https://homeostasis.scs.carleton.ca/wiki/index.php?action=history&feed=atom&title=DistOS_2021F_2021-11-02 2026-05-14T08:42:19Z Revision history for this page on the wiki MediaWiki 1.42.1 https://homeostasis.scs.carleton.ca/wiki/index.php?title=DistOS_2021F_2021-11-02&diff=23432&oldid=prev 2021-11-03T00:04:50Z

<p>Created page with &quot;==Notes== &lt;pre&gt; Lecture 13: Cassandra &amp; Dynamo ------------------------------ - haven&#039;t graded midterm, will do so by next week - I had to submit a grant app on Monday...&quot;</p> <p><b>New page</b></p><div>==Notes==<br /> <br /> &lt;pre&gt;<br /> Lecture 13: Cassandra &amp; Dynamo<br /> ------------------------------<br /> - haven&#039;t graded midterm, will do so by next week<br /> - I had to submit a grant app on Monday<br /> <br /> - will have grades on midterm before proposal is due<br /> - you have until Nov 9th<br /> <br /> Algorithm for making a proposal (for a lit review)<br /> - find a paper you like/find interesting that is related to the course<br /> - make sure it is a good one, published somewhere reputable<br /> (ACM/IEEE isn&#039;t sufficient)<br /> - preferably with reasonable number of citations<br /> <br /> From that paper, find a related set of papers<br /> - related to *one aspect* of the paper<br /> - see citations (who the paper cites, who cites the paper)<br /> - follow graph and search keywords to expand out<br /> <br /> I prefer using a CS-standard citation format (like the papers we read use)<br /> <br /> I prefer individual projects. If you want to do pairs it is possible -<br /> but there has to be a clear division of responsibilities<br /> (so assume no unless it makes sense for the topic and you ask me)<br /> <br /> (Partners make more sense if you&#039;re building something.)<br /> <br /> related to this class: can relate to &quot;distributed&quot; and &quot;operating system&quot;<br /> - can&#039;t just be distributed or OS related<br /> <br /> Try not to start with papers we&#039;ve covered in class<br /> - branch out, search!<br /> - use ones related to your other interests<br /> <br /> You should look for patterns amongst the papers you identify as being related. Your paper&#039;s argument is showing that that pattern exists and how it connects to the papers you find.<br /> - don&#039;t just list summaries of papers, that&#039;s not a lit review<br /> <br /> <br /> On to the papers<br /> <br /> relational databases were inspired by the needs of airline reservation systems<br /> <br /> dynamo was inspired by the needs of e-commerce shopping carts<br /> - so writes (users selecting something to buy always gets saved)<br /> <br /> other systems will refuse writes when they&#039;ll make the system inconsistent<br /> - so client has to retry<br /> <br /> dynamo always accepts writes, even when they&#039;ll make things inconsistent<br /> - so how do they deal with inconsistency?<br /> - when data is read<br /> - use application specific semantics to reconcile conflicts<br /> in the client<br /> (like with source code version control - the programmer<br /> has to manually figure out how to reconcile conflicts)<br /> <br /> A hard problem in any distributed system is determining a canonical order of events<br /> - in principle this can not exist, but we may need an ordering anyway<br /> <br /> A ring is the simplest topolgy that remains connected when you lose a node<br /> &lt;/pre&gt;</div>

Soma