Cabling the Annexe: Three Rooms on One /24
St Xavier's annexe building is being cabled for a small computer laboratory. The ISP allocates one block, 192.168.50.0/24, and the network is split by room: Laboratory 58 PCs, Library 20 PCs, Admin office 12 PCs, plus 3 shared printers that sit in the Admin room. The electrician's nephew, who did the first configuration, gave every room the mask 255.255.255.224 and the addresses 192.168.50.1–192.168.50.30 for Admin, 192.168.50.33–192.168.50.62 for Library and 192.168.50.65–192.168.50.120 for the Laboratory. In the Laboratory, PCs 192.168.50.70 and 192.168.50.110 cannot ping each other even though both switch port LEDs are green. The Admin printer at 192.168.50.20 cannot be reached from any Laboratory PC. No router is installed yet; rooms are on separate switches joined by one uplink cable.
1. With mask 255.255.255.224, explain why 192.168.50.70 and 192.168.50.110 cannot ping each other, and why the green LEDs prove nothing.
Read the worked answer
A /27 mask (255.255.255.224) leaves 5 host bits, so each subnet holds 2⁵ = 32 addresses in blocks of 32: .0–.31, .32–.63, .64–.95, .96–.127, and so on.
- 192.168.50.70 lies in 64–95 (subnet 192.168.50.64/27).
- 192.168.50.110 lies in 96–127 (subnet 192.168.50.96/27).
The two PCs are therefore on different IP subnets, even though they are plugged into the same physical switches. A host checks the destination against its own mask: if the destination is inside its subnet it ARPs for it directly, otherwise it sends the frame to its default gateway. No router and no gateway exist, so the frame is dropped and the ping fails.
The green LEDs only report that the physical link and the switch port are up (layers 1 and 2). They say nothing about IP addresses, masks or routing. A flat layer-2 segment will happily carry frames between two hosts that IP will never let talk to each other.
Two valid fixes: (a) put both PCs in one subnet by using a wider mask, e.g. /25 covering .0–.127, which is enough for the 58 lab PCs; or (b) install a router, give every PC a correct gateway, and let it route between subnets. Since the lab needs 58 hosts, /26 (62 usable) is the smallest mask that holds the whole lab in one broadcast domain.
2. Redesign the addressing with VLSM, starting at 192.168.50.0. For each room state the network address, mask and usable range, and give the Laboratory's broadcast address. Justify the mask chosen for the Admin room.
Read the worked answer
Size each subnet by the number of devices, remembering that usable hosts = 2^h − 2.
| Room | Devices | Smallest mask | Network | Usable range | Broadcast |
|---|---|---|---|---|---|
| Laboratory | 58 | /26 (62 usable) | 192.168.50.0/26 | .1 – .62 | .63 |
| Library | 20 | /27 (30 usable) | 192.168.50.64/27 | .65 – .94 | .95 |
| Admin | 12 PCs + 3 printers = 15 | /27 (30 usable) | 192.168.50.96/27 | .97 – .126 | .127 |
Allocate largest first, so the /26 does not get fragmented. Mask values: /26 = 255.255.255.192, /27 = 255.255.255.224.
The Admin room is the trap. Twelve PCs alone fit in a /28 (16 − 2 = 14 usable), but the three shared printers are also layer-3 devices needing addresses, making 15 hosts. A /28 gives only 14 usable addresses - one short, and the usual symptom is that the last device configured simply cannot communicate or duplicates an address. So Admin needs /27. If the printers were moved to a separate management subnet, /28 would become viable, but as described they sit in the Admin room and must be counted.
This plan still leaves 192.168.50.128–192.168.50.255 free for future growth and for point-to-point links to a router.
3. The Admin printer at 192.168.50.20 must be reachable from the Laboratory. Explain why its current address is now a problem and list the changes needed.
Read the worked answer
Under the VLSM plan, 192.168.50.20 falls inside the Laboratory's range (192.168.50.0/26, hosts .1–.62). That is a conflict on two counts: the printer would either duplicate an address already given to a lab PC, or it would sit on a subnet different from the Admin switch it is physically plugged into. Either way, lab PCs would ARP for .20 locally and reach the wrong device or nothing at all.
Changes needed:
- Renumber the printer into the Admin subnet, e.g. 192.168.50.100 with mask 255.255.255.224. Its gateway is the router interface on that subnet, e.g. 192.168.50.97.
- Connect it to the Admin switch, not to the lab uplink, so its layer-2 location matches its layer-3 subnet. If it must stay on the lab switch, the switch port has to be configured for the Admin VLAN.
- Install the router with one interface per subnet (192.168.50.1/26, 192.168.50.65/27, 192.168.50.97/27) and enable routing between them.
- Set the default gateway on every PC and on the printer to the router address in its own subnet; without a gateway the lab PC sends nothing off-subnet even with correct masks.
- Check for duplicate addresses across the old .1–.120 plan before reusing any number, since the old configuration overlapped the new subnets.