Form groups of 2-4 people to discuss the following exercises.

To celebrate a great thesis defense on Wednesday, the topics connect to
Samu Varjonen's thesis "Secure connectivity with persistent identities".
You can find it at http://urn.fi/URN:ISBN:978-952-10-8341-9
(Note the URN link!)

1. We discussed location-based and location-independent naming on the
lecture slides.

Samu's work aims to improve HIP (Host Identity Protocol), which has
as a central goal the separation of a host's identity from its location
information on the Internet. IP addresses are currently used as a sort
of cheap identity ("Oh, you're 129.23.2.1? Welcome back, let's continue
this session!"), while it's primarily a location ("this packet should be
routed to the 129.23.*.* network, obviously").

What are the benefits of providing hosts with location-independent
names? Can you think of applications where this would be particularly
beneficial?

2. Samu's thesis (chapter 2.1) also discusses DNS and problems with it
in particular applications that have appeared only long after the
original design. Can you identify where the original assumptions made at
the time show in the design of DNS and the Internet in general? (For
example, the DARPA network was a military network, one of the expected
threats was that several routers might go down even due to a war, but
that the hosts in the network were all controlled by essentially the
same organization.) Can you identify any assumptions that are clearly
incorrect these days?
 

3. The cloud and services like Akamai are an example of intelligence
moving to the network, while so far the intelligence has been in the
hosts themselves. What kinds of problems and benefits does this bring?

4. Go through the first topic row of the learning goals of the course.
If you were an examiner trying to gauge the students' learning on these
topics, what kinds of questions would you ask from this theme?