Retro networking / WAN communities

[Paul Koning]

> On Apr 12, 2022, at 12:42 AM, Grant Taylor <cctalk at gtaylor.tnetconsulting.net> wrote:
> 
> On 4/11/22 6:16 PM, Paul Koning wrote:
>> I think "hub" is another word for "repeater" (just like "switch" is another word for "bridge").
> 
> Interesting.
> 
> Do you know of any documentation, preferably not marketing materials, that used "repeater" in lieu of "hub"?

DEC documentation.

> From my naive point of view, hubs came about when multiple stations connected to a central location, the center, or hub, of the start if you will.  Conversely, I remember reading (after the fact) about repeaters as something that existed in pure 10Base5 / 10Base2 networks, predating hubs.
> 
> I'm questioning form a place of ignorance.  Like a child asking why fire is hot.

The concept of a repeater goes back to day 1 of Ethernet; you'll find them in the D/I/X Ethernet spec.  And they were part of the first batch of Ethernet products from DEC.  Yes, AUI based devices, two port.

But the next thing out the door was the DEMPR, "Digital Multi-Port Repeater", an 8 port repeater.  I think that's 10Base2.

I first saw "structured wiring" -- the star wiring with a hierarchy of wiring closets and devices -- around 1986, in the new Littleton King Street DEC building.  It had distribution cabinets at the end of each row of cubicles.  These looked just like standard office supplies storage cabinets, with shelves; inside you'd find a bridge and a couple of DEMPR repeaters, connected to 10Base2 coax drops to each cubicle.

> I think there is a large, > 80%, overlap between switch and bridge, but they aren't perfect.  Bridging some traffic between otherwise incompatible networks comes to mind; e.g. SNAP between Token Ring and Ethernet or Ethernet to xDSL (RFC 1483).

That's not where the term "switch" was introduced.   And devices like that were called "bridge" by market leaders like DEC -- the two generations of FDDI to Ethernet bridges I mentioned were both called "bridge".  

Also, the general operation of the device is the same whether it does MAC frame tweaking or not, 802.1d applies unchanged.  Ethernet to non-Ethernet bridges have to do some tinkering with Ethernet protocol type frames (which is where SNAP comes in, all nicely standardized in the FDDI days).  For 802.5 they also have to deal with the misnamed "functional" addresses, but that's not hard.

There also was such a thing as a "source routing bridge", an 802.5 only bad idea invented by IBM and sold for a while until the whole idea faded away.

>> The device I have is a small standalone box, about the size of today's small 4-6 port switches you buy at Staples for $100.  But it's actually a repeater, not a switch, and one of its ports is a 10Base2 connector (BNC jack).
> 
> I would firmly consider what you describe as a "hub".

I think "hub" is what DEC called the chassis that these boxes could plug in to. 

>> ...
>> That's rather odd because even if someone doesn't obey the letter of the law you'd think they would at least support 100BaseT. Or was the problem lack of half duplex?  Do those management interfaces want to run half duplex?
> 
> No.  It's more nefarious than that.  You mentioned supporting n - 1 generation.  I'm talking about switches that support 1 Gbps / 10 Gbps / 25 Gbps / 40 Gbps / 50 Gbps / 100 Gbps.  They quite simply don't scale down to 100 Mbps much less 10 Mbps.  --  Why would someone want to support those slow speed connections on such a high speed switch? Devices like intelligent power strips or serial consoles or the likes in a cabinet that uses said switch as a Top of Rack device.  --  Our reluctant solution has been to put in a lower end / un-manged 10 Mbps / 100 Mbps / 1 Gbps that can link at 1 Gbps to the main ToR.

I understand now.  Yes, that's annoying indeed.

>> I think I saw in the standard that Gigabit Ethernet in theory includes a half duplex mode, but I have never seen anyone use it and I wonder if it would work if tried.  Perhaps I misread things.
> 
> My understanding is that Gigabit Ethernet (and beyond) only supports full duplex.  Maybe I'm mis-remembering or thinking about what is actually produced vs theoretical / lab experiments.

I took a quick look in the 802.3 spec.  In the 2002 edition, Part 3 describes gigabit Ethernet.  The intro ("clause 34") has this to say:

"In full duplex mode, the mini- mum packet transmission time has been reduced by a factor of ten. Achievable topologies for 1000 Mb/s full duplex operation are comparable to those found in 100BASE-T full duplex mode. In half duplex mode, the minimum packet transmission time has been reduced, but not by a factor of ten."

So yes, it's theoretically part of the spec.  As you said, it doesn't seem to be in actual use.

> Similarly, I know someone that has 100 Mbps Token Ring, a.k.a. High Speed Token Ring (HSTR) equipment for their mainframe.  And 1 Gigabit Token Ring was designed in the lab but never actualized.

Curious.  Clearly such things are possible.  But FDDI came out well before HSTR, and it was crushed by 100 Mb Ethernet.  All the reasons for that to happen would apply much more so for HSTR.

Does anyone still remember the other 100 Mb Ethernet-like proposal, I think from HP, which added various types of complexity instead of simply being a faster Ethernet?  I forgot what it was called, or what other things it added.  Something about isochronous mode, perhaps?  Or maybe I'm confused with FDDI 2 -- another concept that never got anywhere, being much more complicated even than regular FDDI.

	paul