Compsci 82, Fall 2009, Final II Questions

These are the class answers to the second set of online final questions.

Question 1 text   Question 1   Multiple Choice Average Score 0.52 points  
  Netflix, Youtube, Facebook and other hosts that serve lots of content use Content Delivery Networks (CDNs) like Akamia, Limelight, and Level 3 -- we mentioned Akamai and CDNs in class. CDNs get content to users/clients quickly by replicating the content in many places throughout the world as well as using protocols other than IP to serve packets. Which of the following is LEAST likely to be a reason for using a CDN to host a website?
Question 1 answers
Answers Percent Answered
When content is replicated by a CDN it is likely that Google's (or Bing's) web-crawling spiders will find the content multiple times, making a site distributed using a CDN more likely to appear high in Google's rankings than a site that doesn't use a CDN. So, using a CDN will increase a site's Google page-rank or a Bing ranking because a site replicated 100 times will be 100-times better (or some fraction better) just because the site is stored on multiple servers throughout the world even though the site has a single URL, e.g.,
Websites need to cope with unexpected surges in demand, e.g., when Michael Jackson dies, Tiger Woods gets in an accident, or President Obama gives a major speech. Replicating content and using CDNs helps ensure that spikes/surges can be coped with on an as-needed basis.
Websites that are media-rich, or content-rich, in terms of hosting photos, videos, sound, etc. need CDNs to get the big, rich content to users quickly. If every YouTube video was served from one building, coping with the constant demand would be difficult in terms of quality of service.
If YouTube and Hulu both use CDNs to ensure fast, non-jittery access to their video content, then other sites will likely need to use CDNs too so that they're competitive in the marketplace. Similarly for Flickr, Smugmug, or other Photosharing sites --- basically once CDNs are used by one segment of a marketplace, the likelihood of their continued use is increased.
Unanswered 2.273%

Question 2 text   Question 2   Multiple Choice Average Score 0.71 points  
  In his guest lecture Prof. Cox told us that 2 billion photos were uploaded to Facebook each day. This website/blog:, claims that 2.5 billion photos are uploaded each month (as of Dec 2, 2009). This is roughly 83 million photos/day. If each photo averages one Megabyte (one million bytes) then there are 83 million Megabytes uploaded each day. How many terabytes is this? Note that a Gigabyte is a billion bytes and a Terabyte is a trillion bytes (roughly). A billion is a thousand million and a trillion is a thousand billion.
Question 2 answers
Answers Percent Answered
83 Terabytes
8,300 Terabytes
83,000 Terabytes
83,000,000 Terabytes
Unanswered 2.273%

Question 3 text   Question 3   Multiple Choice Average Score 0.71 points  
  Which one of the following was NOT mentioned by Prof. Cox as a reason to be concerned about Facebook?
Question 3 answers
Answers Percent Answered
Facebook knows who your friends are, that's a cause for worry compared to an online advertising company like Doubleclick which doesn't have that knowledge.
It's hard to really remove data (e.g., photos) from Facebook, it may not appear in your Profile/info space, but the data is likely still stored in Facebook's servers.
Facebook is NOT protected by the Communications Decency Act (CDA) section 230 from posts that are defamatory because only your friends can post information about you, so Facebook isn't protected against defamation.
Facebook has sacrificed some security to make the site more usable. It's hard to get fast, easy-to-use sites with security.
Unanswered 0%

Question 4 text   Question 4   Multiple Choice Average Score 0.79 points  
  In his guest lecture Thomas Narten talked about issues with IPv4 and why we should change to IPv6. Which one of the following was NOT mentioned as a reason to make a switch and to make it soon?
Question 4 answers
Answers Percent Answered
With IPv4 many people use NAT (Network Address Translation) to cope with the need for multiple devices to share one actual IP address. But NAT has problems, often it's at fault when the Internet doesn't appear to work from a consumer perspective.
It's possible to make a switch to IPv6 without doing away with IPv4 -- the two protocols can co-exist making a transition more possible than it might first appear to be.
A market for blocks of IP addresses might develop that would/could cause problems with both engineering and social aspects of the Internet. It's possible such a marketplace could be illegal, but it would still happen if some users/groups were advantaged by it even if it meant trouble from an engineering perspective to swap blocks of addresses.
The Internet will likely stop working in two years because it's unlikely we'll make a switch to IPv6 by then. When the Internet stops working, consumers will be inconvenienced, businesses will fail, and in a worst-case scenario countries might find themselves bankrupt or at war because of difficulties in communicating without the Internet.
Unanswered 0.568%

Question 5 text   Question 5   Multiple Choice Average Score 0.55 points  
  Mark Webbink, the former chief counsel for Red Hat, gave a guest lecture to our class. Which one of the following is NOT true about him and his career?
Question 5 answers
Answers Percent Answered
He is both a lawyer and a hacker -- writing serious and widely-used open source software before going back to law school to become a lawyer.
He was the statistician for the men's basketball team for 20 years.
He uses an iPhone.
When he first started at Redhat the company lost more than $200 million/year when it went public and before they used a subscription-based business model.
Unanswered 0%

Question 6 text   Question 6   Multiple Choice Average Score 0.83 points  
  Brian Williams, the anchor for NBC news, skyped into class on the last day of class. Which one of the following is NOT true about the discussion that occurred?
Question 6 answers
Answers Percent Answered
He talked about some of the serious problems facing the country, such as that of a bridge in New York City/Harlem that needs replacement now.
He mentioned that Wikipedia doesn't always get facts straight, though it's sometimes a good place to get started with information-gathering.
He spends so much time reading print newspapers and magazines that he almost never catches up with blogs or online sources of news.
He made light-hearted references to Prof. Astrachan's abilities as a touch-football player.
Unanswered 0.568%

Question 7 text   Question 7   Multiple Choice Average Score 0.78 points  
  In class we saw a comparison of the sites and images returned by (The Chinese Google site) and (the U.S. Google site) using this website: for the search queries tianaman and tiananmen. The results returned were very different for the second of these queries, and substantially similar for the first query. Which of the following best explains this observed phenomenon.
Question 7 answers
Answers Percent Answered
The Google in China site does not store photos or results of anything that the Chinese Government would view as presenting China in a bad light. The query tianaman is misspelled, so the Chinese site shows images from the U.S. site since the Chinese site cannot handle misspelled queries in English.
The Google in China site is looking for specific queries and treating them differently, essentially censoring the returned results when a search query matches a list of those for which different results should be returned, like tiananmen. Since the query tianaman is misspelled, it doesn't get caught by the literal-matching filter that looks for specific key words, thus unfiltered/uncensored results are returned for the misspelling.
The English words and spelling of tiananmen and tianaman are actually translated into completely different words in Chinese. The first translates into 'Gate of Heavenly Peace' which is why the Chinese site shows images of beauty and color whereas the U.S. site doesn't because the English word just matches the square that is the site of an event in 1989 in which hundreds were killed.
The Google did-you-mean query fixing mechanism doesn't work in the Chinese Google site, so they query isn't corrected on that site as it is on the U.S. Google site.
Unanswered 0%

Question 8 text   Question 8   Multiple Choice Average Score 0.89 points  
  BGP, or the Border Gateway Protocol, is used as the protocol for which of the following?
Question 8 answers
Answers Percent Answered
It is the protocol for communication between the large Autonomous Systems (ASs) that make up the Internet, i.e., how the networks that make up the network of networks (that is the Internet) communicate.
It is the protocol for streaming videos that don't require the same quality of service as do services like email, e.g., streaming videos can cope with a few dropped packets without adversely affecting what consumers see, whereas dropped packets in email would be a problem.
It is the protocol by which ISPs in one country communicate with ISPs in another country across international borders.
It is the protocol that YouTube now uses in Pakistan as a result of Pakistan shutting down YouTube accidentally previous to the use of BGP.
Unanswered 0%

Question 9 text   Question 9   Multiple Choice Average Score 0.87 points  
  Guest speaker and Duke alum Luis von Ahn told a story about a five-letter word used by the group Anonymous that was trying to circumvent the reCaptcha system put in place by Time Magazine when the group was working to get Moot voted the most influential person in 2008. What was the five letter word that the group used to try to poison the reCaptcha system?
Question 9 answers
Answers Percent Answered
peter 2.273%
proxy 9.659%
Unanswered 0%

Question 10 text   Question 10   Multiple Choice Average Score 0.29 points  
  Sarah Cohen, former database editor for the Washington Post and current Duke Professor of Public Policy, spoke to the class. Which one of the following was NOT one of the claims she made in her talk?
Question 10 answers
Answers Percent Answered
CORRECT Reporters use data-mining techniques to look for trends, statistical patterns, and evidence when looking at data sets -- these statistical patterns could be the smoking-gun an investigative reporter looks for.
Despite open records laws, has delivered fewer than 1,000 of the 100,000 datasets promised for delivery to the public.
Most information for news and journalists isn't online, it's in filing cabinets or on someone's desk, or just not accessible without looking hard.
The Washington post has created a few websites for understanding data and news that get reasonably extensive hits and interest -- these sites were created for business interests by the Washington Post -- they benefit the Post in economic terms.
Unanswered 1.136%