Homepage CptS 565:
Advanced Topics in Distributed Computing Systems
Taught Fall 2015 as CptS 580 (Section 2)
Fall MW(F) 415-530 in Jackson 52/53 (Pullman) and TBD in Tri Cities
Today we will talk about IPFS and how it can be extended.
Today we will finish a little on GridStat and then look at some coordination languages, Linda, FT-Linda, and Jini.
Wednesday we will have a very interesting discussion and design session involving IPFS and how to extend it to include discussion forums. Please read this article (and poke around at a few of the links) ahead of time. Reminder: 25% of your grade is participation, and this is a great opportunity.
Wow, it is October already!
Today we will go over wide-area middleware for the power grid, GridStat. Next week we will have topics TBD.
Today we will discuss the literature survey report (“LitSurv”) assignment, discuss upcoming paper presentations, finish QuO, and discuss each student’s particular interests in this class.
Reminder: we WILL have a lecture on Friday.
Today we will discuss runtime issues per the question I posed on the 20th. We will then briefly review middleware, and possibly start on QuO. I’ve decided that we will discuss the Schantz retrospective paper after we do QuO so you have a much broader perspective.
Here is the last page of a curious final exam given last fall by some unnamed professor somewhere.
Here is Homework #1.
Today I will cover an overview of dependable computing for most or all of the lecture. Here is the source for this, but I do not expect you to read it unless you have a specific need to (its very dense verbiage).
Reading assignments:
1. Look through the distributed systems paper sources below, and write up a half or full page on what topics interest you, and why.
2. Rick Schantz paper on Distributed Computing at BBN 1970-1995 (try to finish by next Monday)
3. QuO paper (try to finish by next Wednesday)
Now to runtime issues. When I teach 564, I always have this question on HW1:
Choose [1 or 2] real distributed application program(s) that you are
familiar with (or could become familiar with if you don’t know any (!!!)). For each application,write a page that includes the following:
·
A
paragraph or two giving an overview of the application. Make sure to include which generic architecture
(remote access, client-server, publish-subscribe/event-based,
…) you believe best categorizes the application, and why it does
so. Be sure also to describe what
application-specific activities the various pieces in that architecture
(clients, servers, whatever) are doing: what they are requesting, etc.
·
A
paragraph or two describing whether you believe the following runtime issues
are important to users of this application, and why you believe it (e.g., what
happens in its absence):
o
Low
latency network connections
o
High
bandwidth available
o
Perfectly
“consistent” and “correct” replies/answers/service versus an inconsistent or
approximate reply/answer/service
So think about this, and come prepared next time to discuss an application you think of in the above terms. Note: I am NOT asking you to write ANYTHING, the verbiage above is what they turn in with HW1.
Today we will go over the syllabus. We’ll also talk about the history of distributed computing.
Next few lectures will cover:
· Your backgrounds and interests
· Distributed Systems Paper Sources
· Discussion of applications and run-time issues (see below)
· Discuss the literature survey project
This probably gets us through the end of Thursday. After that the class will consist of
· Me lecturing on an overview of dependable computing
· Me lecturing on QuO (probably 3 lectures)
· Me preaching on GridStat (1 lecture)
· Me maybe lecturing on Cloud Computing
· Student paper presentations (with vigorous discussions)