IP Ranges & Netmasks

I am trying to assign ip addresses to a number of depots and want each on a seperate subnet (as all are connected and want to distinguish between them easilly).

Could somebody please tell me if the start addresses and netmaskes are correct for the ip address ranges i have specified in each case...? I'm pretty sure they are but want to make sure as a lot is at stake if it all goes wrong!

Private IP's:
Depot 1:
Range: 192.168.10.0 - 192.168.19.255

Start: 192.168.10.0
Netmask: 255.255.245.0

Depot 2:
Range: 192.168.20.0 - 192.168.29.255

Start: 192.168.20.0
Netmask: 255.255.245.0

Depot 3:
Range: 192.168.30.0 - 192.168.39.255

Start: 192.168.30.0
Netmask: 255.255.245.0

etc.

Thanks

Nick.

change your mask to 255.255.255.0 it will save alot of explanation. unless you have more than 253 hosts per network segment. change the mask to 255.255.248.0

nickcol, if your like me you come to the devshed to learn the details, get the how and why on your question. I don't mind giving you a full explanation if you dont mind providing a few more details.

If you can please answer the following:
1. How many hosts do you have in each depot. (Factor in future expansion/growth)
2. What type of network equipment are you using to internetwork these depots (List the routing protocol as well)
3. How familiar are you with basic TCP/IP Addressing (any knowledge on VLSM or more importantly Supernetting?)

Without going into to much detail before you can reply to my questions.

1. What you are trying to do is called supernetting (conceptually the reverse of subnetting). You are combining multiple network addresses of the same class into blocks. To get this to work across your network you need to either use a classless routing protocol like RIP2, EIGRP and OSPF or you need to static route everything (that could turn into a major headache).

2. The subnet mask 255.255.245.0 is not correct...basically it cant exist (actually it can exist but for our purposes understand it is an illegal subnet mask).


3. Using the subnet mask 255.255.248.0 is also incorrect with the supernets you have defined. Once you have provided the requiste amount of hosts per depot the correct mask can be given...and I'll also demonstrate how to compute that.

A few links for your perusal:

Supernetting: http://encyclopedia.thefreedictionary.com/CIDR
Subnetwork Mask: http://encyclopedia.thefreedictiona...bnetwork%20mask
Classful Addressing: http://encyclopedia.thefreedictiona...ssful%20network

I do realize that this is incorrect with the subnets he has listed as if you read it it will say if you have more than 253 hosts then change to using this mask ( meaning next ) with the assumption he would know the networks associated with this mask without me having to go into detail on why 255.255.245.0 is not a valid subnet. also I noted that he should use a /24. this is because im lazy and he would not have to worry about what equipment he has, whether he knows VLSM which obviously he doesnt ( no offense intended), which routing protocols he is using, or explaining IP classless routing.
a WAN utilizing only three networks does not require any routing protocol, infact statics are preffered due to bandwidth overhead of the routing protocols, or using ODR (based on CDP in hub spoke topology) if cisco routers. The use of RIP or better would come into play when the topology gets more complex ( away from a hub spoke topology or if multiple routes exist where you would want to make routing decisions based on hops, bandwidth, or administrative distance or controlling LSA sync due to bandwidth size) The only valid question to ask here is how many hosts are in each segment?




I really do not believe he is trying to supernet anything he is not wanting to aggregate routes he wants 3 seperate networks.

I think the link for CIDR is a little over kill hehe!

Why make that statement at all? You provided incorrect information to someone who asked for a verification his subnet masking. Yes, his setup is wrong. What he tried to do is called supernetting. What you gave him for an answer was an incorrect mask for the amount of networks he listed for each depot (10 per). A 21 bit mask will only allow for 8 supernets. The correct answer in the case of supernetting is a 20 bit mask which would give him 16 supernets. And telling him to switch all to a 24 bit mask for lack of better explanation may or may not work depending on his current network architecture...but it certainly is a lazy way of answering a post.

You're making alot of assumptions about nickcol's network setup with minimal of information. Granted I agree most likely what he wants to do is not supernetting and that a mistake was made while formulating his subnet masking. That's why I asked the questions I did. Which are still quite valid. Nickcol may or may not have the proper understanding to comprehend the quick quip you posted or to realize the garbled instruction you gave was in fact incorrect. Hence I posted the questions and the links along with what could be wrong about his setup. Kinda gives him an opportunity to gain the knowledge on his one not just get spoon fed an answer. Of course I could be spewing all this for no reason, nickcol might just want an answer and not the knowledge :P but I leave that up to him

Not sure where you were going with this statment but unless you know nickcol's network architecture that whole statement is just conjecture.


Again without nickcol's input youre still just guessing. Depending on his network architecture he may or may not want to do route aggregation. I'm not advocating he use it...I did however point out that his current network masking scheme is using supernetting.

Actually the most important question is how many hosts for each depot. The rest of the questions prove valid by informing us of his network setup and his knowledge of the subject.


Then I'd say you may need to look at the link yourself. CIDR is supernetting...and an incorrect supernet mask was one of the suggestions you gave nickcol to verfiy his work

actualy I was going off one of nickcol's other posts so I have a good idea of is network. and I know why he is wanting to change the IP scheme. he has 3 sites one is using APIPA addressing and the other 2 are using publics. He needs to route between the 3 sites not aggregate the routes. I believe giving a quick solution in a forum is the best route untill the person asked for further explanation. my answer was not incorrect at all, using statics with 3 /24 networks would be the easiest and simplest way to change his networks over with little to no explantion, unless he requires more hosts then he could supernet or use a private class B.



Apparently it is you that is guessing, Not sure where you came to this conclusion he only has 3 networks you are reading way to far into his mask error. One thing Ive learned is when someone asks if their config is correct is to not read to far into it, keep it simple.

From his other post..... current IP schemes...
Main office - range 169.254.0.x Mask 255.255.0.0
small office 1via WAN 169.253.0.x MASK 255.255.255.0
Small ofice 2 via Wan 169.252.0.x MASK 255.255.255.0

Yes juniperr I read that post as well. It merely stated he had 3 networks 2 via a WAN and gave his current addressing scheme. In no way does that provide a "good idea" of his current network configuration.

We obviously have differing opinions on how to provide assistance to users in a forum. I believe a detailed response curtails a need to have multiple follow up answers. Providing as much information as possible, asking detailed questions and teaching by example make for a better post...not just for the user who you are answering but for all the other guests that may read this post.


Actually I wasn't reading into his mask error. I pointed out that his mask scheme is using supernetting and that the mask he used is not correct. I also stated the supernet mask you gave was incorrect and that a 24 bit mask may or may not be correct depending on his network configuration.

One thing I know is that a simple answer may not be a well read answer. You may consider what I stated as reading into the errors of a configuration. I consider it as a thoughtful approach to answering his question.

BTW...Pointing out where you were incorrect is not a personal affront...I promise.

Actualy my answer was not wrong and still stands I just failed to elaborate ( my appologies for this) that he would have to change his network addresses when using the /21. How you got 10 networks per depot after reading the other post is really beyond me ( if that was the case he would want a dynamic routing protocol also) . the only way the /24 will not work is if depot 1 which is using a /16 atm would have more than 253 hosts and would only not work for that one network ( as the other 2 are small depots using a /24 atm). I made my conclusion by the fact he stated using win XP as servers which can only have 10 concurrent connections at a time (leaving to believe this is a small network also).

Rewrite original post hehe....
change your mask to 255.255.255.0 it will save alot of explanation. unless you have more than 253 hosts per network segment then change the mask to 255.255.248.0 and use the appropriate networks such as 8, 16, 24, 32 etc..

If someone says they have three networks at different locations and needs a new IP scheme for each its pretty obvious they have three networks which has a router at each point. In my opinion thats a pretty good idea of the WAN topology.


I will leave you to explain for the rest of this thread.

And yes we both have different opinions on how to answer questions in forums. I like small responses at a time leading to more technical answers as it progresses making poster think about each answer to pose the next question not flooding them with a thousand possibilities right from the get go and getting detailed on issues they will never encounter. It never fails when you offer to much the poster response will get way away from what they need to know for their current issue. (usualy out of curiosity, turning to a discussion way off the topic)

Oh you should have elaborated also..



let him know why it is wrong so in the future he will understand his mistake which was the original reason for the post. knowing his network and equipment plays no relavance to this answer.

Hence the reason for my more detailed initial reply. Going back to my last post...a simple answer is not always a well read answer.

To satiate your curiosity on how I arrived at 10 supernets per depot (as shown in his scheme).

Range: 192.168.10.0 - 192.168.19.255 mask 255.255.245.0

networks involved in his incorrect supernet mask
192.168.10.0 supernet1
192.168.11.0 supernet2
192.168.12.0 supernet3
192.168.13.0 supernet4
192.168.14.0 supernet5
192.168.15.0 supernet6
192.168.16.0 supernet7
192.168.17.0 supernet8
192.168.18.0 supernet9
192.168.19.0 supernet10

You listed 255.255.248.0 which would cut him short two supernets. Now I doubt he really needed over 2500 host addresses but thats what he put down and right now its all we have to work with.

And as far as the 24 bit subnet mask. I agree with you. If he has less than 254 hosts at each depot he should stick to a 24 bit subnet mask.

BTW...its 254 host per 24 bit subnet not 253 host. 256 addresses...1 for the network number, 1 for the broadcast address and 254 left for hosts. Where do you come up with 253??

All in all we are just wasting time bickering back and for at each other until nickcol replies with how many hosts he needs per depot network. Do you think we scared him off?


Well perceptions are what drive us. I perceived your answer was lacking and you perceived mine was overkill. Together it might just make an informative thread.

If you read my statement above the one you qouted



You'll notice the intent to elaborate after receiving his answers.

Can we possibly leave the argument now? You both seem to have posted some excellent and detailed information, but it's starting to turn more into an argument...

Agreed