Networks

There are a lot of ways to do this, but my first take is to group the clients into clusters of 15 or fewer hosts. Each host in a cluster is connected to a hub (hubs are generally 16-port devices, and I like spares). Each of the 7 hubs required for a subnet is connected to a router, and the router is configured to provide DHCP services for its respective subnet. This is commonly available in routers, and makes client configuration easier, as well as relieving some of the load on your (presumed) one server. These 4 routers are then connected to a single router which connects to your server. If the server handles Internet Connection Sharing for the network, the clients should be configured to use the server local address for their default gateways. If not, the Internet access device in use should be connected to the top level router and the clients configured to use its local address as a gateway.

I suspect, though, that the interviewer will be more interested in your skill at crafting a subnetting scheme using TCP/IP. If your server or other Internet connection device supports NAT, you can use a private address range without having to register all of your client machines. The IP range 192.168.x.x is reserved for that purpose. If you use 192.168.x.x as your basic network address, and 255.255.192.0 as a subnet mask, you should be able to set up 4 internal subnets with 254 hosts on each - lots of room for growth. Remember that the first and last address in the range are not available for use by individual hosts. It's too late, and I'm too tired to to lay out a specific plan for you, but I'm sure you can research it yourself. The hardware options are many, but I think the TCP/IP configuration is going to be what your potential employer is really going to be looking closely at. At least, that's what I'd be looking for. Good luck, BTW - I just got my first job interview in 2 years on Monday! It's cold out there...

hi the guy above me gave some great advice but you need to use a class B(150.156.x.x) address not a class C (192.168.1.X), as you know only one octect can be changed in this class
well good luck on the interview

4 subnets with 25 IPs equals can use the follwing IP scheme:

Network: (random private network address)10.0.0.0 Mask: 255.255.255.192
The four subnets would be, with 62 USABLE addresses in each subnet:
10.0.0.1-10.0.0.62
10.0.0.65-10.0.0.126
10.0.0.129-10.0.0.190
10.0.0.193-10.0.0.253
Remember the mask in these network is always the same (255.255.255.192)

Or you can use 10.0.0.0 mask 255.255.255.224. This would give you 30 USABLE adderesses in each subnet, but only give you five extra IPs:
10.0.0.1-10.0.0.30
10.0.0.33-10.0.0.62
10.0.0.65-10.0.0.94
10.0.0.97-10.0.0.126

Tech Forum:
www.bbfree.com/Tech

I also recommend using /24 (255.255.255.0) subnet mask - playing around with Variable Length Subnet Masks (where the Network/Host boundary falls in the middle of a byte) is a real pain - don't do it unless you need to! As general guidelines recommend a maximum of c. 200 nodes per subnet (so broadcasts don't overstretch the Layer-2 domain) this will give plenty of host addresses. If 192.168.x (c. 250 subnet) isn't enough, try using 172.16-31, whic are also unregistered IP ranges (see RFC 1**.

Next, don't use hubs, use switches NO-ONE installs hubs any more (Cisco don't even make them!)

Also, if it's truly a LAN (i.e. in a Local Area) dn't inter-connect your subnets with a router - they cannot process at anything near LAN wire rates. Use a Layer-3 switch, it still allows you to sub-divide the network into LANs (usually Virtual LANs) but forwards at Millions of Packets per Second instead of Thousands for a traditional router.

Best of Luck,
Paul (CCDA)