Need insight for my Networks presentation

[Sidewinder]

Hey forummy-baby!!!!

I am going to give a presentation to our division about the basics of How A Network Works. It's meant to be a plain, enjoyable overview of basic networking technologi for the interested layman; it's not a real technical course.

I'm excited to put together some visual aids and I find myself in a quandary which I hope ye designerfolk can help me with. Sorry to do this, but to explain the quandary I have to give a bit of background on the technologies I'm talking about - I'll try to keep it as simple as possible:

It used to be that each computer maker had their own way of networking; IBM had SNA, DEC had DECnet, Apple had (nay, has!) AppleTalk. Those protocols were unique in every way, from the high-level design right on down to the type of wires and connectors. As interoperability between different vendors' equipment became important, some things had to be standardized. One of the things that got standardized was what we're going to call Layer 2 technology (Layer 1 technology (voltage and wire-gauges and stuff) is important too but most L1 specs are included or assumed based on the L2 spec). Layer 2 specifies the way the software located closest to the physical circuit handles the data coming in and out (as opposed to higher layers; which specify data-handling by the software located closer to the users). Layer 2 technologies include Ethernet, token-ring, FDDI, Frame Relay, and others.

Vendors who used to require the use of their own layer 2 or layer-2-like systems began to support the use of their proprietary protocols over the standardized L2 implementations. Where DECnet traffic would once only use DEC's own "LAT" networks, it was now permitted (and accommodated) on standard Ethernet, token-ring, FDDI, Frame Relay, etc. (DEC was actually one of the inventors of Ethernet, and of Framey-baby as well). AppleTalk, which used to travel only over Apple's own LocalTalk cabling scheme, could now be seen on Ethernet, TR, FDDI, etc. These vendor protocols we are going to call Layer 3 protocols. The Internet's own IP is also a Layer 3 protocol.

A chunk of data intended for transmission, before it is transmitted, is chopped up (painlessly, I promise) into little pieces called packets. Each packet is inserted into a slightly bigger piece of data called a frame; the frame is the series of bits that actually gets transmitted on the physical media. One Ethernet frame can contain a DECnet packet, the next one can contain an IP packet, the next one DEcnet, then a frame with an AppleTalk packet. Widely differing Layer 3 protocols can coexist on the same Layer 2 gear and even utilize it simultaneously (yes, only one frame allowed on the wire at a time, but at 100 megabits per second, that's simultaneous enough!!!)

While multiple Layer 3 protocols are allowed over one Layer 2 setup; only one Layer 2 protocol is allowed per set-up!! That is, you can have IP and AppleTalk running on one Ethernet hub; but you cannot have token-ring and Ethernet running on the same hub!!

Here endeth the lesson. Now beginneth the question:

[Sidewinder]

I want to do a visual in which the L2 systems are represented by railroad flatcars. The L3 payload will be containerized freight; one container in each car (since only one L3 packet is allowed in a frame).


Do I want the flatcars representing the different media to be different colors; and the containers to be different shapes /sizes (Figure 1) ?

Or do I want the L3 containers to be different colors, and the railcars can be different lengths - or even different shapes; i.e. Ethernet could be flatcars, while Frame Relay could be trapezoidal coal-cars, token-ring could be rounded, kinda like tank cars (Figure 2) ?

NOTE: The attached visuals do NOT belong in The Crit Pit because they are only meant to help me ax my question. I consider them nowhere NEAR finished, and depending on the answers I get here, they may not even be started.

I wonder if there's a grand unifying design principle to answer a question like this. The lower-level, infrastructure-type stuff has more mass and less variety (the Ethernet segment contains only Ethernet, and more of Ethernet than of L3 stuff); the upper layer stuff makes less of an appearance (packets are smaller than frames) but each appearance is potentially much more varied than the infrastructure. In the instances in the design world where that's generally so (less variety below, less quantity/mass above), does color-coding generally belong at one level and not at another?

[Sidewinder]

So we can't upload .tif files, or what?

.bmp, .jpg, .gif, and .png versions of these files all exceed the 19kb limit...

[garricks]

You can use an image-hosting site like flikr, photobucket, picasa, etc., then link to the photos by using the img tags instead of html tags.

[ IMG]http://www.imagehosting.com/123/image.jpg[/IMG ]

but without the spaces inside the brackets.

[flutterby nut]

i think shape carries greater impact and weight than color in your analogy...

boxcars and flat cars are the more common bits of railroading, followed by more specialized things like coal cars, oil cars, and pickle cars, followed by the rare things like say, nuclear, or military, or some really bizarre manufacturing equipment...

choose your cars based on specificity of function, then color accordingly, perhaps realistically to aid greater recognition from your audience...you don't want them to stop and think what a car might be--make it obvious...

[Sidewinder]

Figure 1:

http://farm4.static.flickr.com/3594/3408316180_0e77f92c87_m.jpg

Y Figure 2:
http://farm4.static.flickr.com/3308/3407507383_b93780090b_m.jpg