(stupid?) Question assigned from college Cisco course

Status
Not open for further replies.
C

colinstu

2[H]4U
Joined
Oct 11, 2007
Messages
3,563
My university has a "Cisco" course and recently one of the questions presented was this. I don't intend for you to tell me the answer (of it that's "legal" in this forum) ... but I am frustrated by it.

The question reads like this:

Your supervisor has given you this class C address 192.168.50.0 and asked you to design an internetwork for their new client which is a small company that consists of the following five departments:

Engineering: 58 employees (58 nodes.)
Research & Development: 26 employees.
Sales & Marketing: 12 employees.
Human Resources: 8 employees.
Security: 4 employees.

There is a possibility of adding two more departments in the future. Thus, in your design you must account for future growth.
Click to expand...

We're supposed to use Cisco Packet Tracer to design the network.

My number one gripe about this question is the fact we have to use a Class C network for this! What is the deal with that?! As far as I know, any guy in this position would tell this "supervisor" to sit on it and use at least a Class B. What would really prevent them from not doing so? Anyways. (Yes, I did ask if we could use Class B, I was told we cannot).

As far as I know, you'd have to subnet the crap out of this thing. These "departments" must also be on different subnets... or not be able to communicate with each other at all anyways.

We learned a little bit about subnetting... and learned some router protocol stuff (RIP v2, EIGRP... that's it). Those are the tools I'm given and the only ones I'm allowed to use. No VLANS... or anything fancier than that.

In packet tracer we're typically using 1841 or 1941 routers (and can put modules into them).

Any suggestions on how to proceed with this ridiculous question?
 
Dude, get on Google and get you a subnet cheat sheet and break it down into 7 subnets.

Its really that simple.

Then use packet tracer to set up your network.

I could probably do it in my head.
 
Is it really that easy? I guess I was trying to overcomplicate it.

/27 subnet I take it
 
This is really the final project for the class. It being ... "the final" made me think it would be harder!

thanks for the help everyone.
 
I'm betting they're looking for VLSM.
In a "real" work environment, the subnets are likely already carved out so you will have to work within those confines as well. If the company you work for is class C, then you'll be working in class C while you're there. Ripping out infrastructure to change is unlikely to occur without serious business analysis and a huge project with big budget ;)
 
No such thing as class A/B/C networks anymore!

Sounds like they want to keep only the single /24 but have you use VLSM.
 
/usr/home said:
No such thing as class A/B/C networks anymore!

Sounds like they want to keep only the single /24 but have you use VLSM.
Click to expand...

What do you mean there's no such thing as class A/B/C networks anymore? OP has IPv4 not 6 :confused:

Yes OP, as mentioned before, VLSM that thing
 
*The*Dark*Lord* said:
What do you mean there's no such thing as class A/B/C networks anymore? OP has IPv4 not 6 :confused:

Yes OP, as mentioned before, VLSM that thing
Click to expand...

He means exactly that. Classful routing has been deprecated since the early to mid 90s. It was replaced with CIDR. I don't know why anyone would still teach classful as it is as dead as dinosaurs.
 
Nicklebon said:
He means exactly that. Classful routing has been deprecated since the early to mid 90s. It was replaced with CIDR. I don't know why anyone would still teach classful as it is as dead as dinosaurs.
Click to expand...

My school taught both so students would know the differences between the two. However, we mainly focused on classless routing protocols.
But yes this is just a lesson to demonstrate your understanding of VLSM.

Anyways what your going to want to do is create 5 subnets that are just big enough for each department. Then fit those into the 192.168.50.0/24 space. Best practice is usually to fit them in from the largest to smallest network.

For example say the Engineering department needs 82 addresses and the R&D department needs 56 addresses.
The best subnet/CIDR for engineering would be /25 as this would have enough addresses to fit 82 addresses while wasting the least amount. The best for R&D would be /26 as it has enough to fit 56 addresses while wasting the least amount of available addresses.

Next you will fit them into your prescribed 192.168.50.0/24 address space

Using the example above the engineering department would fit first seeing as it is the largest.

So the engineering department would take the addresses 192.168.50.0 - 192.168.50.127
This would be labeled as 192.168.50.0/25

Then the R&D department would take the addresses 192.168.50.128 - 192.168.50.191
This would be labeled as 192.168.50.128/26

Now you will do this for all of them.

If I have explained anything incorrectly please let me know as I am still fairly new to this myself.
 
Last edited:
/usr/home said:
No such thing as class A/B/C networks anymore!

Sounds like they want to keep only the single /24 but have you use VLSM.
Click to expand...

Sadly the CCNA still makes you learn it. I am currently working on mine, and after about 5 pages of A/B/C they just say forget it because CIDR is here friend!
 
I updated my previous post with an example for the OP. Let me know if you guys see anything wrong with my example.
 
There's nothing unusual or ridiculous about this question. It's a bog standard CCNA subnetting question.

Telling your supervisor to "sit on it" and requesting a /16's worth of space for 108 nodes would be ridiculous. :confused:

Sorry if it sounds harsh, but my first step would be to relax and realize you don't know as much as you think you do.
 
markwo said:
There's nothing unusual or ridiculous about this question. It's a bog standard CCNA subnetting question.

Telling your supervisor to "sit on it" and requesting a /16's worth of space for 108 nodes would be ridiculous. :confused:

Sorry if it sounds harsh, but my first step would be to relax and realize you don't know as much as you think you do.
Click to expand...

Well I wouldn't actually say that. I'd certainly think it though. Why would you purposefully want to limit your network like that? What happens if another 150 nodes come in across all departments? You would have to spend a whole load of time moving to a network that actually makes sense.

I'm not saying I know everything. I just don't see why a company would want to limit themselves. There'd be nothing stopping them from using a 172 network. You can set a router to do whatever you want, why use something that just barely works for their current needs let alone their future needs? That's my problem.
 
colinstu said:
Well I wouldn't actually say that. I'd certainly think it though. Why would you purposefully want to limit your network like that? What happens if another 150 nodes come in across all departments? You would have to spend a whole load of time moving to a network that actually makes sense.

I'm not saying I know everything. I just don't see why a company would want to limit themselves. There'd be nothing stopping them from using a 172 network. You can set a router to do whatever you want, why use something that just barely works for their current needs let alone their future needs? That's my problem.
Click to expand...

it's just a test question.... a real world scenario where you would use these skills to solve the problem may be difficult to find... but it's a question who's answer is supposed to reveal whether you know the pertinent information...

for all you know, that 192.168.50.0/24 network is nestled between a .49/24 and a .51/24 that other parts of the company are using...
 
colinstu said:
I'm not saying I know everything. I just don't see why a company would want to limit themselves. There'd be nothing stopping them from using a 172 network. You can set a router to do whatever you want, why use something that just barely works for their current needs let alone their future needs? That's my problem.
Click to expand...

In a real environment this is a valid point. I've got a customer that loves lots of tiny networks and worse still doesn't understand route summarization so they have a /29 in Amsterdam and the adjacent networks in Tokyo and Seattle. They've ended up with a an enormous routing table and distribute it via OSPF. They also insist on running dynamic routing on their firewalls. This has caused no end of nightmares as the routing were too large for the OSPF running on the firewalls.

There are lots of factors one must consider when planning a real network. In this case however none of these apply to you as it isn't a real network and your instructor has told you what he expects you to do. He is doing this to see that you understand subnetting and that is all he is interested in.
 
99vw said:
My school taught both so students would know the differences between the two. However, we mainly focused on classless routing protocols.
But yes this is just a lesson to demonstrate your understanding of VLSM.

Anyways what your going to want to do is create 5 subnets that are just big enough for each department. Then fit those into the 192.168.50.0/24 space. Best practice is usually to fit them in from the largest to smallest network.

For example say the Engineering department needs 82 addresses and the R&D department needs 56 addresses.
The best subnet/CIDR for engineering would be /25 as this would have enough addresses to fit 82 addresses while wasting the least amount. The best for R&D would be /26 as it has enough to fit 56 addresses while wasting the least amount of available addresses.

Next you will fit them into your prescribed 192.168.50.0/24 address space

Using the example above the engineering department would fit first seeing as it is the largest.

So the engineering department would take the addresses 192.168.50.0 - 192.168.50.127
This would be labeled as 192.168.50.0/25

Then the R&D department would take the addresses 192.168.50.128 - 192.168.50.191
This would be labeled as 192.168.50.128/26

Now you will do this for all of them.

If I have explained anything incorrectly please let me know as I am still fairly new to this myself.
Click to expand...

How exactly does this get implemented in Packet Tracer? I keep trying to use different subnet masks and keep getting told that there is overlap. (255.255.255.0 overlaps with everything else). This makes sense... but on the same token every "cheat sheet" out there suggests using 255.255.255.0 mask for the first subnet, something else for the second, etc.

And if I set everything to use the same subnet mask (different gateways)... well than every PC can still ping every other one still for some reason. No routing protocols have been configured yet pinging can still happen between PCs in different networks. I don't get it :confused:

Also. What exactly should the topology for this look like? I was originally thinking a single "main" router (in reality, a L3 Switch I think...? but those can't be emulated in packet tracer. maybe they can... but we haven't talked abotu VLANing at all) with 8 routers connected to it. But.... since L3 Switches don't exist on PT I scrapped that plan.

Second idea was daisy chaining 8 routers in a row, with each router connected to the next, with one switch hanging off each router. This sounds like would be terrible in practice (if one PC ever had to talk to another PC all the way on the last router) but seems to be able to be done in PT.

Finally my last idea was configuring one router to actually have enough ports that I can configure IPs on. The Cisco 2811 with NM-4E and enough WIC-1ENETs seems to do the trick.

Another issue I'm running to is that if I do have to daisychain all the routers, I need extra IPs just so I can connect each interface with the next one (as well as the interface that connects to the switch). Would one subnet be required just so the routers can be properly connected with each other?
 
colinstu said:
How exactly does this get implemented in Packet Tracer? I keep trying to use different subnet masks and keep getting told that there is overlap. (255.255.255.0 overlaps with everything else). This makes sense... but on the same token every "cheat sheet" out there suggests using 255.255.255.0 mask for the first subnet, something else for the second, etc.

And if I set everything to use the same subnet mask (different gateways)... well than every PC can still ping every other one still for some reason. No routing protocols have been configured yet pinging can still happen between PCs in different networks. I don't get it :confused:

Also. What exactly should the topology for this look like? I was originally thinking a single "main" router (in reality, a L3 Switch I think...? but those can't be emulated in packet tracer. maybe they can... but we haven't talked abotu VLANing at all) with 8 routers connected to it. But.... since L3 Switches don't exist on PT I scrapped that plan.

Second idea was daisy chaining 8 routers in a row, with each router connected to the next, with one switch hanging off each router. This sounds like would be terrible in practice (if one PC ever had to talk to another PC all the way on the last router) but seems to be able to be done in PT.

Finally my last idea was configuring one router to actually have enough ports that I can configure IPs on. The Cisco 2811 with NM-4E and enough WIC-1ENETs seems to do the trick.

Another issue I'm running to is that if I do have to daisychain all the routers, I need extra IPs just so I can connect each interface with the next one (as well as the interface that connects to the switch). Would one subnet be required just so the routers can be properly connected with each other?
Click to expand...

Show us what you have first. Your Layer 3 logical topology can be done with various Layer 1 physical topologies.

I can give you all the answers but what good does that do you. I can give you a fish or I can teach you to fish as the old saying goes.
 
I'm confused by this thread as well.

I thought each subnet had to have its own gateway?
 
Status
Not open for further replies.
Back
Top