how can computer communicate with different octect?

imported_itr

Senior member
Mar 2, 2005
900
0
0
i have a computer that has an ip of 192.168.15.x and would like it to talk to a server that has an ip of 192.168.1.x. as you can tell, the 3rd octect is different. a friend told me they can talk to each other as long as their on the same subnet mask, however, i have had no luck with that.
 

spidey07

No Lifer
Aug 4, 2000
65,469
5
76
Are they on the same switched network?

If so the mask needs to be changed to something like 255.255.240.0 - that would place both of those addresses on the same IP network (subnet).

If they are not on the same switched network a router is required.
 

imported_itr

Senior member
Mar 2, 2005
900
0
0
what exactly do you mean if they are on the same switched network? both computers are on the same subnet mask: 255.255.255.0
 

spidey07

No Lifer
Aug 4, 2000
65,469
5
76
A requirement for proper IP communication is the concept of a broadcast domain. Meaning if one station does a layer2 broadcast it must reach every single node. A switch or hub is a single layer2 network. Heck 100 switches all connected to each other is a single layer2 network - a single broadcast domain.

This is really what we call a "network" or a "subnet"

So if both of your hosts are connected to the same switch without any VLANing involved then they are on the same layer2 network and the same broadcast domain.

IP will look at the destination address of a packet going to another host. It will then determine via its own subnet mask if this IP address is on the same "network" or "subnet". If it is then it uses a broadcast to locate the other host. If not then it just sends it to its default gateway (router) and the router handles it from there.

There are other more complicated scenarios where you can and do have multiple subnets on the same broadcast domain, but it is a highly discouraged practice.

-edit- what I mean to say is the switches define the broadcast domain, bounded by routers. Routers set the mask of that IP subnet - they are the gods, the master postmaster for delivery of IP packets...they DEFINE the addressing and mask. The hosts must then be properly configured with the correct IP range, mask and gateway.
 

imported_itr

Senior member
Mar 2, 2005
900
0
0
my computers are connected to a 3com superstack switch, including my server. which means they are on the same network? i still can't talk to the server which has a different 3rd octet. =\
 

spidey07

No Lifer
Aug 4, 2000
65,469
5
76
it won't happen until you change your mask to 255.255.240.0

But in all honesty what you are trying to do breaks every good practice in networking.
 

imported_itr

Senior member
Mar 2, 2005
900
0
0
nothing really, just trying to obtain more knowledge of networking. thanks for the help nonetheless. just curious, don't most major companies seperate their network into different subnets? what are the advantages of this?
 

spidey07

No Lifer
Aug 4, 2000
65,469
5
76
Originally posted by: itr
nothing really, just trying to obtain more knowledge of networking. thanks for the help nonetheless. just curious, don't most major companies seperate their network into different subnets? what are the advantages of this?

segregation of broadcast domains, structured IP addressing for route summarization, minimizing failure domains, eliminating large spanning-trees are the main reasons.

I know it sounds like a bunch of gobblty-gok, but those are the big drivers. We're getting into network design however and it can be quite a lot to swallow.
 

nweaver

Diamond Member
Jan 21, 2001
6,813
1
0
let me try to explain...

There are 2 parts to an IP Address, Network portion and Node portion. Think of the network portion as your zip/city/street and the node as your name/house number. A computer uses a subnet mask to determine what parts of IT'S OWN ip are network/node. So for 255.255.0.0, the first two octets are network (the 255's) and the second two are node (the 0's). When you get outside of 0 or 255, you are getting into binary stuff, and it's a bit rough, I would suggest google for a decent site on bitwise anding :p

So, if you are 192.168.3.8, and the server is 192.168.1.7 (just making these up) then you need to use your SUBNET MASK to determine what part is network, and what part is node. lets assume for now, you have a subnet mask of 255.255.255.0. That would make your address 192.168.3 (network) and 8 (node). The server would be 192.168.1 (network) and 7 (node). So your PC looks at the network portion, and compares them. If they are the same, it sends a broadcast (who as X.X.X.X, tell X.X.X.X), if they are different, it goes to it's default gateway (please send this to X.X.X.X). In this example it's not the same, so it goes to the router.

Change the subnet mask to 255.255.0.0, and it's a different ballgame (or network). Now you have 192.168 (network) and 3.8 (node) looking for 192.168 (network) and 1.7 (node). The network portions match, so the computer assumes it is local, and just "hollers out the window" to find the local computer (aka broadcast, who has 192.168.1.7, tell 192.168.3.8).

Make sense?

Spidey is very smart, and confuses some of us n00bs from time to time. As you hang in networking more, you will realize what a good geek he is :p
 

Imdmn04

Platinum Member
Jan 28, 2002
2,566
6
81
I suggest the OP google for CIDR. Its much easier to explain with calculations of AND operations.
 

nweaver

Diamond Member
Jan 21, 2001
6,813
1
0
Originally posted by: spidey07
And you can't do/teach networking without a whiteboard.

:(


when moving from our remote office to corp, I picked the cube based on Whiteboard realestate (got a huge one on one wall, and a small one on the cube wall)
 

okb

Member
Mar 9, 2005
126
0
0
Originally posted by: spidey07
And you can't do/teach networking without a whiteboard.

:(

No kidding. Try explaining TCP/IP to a bunch of guys who have a hard enough time with English.

itr, if you really want to know more about it then I suggest picking up a good book. It doesn't have to be an ubergeek book, something that does a basic understanding in fairly normal language. Maybe a "For Dummies" book? I've never read any of that brand but I know some people that like them. Pretty interesting stuff once you get over the first hump of understanding how the addys break down.
 

imported_itr

Senior member
Mar 2, 2005
900
0
0
i would like to add security to my local server just in case some hacker breaks into my network. do i add a hardware firewall in front of the server? where would i start?

edit: this server is a web server so i'd like it to be isolated from my internal network.

edit2: the only way i can think of adding security, for now, is to have the webserver on a different subnet 255.255.0.0 and for client computers to have a subnet of 255.255.255.0. this way the clients can't talk to the web server iternally. am i going in the right direction? hmm, with that setup, it looks like the server can talk to the clients but the clients can't talk to the server?

this seems to be a different question than the original post. i'll start a new thread.