Real-Time and Embedded Systems (M)

People say Peter Dickman runs a hard course! Of course he does, he's pd! But if anyone has the opportunity to participate in a course given by Colin Perkins, take heed as he can cook up some steamers too.

I'm now at home drinking a 710ml bottle of Beck's, but no less than 35 minutes ago I was walking out of a Real-Time Embedded Systems exam. The once project supervisor turned torturer lecturer was in style today, producing a not-so-bad exam (says he who knows not his result!).

The paper consisted of an answer 3 out of 4 structure with questions roughly falling into the categories of:

Question 1

Question on schedulers as was expected. Did this question as it's the typical RTES question. The content was from the start of the course, slightly easier than the rest and covered to the death in the lectures. Contained a some server points but on the whole focusing on the scheduling.

Question 2

The second of the typical RTES questions. Again, as expected this question was on servers, mainly sporadic ones in my answers (not sure if this should have been the case though).

I had a few concerns with this question as the section of the course on servers was all pretty much similar throughout. This meant that it was slightly like writing the same answer for both of the 8 mark parts at the end of the question (one of which I am sure I got totally wrong!).

Question 3

The third main section of the course was covered in this question, Real-Time networking. Being a toss up between this and question 4 for my final choice of question, and due to my slightly greater knowledge in this area, this question won.

Parts of this question were about networking problems and quality of service concerns, implementations of weighted fair queuing algorithms (WFQ) and comparisons of that with the weighted round robin algorithm (WRR) with respect to RT networking.

The last part of this question was a random 4 marker about the behavior of traffic flow through the network when using WRR vs. WFQ, emphasising on the load on the servers. Took a wild shot in the dark and hopefully scored a few marks.

Question 4

Question 4 covered the RT operating systems section of the course with a mention of the word POSIX was greeted with a pass from me. Mainly about choice of schedulers to run on a Linux based set top box, some stuff on two-level schedulers and the performance comparisons of RT operating systems against general-purpose operating systems with RT extensions.

This paper seemed a whole lot easier than the sample that Colin produced, however that may just be because I don't know what the answers to this exam should be. Well, 1 down 7 to go!
- Chris