Cerf's thesis is that the internet needs to be re-imagined for a variety of reasons - the increasing lack of address space, the security problems that haunt the net, and the scope for adding features among others.
Today's talk was similar to one he gave last year; some of the highlights can be seen in this video clip.
His presence, suited and booted, was inspiring of its own accord, more so given that a man of his eminence is willing to admit to mistakes, one of them being his decision to go with a 32-bit address space (IPv4).
His keynote was dogged by one factor common to these speeches - too much content which could not be adequately compressed into the time allocated to him.
However, he did enumerate a number of issues that should give students direction for their next project, such as the interplanetary internet - which is not some wild scientist's dream.
Cerf gave a brief history of the internet from 1969 when it had just four nodes.
From that point onwards, it has become an incredibly complex organism, with 2010 stats showing up nearly 2 billion users, more than 750 million hosts, a billion PCs and 4.5 billion mobiles with connectivity.
Some of the ideas Cerf sees in the future of the internet include sensor networks, similar to the one he has set up in his house.
He cited this as an example of an expected wave of sensor data making up the internet of things. In other words, wherever there is data, the potential exists to value-add as an innovation.