DOC PREVIEW
Duke CPS 214 - Problem Set 1

This preview shows page 1 out of 4 pages.

Save
View full document
View full document
Premium Document
Do you want full access? Go Premium and unlock all 4 pages.
Access to all documents
Download any document
Ad free experience
Premium Document
Do you want full access? Go Premium and unlock all 4 pages.
Access to all documents
Download any document
Ad free experience

Unformatted text preview:

DNSNetwork Operator for a Day (with rcc)Understanding BGP using table dumpsUnderstanding IS-IS Using Packet TracesCPS 214: Computer Networks Distributed Systems January 23, 2007Problem Set 1Instructor: Prof. Bruce Maggs Computer Science Department, Duke UniversityThis problem set has f our questions, each with several parts. Answer them as clearly and conciselyas possible. You may discuss ideas with others in the class, but your solutions and presentationmust be your own. Do not look at anyone else’s solutions or copy them from anywhere. (Pleaserefer to the Duke University honor code).Turn in your solutions in on February 6, 2007 in class.1 DNSIn the first part of this question, you will perform some hands-on DNS queries using dig and playwith DNS lookups from various applications to understand more about the DNS. In the secondpart of this question, you will implement a variation on a stub DNS resolver.RFC 1035 may be helpfu l for answering some of these questions.1. In this question, we’ll warm up by learning a few things about Duke University DNS setup.(a) What are the authoritative nameservers for duke.edu? How long will your resolver cachethe records pointing to these n ameservers? What are the Computer Science Departmentauthoritative nameservers (i.e., , for the domain cs.duke.edu)? Give two benefits oftopologically diverse authoritative nameservers. Why do NS records return names, ratherthan IP addresses?(b) What is another “canonical name” for the Duke University Web server?(c) What is the primary mail exchanger for cs.duke.edu?2. Now that you’ve had some experience playing with dig, in this part of the problem, we’llimplement a stub r esolver that performs iterative DNS queries. Most of the time, stub resolverssend queries with the “RD” (Recursion Desired) bit tur ned on. In this problem, you are notallowed to use the recursion bit. Of course, you are welcome to solve this problem any wayyou like (C,C++,JAVA,Phyton,Perl,shell scrip t, etc). If you prefer, you may use the Rubyskeleton code provided athttp://www.cs.duke.edu/courses/spring08/cps214/hw/ps1/dns-resolv-rb.tgz. This may save you the trou ble of figuring out which modules to use,instrumenting your own performance measurements, etc.(a) Why do stub resolvers typically set the RD bit?(b) Implement a stub resolver that performs only iterative queries to resolve A records. Toanswer the next question, you’ll want to make it possible to provide an option to yourprogram to control the root nameserver.Your resolver need not do anything special as far as caching, etc., but you should handletimeouts (e.g., , querying the next preferred authoritative nameserver if the first doesnot respond).P1-1Just make sure you can (1) point it at different root nameservers and (2) measure thetime taken to resolve a query (the skeleton code is instru mented for this).(c) Use your query to resolve (1) www.cs.duke.edu and (2) www.nytimes.com at the fol-lowing nameservers.• a.name-servers.net (198.41.0.4)• f.name-servers.net (192.5.5.241)• m.name-servers .net (202.12.27.33)• a.gtld-servers.net (192.5.6.30)(i) Through what sequence of nameservers was each query referred? How long did eachreferral step take? Based on this, what fraction of DNS query time is saved bycaching at local resolvers?(ii) What is the first referral when you send a query www.cs.duke.edu to a.gtld-servers.net?Is the answer the same everytime? Why or why not?(iii) How do stub resolvers typically choose root nameservers?Please hand in your code to this p roblem as well. The co de should be commented. Also, include aREADME file with instructions for compiling/executing your co de.2 Network Operator for a Day (with rcc)To work on this problem, you will need the following three resources:• The first routing table dump (or dumps , if you need them) on J anuary 23, 2008 from the In-ternet2 backbone network. BGP table dumps are available at:http://ndb2-blmt.abilene.ucaid.edu/bgp/STTL/2008.01/RIBS/. You should choose th e first dump made on January23, 2008 (rib.20080123.0153.gz).1. To parse rib* files you may use Marco d’Itri zebra parser athttp://www.linux.it/~md/software/zebra-dump-parser.tgz(a) Other than the sessions to private AS numbers, what are the ASes with the most numberof eBGP sessions?(b) At what routers does Microsoft have eBGP sessions to Internet2? (Hint: You will firsthave to figure out Microsoft’s AS number!)(c) Note th at Microsoft is corporate, but Internet2 is supposedly a research and educationnetwork; why might Microsoft have eBGP sessions to Abilene?(d) What prefixes that are advertised by Microsoft are reachable from Internet2? Whichrouting table did you look at to answer this question (and d oes it matter)?2. Observe an output of running rcc verifier at:http://www.cs.duke.edu/courses/spring08/cps214/hw/ps1/rcc-html/(a) Click on “IS-IS Errors” and then on “MTU Mismatch Checks”. What is an MTU mis-match, and why could it cause a problem? The pair of interfaces in question start withge-*, which typically stands for “gigabit ethernet”. Which value is likely the correctvalue for the MTU?P1-2(b) Under “BGP Errors”, click on “Information Flow”. These warnings indicate places wh erean import or export policy was configured in different ways on different routers f or thesame neighboring AS. What is a reasonable explanation for why “anomalous import”(i.e., , different import policies on different neighboring routers) might be a reasonablething for an operator to do?(c) Under “BGP Errors”, click on “iBGP Signaling”. What is meant by an “iBGP SignalingPartition”, and why is it bad?3 Understanding BGP using table dumpsFor this question, you will need to download the Routeviews routing table fromhttp://www.cs.duke.edu/courses/spring08/cps214/hw/ps1/oix-full-snapshot-2008-01-20-1800.dat.bz2Thisfile contains a Cisco BGP4 routing table snapshot, taken at Oregon Route Views (http://www.routeviews.org/) on January 20, 2008. (Beware: Th is is a text file that is 13MB, compressed.You sh ou ld be able to analyze it without uncompressing it using, for example bzcat, grep, less,searching into the file - be patient when searching this is a really huge file !!!)If you are curious about what other snapshots look like, you can find daily snapshots athttp://archive.routeviews.org/1. Find the routing table entry for Duke University network.(a) What is the IP address of the best next hop from this router to Duke? How


View Full Document

Duke CPS 214 - Problem Set 1

Download Problem Set 1
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view Problem Set 1 and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Problem Set 1 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?