Subnetting has to become second nature to any successful network engineer. In studying and reviewing my subnetting for an upcoming exam, I came across what I believe to be an incorrect subnetting answer from a Cisco book.
If you have a 19-bit subnet mask, how many subnets and how many hosts do you have?
A. 8190 subnets, 126 hosts
B. 524,288 subnets, 32 hosts
C. 524,286 subnets, 30 hosts
D. 65,234 subnets, 62 hosts
I'm pretty sure the answer should read 8190 hosts in a class B, and 6 subnets because after I did the calculation the mask is 255.255.224.0
I agree with you that there appears to be no valid answer there. There are two ways to think of this question.
1. If we assume it's a class B range, then we would have 8 subnets (2^3) including the zero subnet. And 8190 (2^13-2) hosts.
2. If we assume it's a class A range, then we would have 2048 subnets (2^11) including the zero subnet. And 8190 (2^13-2) hosts.
Usually when there is more than one correct answer, Cisco exams will give you only one of them among the answers, so you just choose it. Otherwise, the question will indicate directly or indirectly which class it means, A, B or C so you can decide.
that means your assumption is based on the router configured with IP SUBNET ZERO sometimes some questions will disregard subnet zero
Thats absolutely correct. But since the question did not mention any thing in this regard, I had to make an assumption and search for the answer, if the answer is not there, I could have tried without zero subnet. Any way, no answer is there, with or without!!
By the way, recent routers (IOS 12.0 and above) have the ip subnet-zero enabled by default. So it's the default behavior to accept subnet zero.
One other way to look at this. Assume a class A network, so the first 8 bits are the network mask. Add 19 bits of subnet bit mask so a total mask length 19 + 8 = 27. From this view the answer would be B or C. Last time I studied for exams it was always (2^19) -2 so I would lean towards C.
19 bit subnet mask it would have to be a class A network.
Bravo Perlhack!! Thats quite an interesting approach. But I have to say the question is extremely vague/ambiguous to say it's a "19-bit subnet mask" with an extra 8 bits for the class A. I've never seen it said this way.
I'm more towards answer C too. However, the -2 shouldn't apply to number of subnets, only to the hosts. Subnets should be either 2^19 or 2^19-1 (-1 for the zero subnet). This adds more doubts about the question I guess.