ECT265: Lab 10.6.6 Input & Revisement

Sent: Wednesday, May 07, 2003 3:23 AM

Subject: ECT265-Dirk Fretz: Lab 10.6.6

Hi Professor Gulledge,

Let me start this off first by quoting an excerpt from "Feedback Information" on page iii of the CNAP: Lab Companion Vol I, 2nd Ed.

"At Cisco Press, our goal is to create in-depth technical books of the highest quality and value. Each book is crafted with care and precision, undergoing rigorous development that involves the unique expertise of members from the professional technical community."

All I can say to that is, "BULL!". I don't know what the deal is, but it seems as if the author of this lab manual wrote up 10.6.6, ran it by another networking knowledgeable person, and that person said, "Yeah, I understand what you're trying to say.". There are so many vagaries and a complete lack of visual examples as to make the concepts of subnetting almost completely unclear. I can't believe that subnetting is as difficult as the author has made it by not providing some "fundamental" examples/visual cues. After turning to page iv, I see that the author is a local person that teaches at Chandler-Gilbert CC. Maybe Jim Lorenz is better in person than he is in print or maybe I'm just dense. Although, Chapter 10 seemed pretty clear on the subnetting procedure.

On and off, based on life and academic needs of the moment, I've been reading, re-reading, and re-working Lab 10.6.6. I refuse to waste time on rewording whole sections of the lab, but I have made the effort to try and shore up what I perceive to be some serious short-comings. Granted, I have added wordage, underlined, bolded, and tried to make key points jump out at the reader. I've also taken this one step further by referring to Chapter 10, specifically page 408, in order to help clarify some things that are just too damn vague. I know, I know, you're probably saying to yourself, "That's why you have me!". ;0)

You know me. I'm not perfect and I don't have a problem with saying, "Oops. I screwed up.". So, maybe this revamp isn't 100% accurate, but I think it's pretty close. Of course, I'd like to have you take a look at it, especially pages 86 & 89 (Word: 2 & 5). Maybe I'm far off the mark in terms of understanding just how to do subnetting. I haven't completed Lab 10.6.6 yet or 10.7.5 and 10.7.7 simply because I've been working on improving 10.6.6 when I've had the chance. Anyway, let me address some of the lab sections. Of course, snag lab 10.6.6 to compare the two versions.

Page 85

This page is pretty much the same, but I made some of the key points more visible as to why and what is done during subnetting. It just doesn't make sense to have that information buried in a gigantic paragraph. We're trying to learn the material and having key concepts crammed into a paragraph like this doesn't help.

Page 86

Here's where I started some major modifications in the chart. Based on the first page, it's the same problem all over again. The author says a lot, but he doesn't express it in the visual sense to reinforce and help a person grasp the concepts.

Rearranged some of the columns in this chart.
Octet Bit Pattern: Made the column in order to show just what bits from the host are used, their binary weight, and binary class association.
Octet Decimal Range: Used a dividing bar in order to separate subnet bits from host bits. Visually, this helps me get an idea of what's going on in the process.
Number of Networks: Demystified just where the powers of two were coming from based on Fig. 10-14 page 391 and showed binary class association.
Hosts Per Network: Followed through with the same type of presentation style as with "Number of Networks".

Page 87

This is where some of the problems started and that I only caught onto tonight. Apparently, the author starts talking about default and custom subnet masks. Specifically, he uses Step 3 "Two Class C networks using the default subnet mask" as on example. It was pretty clear and I didn't have a problem understanding the idea based on that example. Where things got out of kilter was at Step 4 "One Class C network using a custom subnet mask".

At first, the problem was that I didn't really appreciate that these two steps dealt with similar, yet different types of subnets. Maybe I've misinterpreted something, but it almost seems like the reader needs to extrapolate/interpret the need for a host to AND a custom subnet mask in Step 4 with it's IP address, like it does with a default subnet mask in Step 3, when the host is on a subnet. The problem is that....this just isn't really clear. We've got Steps 3 & 4, lots of blurbage about this and that, but not nearly enough distinction between the two steps as they apply to the final exercise. I'm not even sure where the "Step" part comes in with these examples since they seem to be mutually exclusive instances and not logically progressive "related" steps.

Granted, the words were in the headings: "default" and "custom". Still, they didn't register at first. I also felt like I was running into terminology continuity errors based on the way the author was explaining Steps 3 & 4 versus how Chapter 10 explained subnet masking. Maybe I went a little overboard with my underlining, quotation marking, and italicizing, to bring out "default" and "custom", but I really felt that these were things that need to jump out at the reader and say, "Hey, pay attention!!! I'm not the subnet you think I am! I'm a "default/custom" subnet.".

Added: Custom Subnet Bit Borrowing Rule: The least number bits you can borrow from a host octet(s) is 2. The least number of bits you can leave in the host octet(s) is 2.

Page 89

This was another place where I tried to make things more clear with better detailed chart information. I took a lot more liberty on this page in terms of plastering it with decimal and binary representations based on what the author mentioned. To me, it really doesn't help to just "say" some octet is full of 0's or 1's for some reason. To drive the point home, I need to see visual examples/representations. Page 408 really came in handy for the chart on the bottom of the page.

Overall, I just don't think there was enough supportive and clear information in this section of the lab. That's the opinion of a couple other students as well. Again, maybe I've goofed up on something, misinterpreted some concepts, or am close in the "Maxwell Smart" sense. If you've got time (yeah, right) look over this revamp of 10.6.6 and an original copy to compare the two. Then, get back to me by e-mail or in class to show me where I might have made some mistakes or any other feedback. My hope is that this is a better version of 10.6.6 that you could use for future classes.

As far as I'm concerned, copyrights be damned. If you're Cisco Systems, aka a juggernaut in the networking business, you should be policing and putting together better quality lab manual material that jives with the CNA book. You're a world leader. Now act like it!

Hopefully the other labs will be better.

Thanks for your time,

Dirk Fretz

From: Roger S. Gulledge

To: 'Dirk Fretz'

Sent: Wednesday, May 07, 2003 8:42 AM

Subject: RE: ECT265-Dirk Fretz: Lab 10.6.6

Hi:

May I forward this on to Cisco Press?

RSG
===========================================
Roger S. Gulledge
Associate Professor
of Electronics and Computer Technology
PC and Networking Curriculum Manager
Local Cisco Networking Academy Main Contact
DeVry University
2149 W. Dunlap
Phoenix, AZ 85021
Phone: 602-870-9222, x870
Fax: 602-734-1999
Email: rgulledge@phx.devry.edu

From: Dirk Fretz

To: rgulledge@phx.devry.edu

Sent: Wednesday, May 07, 2003 3:38 PM

Subject: Re: ECT265-Dirk Fretz: Lab 10.6.6

Hi Professor Gulledge,

I was thinking of doing the same thing (networkingacademy@ciscopress.com), but I wanted to get your input on this revamp in a proofreading sort of way to make sure I didn't screw up any concepts. If this came from you to Cisco Press, it would carry more weight. What I don't want to do is hastily submit something with errors that I created and that you might not have caught in your desire to give Cisco Press immediate feedback. Then again, maybe you meant to submit it after proofing it. Assumptions, assumptions.

Case in point, as I looked over the document this morning, with caffeine and allergy medicine coursing through my system, I noticed where I think I made a conceptual mistake on page 89. 4th paragraph, second to last sentence, I wrote this:

When a host ANDs the “192” net with the custom subnet mask to find out what network it’s on, the result is 200.1.1.255, which is equal to the Class C network’s broadcast address.

Well, that's not true. I wrote that in the wee hours of the morning and made the accompanying Net 0 custom subnet and custom subnet mask address comparison chart, but I forgot to change that last sentence to correspond with what I finally implemented. If I followed my own directions in that above sentence, this would be the result:

Assigned Class C Network Address 200.1.1.0

The “192” Subnet Address 200.1.1.192 = 11001000.00000001.00000001.11000000

Host ANDs with “Custom” Subnet Mask 255.255.255.192 = 11111111.11111111.11111111.11000000

ANDing Result = Network Address (200.1.1.192) = 11001000.00000001.00000001.11000000

Hey, what do you know? That's totally incorrect and misses the point of the Net 0 custom subnet and assigned network broadcast address comparison chart. More importantly, based on Step 3, the host doesn't AND subnet addresses with custom subnet masks. It ANDs its own IP address with the default or custom subnet mask. What that 4th paragraph, last and second to last sentence should have said was:

When the last subnet address on the “192” net is reached (200.1.1.255), it equals the Class C network’s broadcast address (200.1.1.255). In short, you can’t violate an assigned network number’s network address or broadcast address in this way.

Those two sentences correspond with this comparison chart.

Class C Network Broadcast Address	The “192” Net Broadcast Address
200.1.1.255	200.1.1.255

Now, while typing this response, here's something else I've found in that paragraph that I missed. 4th paragraph, second sentences says:

This is because the “0” net has all binary zeros in the custom subnet portion of the first address that starts at the range number “0” (200.1.1.0 = 11001000.00000001.00000001.00000000).

The 4th sentence is where I'm trying to clarify the point of the second sentence, but I screwed up:

When a host ANDs the “0” net (200.1.1.0) with the custom subnet mask (255.255.255.192) to find out what network it’s on, the result is 200.1.1.0, which equals the assigned Class C network address.

First address? Which "first" address Dirk?! The "first" SN address for the "0" net or the "first" host address for the "0" net?! What does your example show?! This is the drawback to trying to extrapolate Step 3 info into Step 4 material.

Assigned Class C Network Address	200.1.1.0
The “0” Net First Address = 200.1.1.0	11001000.00000001.00000001.00000000
Host ANDs with “Custom” Subnet Mask 255.255.255.192	11111111.11111111.11111111.11000000
ANDing Result = Network Address (200.1.1.0)	11001000.00000001.00000001.00000000

Well, that's ambiguous. You've got the "0" Net with a "First Address", but you don't clarify it as the "First Host Address". What does the host AND? It ANDs its IP address and the default or custom subnet mask. So, that chart just needs to be more clear by saying the "First Address" is a "Host" address.

Assigned Class C Network Address	200.1.1.0
The “0” Net First Host Address = 200.1.1.0	11001000.00000001.00000001.00000000
Host ANDs with “Custom” Subnet Mask 255.255.255.192	11111111.11111111.11111111.11000000
ANDing Result = Network Address (200.1.1.0)	11001000.00000001.00000001.00000000

Ok, even that chart is incorrect as it applies to custom subnet host IP addresses being ANDed with a custom subnet mask address. I'm getting the impression that there's a two pronged problem. While re-re-re-reading that paragraph, I think I'm beginning to see it from the two vantage points that have been mentioned already.

In the case of creating "custom" subnets, based on the amount of borrowed bits, you can only have 2^n - 2 custom subnets. Personally, I prefer to just call them "subnets". Referring to them as "custom" subnets is like saying I bought a "canine" dog; i.e. the act of subnetting in and of itself is customizing an assigned network address into sub-networks aka subnets without needing to add on the "custom" distinction.

Based on 10.6.6, the reasons for subtracting 2 from the possible number of custom subnets created are:

Eliminating the first custom subnet, Net 0, removes the conflict of it's first custom subnet address and the assigned network address being the same.
Eliminating the last custom subnet, Net 192, removes the conflict of it's last custom subnet address, which happens to be Net 192's broadcast address, and the assigned network broadcast address from being the same.
Host ANDing is irrelevant in both of these instances.

The number of hosts per custom subnet you can have is 2^n - 2. Based on 10.6.6, the reasons for subtracting 2 from the possible number of host address per custom subnet are:

Eliminating the first custom subnet host address, per subnet, removes the conflict of the first host address and that host's custom subnet network address from being the same.

Eliminating the last custom subnet host address, per subnet, which happens to be that subnet's broadcast address, removes the conflict of that subnet's broadcast address and the assigned network broadcast address from being the same.

Host ANDing is relevant at this point.

In a nutshell, there appears to be nested conditions for subnets and subnet host addresses as they relate to assigned network addresses.

Assigned Network Addresses
- Subnets Rules
  - Subnet Host Address Rules

Step 3's ANDing examples influenced the example I put on page 89 for Net 0. Still, based on these little epiphanies I've had during this reply, my example is wrong and irrelevant. You'd never have a Net 0 first custom subnet host address of 200.1.1.0 to AND with a custom subnet mask of 255.255.255.192. This is because the Net 0 first custom subnet host address of 200.1.1.0 is discarded due to conflicting with the Net 0 custom subnet network address of 200.1.1.0.

Assigned Class C Network Address	200.1.1.0
The “0” Net First Host Address = 200.1.1.0	11001000.00000001.00000001.00000000
Host ANDs with “Custom” Subnet Mask 255.255.255.192	11111111.11111111.11111111.11000000
ANDing Result = Network Address (200.1.1.0)	11001000.00000001.00000001.00000000

This is why I wanted you to go over this stuff before you submitted it to Cisco Press. I "think" the assessments that I've made in this reply are right, but again, I'd need your insight to tell me whether I'm right or wrong. If I'm right, it would mean doing a little rework on page 89 to correct the mistakes I've made in the explanations. That's not that big of a deal, but I've got other things to do too and I don't see a royalty check come out of this effort anytime soon.

Well, let me know what you think when you get a chance to peruse this e-mail. I'd like to understand this process and possibly knock out something like a quick-reference guide for other students based on this revamp. Like I said, I can't believe subnetting is that difficult, but I can see where you can get lost in the details.

Thanks for your time,

Dirk Fretz