Hide menu

TDDD66 Mobile Networks

Questions and suggestion log

Questions and suggestion log

This page simply logs some questions and pointers that appear throughout the course. If you have resources that you find useful in understanding the course material, please email me and I can add links.

Also: My apologies if some of the answers are rushed or require further clarification. The idea of the log is simply to provide access to things that otherwise may only be shared with one or a few students. (For various reasons, I will also try to somewhat modify questions sent by students.)

[3/9/2014] Re. in-flight WiFi: It appears GoGo, the world leader of in-flight Internet connectivity (which has most of the North American market), latest hybrid solution "GoGo ground-to-orbit" uses (i) satellite technologies for data transmission to the airplane, and (ii) their own Air-to-Ground (ATG) cellular network (consisting of hundreds of cell towers pointing to the air) for transmissions back to the ground. There are some information in various GoGo announcements and wikis, but there likely are better sources.

[3/10/2013] Note to self for next year: In addition to muddy cards (briefly discussed in class and recorded in separate document), I have obtained offline comments from a group of students regarding: (i) assumption discussion could be added to assignments (e.g., start simple, build as needed to get same/similar conclusions as would be expected with real system), and (ii) timing [and maybe number] of lab slots (schedule). Additional comments and suggestions that may help improve the course are welcome ...

[19/9/2013] Hint for the assignments: To write an event-based simulator you want to identify the different states that the system can be in (e.g., currently downloading segment X and rate r) and what events can causes changes in these states (e.g., a change in the download rate, the download completion of a fragment, or - in the second assignment - the arrival of a new client). In the simulations you simply identify the next event (state change) that will happen and progress your "simulation time" to the time of that event, at which you up-date the current state (including how much the clients have downloaded, for example). Also, note that the second assignment may require you to keep track of more information per state (e.g., the conditions seen by each client may be a sub-state of the overall system state, and you also need to keep track of the state of the proxy).

[10/9/2013] Question: Could you, in the context of caching, for example, explain what the project should be about and what we are supposed to do?

You will have a fair bit of freedom in exactly what you do, as on any of these topics, there are probably an unlimited number of things one could do. Of this reason I expect that each group's problem and report will be fairly different. (As this is the first year for the course and this type of project, I want to see how precisely each group is able to formulate their problem in the first week reports; however, my guess is that some project will start broad and become more and more focused for each week, while others may take an opposite or different routes to reach the end state.)

However, looking ahead, I would like to see that you (either now, or in the next few weeks, if you start broad) try to find a sub-problem that you think that you in some way can analyze, test, or otherwise investigate deeper (e.g., in the case of your caching context, you could imagine doing a performance comparison of two caching policies, look at the value of prefetching, investigate how much of all the content actually can be cached, how much of it results in cache hits, etc.). Many of these are things you can investigate using small-scale experiments (e.g., looking into HTTP headers and caching rules of regular proxies), running simulations, perform some basic calculations, etc., just to give some examples.

I will also try to say a few more words on this in tomorrow's class.