https://homeostasis.scs.carleton.ca/wiki/index.php?action=history&feed=atom&title=DistOS_2023W_2023-01-16 2026-04-08T04:55:22Z Revision history for this page on the wiki MediaWiki 1.42.1 https://homeostasis.scs.carleton.ca/wiki/index.php?title=DistOS_2023W_2023-01-16&diff=24287&oldid=prev 2023-01-16T18:00:33Z

<p>Created page with &quot;Video from the second half of class is available in Brightspace in the Zoom portal. ==Notes== &lt;pre&gt; Lecture 3 --------- Group reports - can be bullet points, but paragraphs are better - should be self-contained, should be readable by someone who hasn&#039;t done the readings - but you don&#039;t have to summarize the entire paper, just make the points you discussed clear Monday group reports are due Friday Wednesday group reports are due Monday (and responses are due...&quot;</p> <p><b>New page</b></p><div>Video from the second half of class is available in Brightspace in the Zoom portal.<br /> <br /> ==Notes==<br /> <br /> &lt;pre&gt;<br /> Lecture 3<br /> ---------<br /> <br /> Group reports<br /> - can be bullet points, but paragraphs are better<br /> - should be self-contained, should be readable by someone who hasn&#039;t done the readings<br /> - but you don&#039;t have to summarize the entire paper, just make the points you discussed clear<br /> <br /> Monday group reports are due Friday<br /> Wednesday group reports are due Monday<br /> (and responses are due Sunday night)<br /> <br /> Multics<br /> - idea of &quot;utility computing&quot;<br /> - that&#039;s why they cared about security<br /> - Multics was supposed to be distrubuted, in a sense<br /> - terminals everywhere connected by data links (phone lines)<br /> <br /> Multics is before computer networking was a thing<br /> - distributed system was one mainframe with lots of terminals in remote places<br /> <br /> <br /> Resource-Sharing Computer Communications Networks<br /> - so, why this paper?<br /> - because their discussion is &quot;distributed operating systems&quot;<br /> <br /> What is this paper arguing for? What is the technical innovation?<br /> - what is the innovation relative to?<br /> <br /> Circuit-switched networks - old<br /> Packet-switched networks - new<br /> <br /> What were circuit-switched networks designed for?<br /> What are their key advantages?<br /> <br /> with circuit switched, as long as someone was on a phone call,<br /> there was a &quot;wire&quot; dedicated to to that call<br /> - and for long distance, that was a significant amount of resources,<br /> hence the cost<br /> <br /> Originally this was all analog, but then it became digital<br /> - but idea was still to have a dedicated connection between callers<br /> - how do you make a &quot;virtual&quot; wire?<br /> - can use packets to send data<br /> - but have explicit bandwidth reservations and latency requirements<br /> - resources are still reserved for each call to ensure quality<br /> - this is circuit-switched networks<br /> <br /> Packet-switched networks are different<br /> - each packet is considered independently<br /> - normally, delivery is &quot;best effort&quot; - no resource reservation<br /> - so packets can get dropped at any time<br /> - endpoints have to handle retransmission, network won&#039;t do it for you<br /> <br /> Packet-switched networks are fundamentally more efficient<br /> - no wasted resources on reservations<br /> - and is okay because we&#039;re transmitting data, delays are fine<br /> (voice data is much more latency sensitive)<br /> <br /> What did they want this network for?<br /> - file transfers<br /> - email (which is a kind of file transfer)<br /> - remote access to computers (remote terminals)<br /> <br /> What was an IMP equivalent to?<br /> - mostly like an ethernet interface<br /> - but also aspects of a router<br /> <br /> Why was it called the ARPANet?<br /> - (D)ARPA - one of the research arms of the US military<br /> <br /> They are talking about the Internet and IP packets, that&#039;s what they built<br /> <br /> How to read a research paper?<br /> - read the intro and discussion/conclusion first<br /> - see what the authors were trying to say<br /> - look at the figures, see what they say<br /> - look at background/related work, what other systems is this connected to<br /> - look at results - why do they think their system is any good?<br /> - then read the paper from the beginning<br /> - a pure linear read is almost never the best idea<br /> &lt;/pre&gt;</div>

Soma