan S. Watt) writes:
>
>The 1.5 GatorShare software makes it possible to connect ethernet-based
>IP networks to the campus network at medium speeds over installed
>telephone wiring.  If it works as it should, it is a far better
>price/performance solution than anything else I'm aware of.  The boxes
>should arrive soon and I intend to try them out in a test setup.

No I don't think so. You can 'tunnel' localtalk through IP networks,
in 1.5 , not the other way around. You can setup 2 GB's far apart
on the Internet, and they will route your AT packets to each other.
Cayman calls these AT tunnels.

Correct me if I'm wrong.

Jos van Wezel
-------------