First IP address of a subnet.

First IP address of a subnet.

by Paul Luc Mc Intyre -
Number of replies: 1

Hello, I was wondering why we sometimes take into consideration the network address for IP address allocation and sometimes we do not.

In the [2019 Final Exam, Problem 2, Question 1], for subnet 1 we start with router interface k: 1.0.0.1 from IP prefix 1.0.0.0/23. Why isn't this k: 1.0.0.0?
In the [2018 Final Exam, Problem 2, Question 1], we assigned 100.0.0.0 to the router interface from IP prefix 100.0.0.0/22. This makes more sense to me.
We consider the first address to be the network address, the router interface's IP address right?
If so, shouldn't we always start with the router interface's IP address as the smallest address?

Thank you very much and I wish you a great afternoon!

In reply to Paul Luc Mc Intyre

Re: First IP address of a subnet.

by Tom Louis Demont -
Hello Paul,

This is a good question.
First, do not forget that the corrections are indicative and do not span the whole correct answers space.
Here, the explicit requirement is to have a broadcast IP address and an address for every interface.

You have seen in class that every subnet also have a network reserved address that corresponds to no end system. Conventionally, this address is the first one of the subnet's prefix (starting by 0). However, you have not seen the use of this address (which I personally only know to be used in OSPF internal routing protocol to designate some subnets in the routing databases, but even there it does not seem so necessary to eat one address per prefix) and this is why those question do not make a big deal of considering you gave the subnet this address or not.

As this is not part of the explicit requirements, and the correction do not make explicit mention of this address, understand that this is because the network address is not here the point the exam wants to evaluate you on. You remark that the prefix size is not changed by +-1 address, this has the benefit to not change the answer between students that considered or not considered network address.

Why is there still a difference between both corrections? Just that one followed the convention reserving the address 0, and the other did not. The prefix allocation procedure is what is being evaluated here and is not fundamentally changed by this address' presence or not.

I hope this answers your concern!
(btw, don't hesitate to join the Piazza forum http://piazza.com/epfl.ch/fall2021/com208 for other questions)