Presentation is loading. Please wait.

Presentation is loading. Please wait.

CS 4700 / CS 5700 Network Fundamentals Lecture 4.5: Review from last week Revised 1/12/14.

Similar presentations


Presentation on theme: "CS 4700 / CS 5700 Network Fundamentals Lecture 4.5: Review from last week Revised 1/12/14."— Presentation transcript:

1 CS 4700 / CS 5700 Network Fundamentals Lecture 4.5: Review from last week Revised 1/12/14

2 Intro to the class 2  Use Piazza for everything  Really Unless it’s something truly personal/private  Don’t cheat  I will catch you  If you have questions about what is cheating, ask  Regrading policy  2 strikes and you’re out  Partners allowed on all projects

3 History of the Internet 3  What were the newideas that revolutionized/distinguished the Internet from other prior networks?  Packet switching (why is this a good thing?)  No global control  Layering to glue together different network types  What was the original Internet called? What was its design goal?  ARPANET, resilient to catastrophic failure (nuclear)  What do you think is the biggest threat to Internet success?

4 Architecture 4  What are the 7 ISO layers? Which ones are used in practice?  What does each layer specify?  How do we combine layers in data transmission?  What are examples of violations of strict layering?

5 Saltzer’s paper 5  Design principle, not a law  Low-level functionality begets assumptions, and when you assume, you …  More generally: Systems will fail with some nonzero probability. The Internet is a very large computer system, so something is almost always going to fail.  Defining endpoint is critical  Some endpoints are at the low level

6 Student comments 6  10/61 students submitted by deadline  40/61 by 10:30am  “I feel that author's assumption of considering an error rate to be acceptable in the system is not the best approach, specially[sic] when today computer systems are used for critical functions like healthcare and can also lead to human fatalities”  Lack of empirical data/analysis/statistics  Not enough solutions, too vague  Makes assumptions about goals of underlying system(s)

7 Clark’s paper 7  Primary goal: Multiplexed use of existing (disparate) communication networks, separately administered  Secondary goals: Priority order shaped the Internet  Make it work now, worry about accounting later (but make it cheap to build)  Keep it simple, but know that it might be less efficient

8 Student comments 8  22/61 on time  37/61 at 10:30am  Why is 100 bytes a reasonable packet size?  What is fate-sharing, why is it easier to engineer?  Location of service-consumer and failure are the same  Why is it hard to support resource management and accountability? Cost-effectiveness & security?  Why TCP?  Why ACK on bytes?  Dedicated network vs. Best-effort

9 Project 1 highlights 9  Key skills  Work with TCP sockets (optionally SSL)  Follow application-layer protocol  Parse data  Use turnin script from (most) any CCS department server  It’s a valid CCS server if you can read /courses/cs5700s14/bin  You can work in teams  Use any language you want (doesn’t have to be C)  Document your code, read instructions carefully!  Get started early. I go to bed early.


Download ppt "CS 4700 / CS 5700 Network Fundamentals Lecture 4.5: Review from last week Revised 1/12/14."

Similar presentations


Ads by Google