Problem 1 (Matt Dearing) Digital radio is ¡°Over-the-air broadcast or cable radio that uses a compressed digital format for transmission. Digital radio effectively increases the capacity of a transmission channel. It also can accommodate data as well as audio transmission¡±1. An example of a big talk show host switching to Sirius is Howard Stern. The advantages are that he can be heard anywhere by anyone who wants to here him by subscribing to Sirius where as before his show would only be picked up by certain affiliates. His show will now also be unedited. With digital radio you also don't have to worry about losing your radio station as you drive around because the signal can be sent much further than regular radio. Also the sound quality of digital radio is near CD quality. An article for digital radio that I found was at http://www.cnn.com/2004/TECH/ptech/01/07/digital.radio.ap/ Digital TV is ¡°Over-the-air or cable television that uses a compressed digital format for transmission. Digital television supports HDTV, or high-definition television, and/or additional channels that can be transmitted in the same bandwidth as a single channel analog signal.¡±2 February 17 2009 is the date that Congress has set for television to switch from analog to digital. By switching to digital television a large number of radio frequencies will be returned to the government that are currently used to broadcast the analog stations. These stations can be used for police and fireman and can also be sold to generate revenue. The radio stations can also be used for cheap wireless broadband, wireless television or other technologies. The only problem is that many television sets will become useless because they will not be able to interpret the digital signal. The government still has to work out what it will do about these estimated 20 million ¡°obsolete¡± television sets. The article for digital TV that I used was http://money.cnn.com/2006/01/04/technology/pluggedin_digitaltv/ Some recent technological developments that I have been exposed to as a student at Purdue are P2P networks and Tortoise SVN. I know that P2P networks have been around for a while but my connection at home was not fast enough to really utilize them. For SVN I am not sure how long it has been around but I did not know about it until last year. It is a way to create a repository for documents that will save all previous revisions and merge new additions together. It was very useful for my software engineering class as we had 7 people on my team working on the same project. I wish I had known about it earlier in my Purdue career because there were many times when I changed a file and wished I could have gone back to an older version but I did not know about SVN at the time. Problem 2 a) TDR (Chanu suh) TIME DOMAIN REFLECTOMETER (TDR) MEASUREMENTS Experiments carried: 1. TDR and 3 cables were provided 2. Connect one end of the cable to TDR¡¯s MAIN jack. 3. Record length of three cables 4. Connect the other end of the cable to TDR¡¯s LOOPBACK jack 5. Press mode of TDR to display wiremap 6. Record wiremap of three cables provided Measurements recorded: Cable label Length Wiremap E 6 ft 21----78 (fault) B 4 ft 12345678 C 5 ft 12345678 Commentary/interpretation: The wiremap function tests twisted- pair cabling for proper wiring. Upper line of fixed digits shows the detected wires at the MICROSCANNER jack, lower line of digits indicates the actual wiring. For cables B and C, MICROSCANNER did not detect any fault which means that they are in working condition. For cable E, the fault indicator is displayed and 21---- blinked. This indicates that wire 2 and 1 is switched and 3, 4, 5, and 6 is not connected in the cable, but wires 7 and 8 in the cable work. b) PING (Chanu Suh) # Packet Size Min Avg Max 1024 0.156 0.159 0.174 2048 0.198 0.206 0.231 4096 0.244 0.250 0.257 Do you observe a trend? As the packet sizes gets larger, the ping response time increases. (TA comment: Transmission delay comes into play. ) PLOT OMMITED Estimated distance (in miles) From Purdue to Bloomington: 115.52 miles From Purdue to San Diego: 1750 miles From Purdue to Oxford: 3901 miles Minimum latency needed for a bit to travel From Purdue to Bloomington: 115.52 miles=185911.42 meters 185911.42 m / 299792458 m/s= 6.201e-4 s= 0.6201 ms Ping time: 4.002ms From Purdue to San Diego: 1750 miles= 2816352 meters 2816352 m / 299792458 m/s= .00939 s= 9.394 ms Ping time: 63.012 ms From Purdue to Oxford: 3901 miles = 6278050.944 meters 6278050.944 m / 299792458 m/s= .02094s= 20.94 ms Ping time: 97.617 ms How much do the calculated numbers differ from the measured numbers obtained using ping? The estimated times are faster than ping time at least 4 times more. In your opinion, what are the top 3 principal factors that contribute to the discrepancy? In my opinion, factors that contribute to the discrepancy include internal networking traffic, external internet traffic, and router passing overhead. (TA comment: In a network point of view, the delays introduced by network devices are the most critical reason; transmission delays, queuing delays at input/output queues of switches/routers) c) TRACEROUTE (Chanu Suh) #of IP hops sslab12: 1 www.purdue.edu: 6 www.indiana.edu: 8 sdsc.edu: 16 ox.ac.uk: 18 Irregular behavior Irregular behavior is found in www.indiana.edu and www.ox.ac.uk where they both do not terminate normally. For www.indiana.edu, traceroute prints out warning, traceroute: Warning: www.indiana.edu has multiple addresses; using 129.79.78.8 traceroute to viator.ucs.indiana.edu (129.79.78.8), 30 hops max, 46 byte packets It looks like www.indiana.edu has multiple addresses and the one forwarded to is disabling traceroute to prevent outsiders from obtaining detailed information about their architecture. For www.ox.ac.uk, it does not terminate and keeps searching for the destination machine even when it reached its destination machine and keeps finding another destination machine that will respond. (Jeremy Porath) You ask if I am able to guess to whom the intermediate routers on the path to www.sdsc.edu belong. The names of these routers, along with all the other output, is stored in traceroute.file. Looking at this data, I find a pattern: the routers begin at purdue.edu, have one stop at indiana.gigapop.net, then pass through abilene.ucaid.edu on their way to cenic.net, before finally coming to sdsc.edu. Looking on-line, I find that indiana.gigapop.net is the Network Operations Center for Indiana GigaPOP, which is owned jointly by Purdue and IU. Abilene Networks, the next stop, is a partnership of Internet2, Qwest Communications, Nortel Networks, Juniper Networks, and Indiana University, according to their website. CENIC?Corporation for Education in Network Initiatives in California?lists the people on its board as belonging to many of the universities in California, as well as members of the Office of Education of some of the Californian counties. So most of these stops appear to be owned, in some part, by universities in joint with private corporations. The route from SDSC back to my computer was not exactly the same, point-bypoint, but for the most part it was very, very similar. The companies that set down the routers between us, such as Abilene, appear to have set them down in pairs, and named them accordingly. For example, while the traceroute from my lab machine went from losang-snvang.abilene.ucaid.edu to CENIC routers, the packet from SDSC went from cenic.net routers to the converse of that, snvang-losang.abilene.ucaid.edu. This was true of not only all the Abilene Network routers, but also of the CENIC routers and the GigaPOP router. On the way to SDSC we hit riv-hpr--lax-hpr-10ge.cenic.net, while on the way back we passed through lax-hpr--riv-hpr-10ge.cenic.net; these are simply the converses of each other. And the GigaPOP router we passed through on the way to SDSC was abilene-ul.indiana.gigapop.net, while on the way back to Purdue it instead was ul-abilene.indiana.gigapop.net. The routers used inside SDSC and Purdue themselves were different?we did not pass through piranha.sdsc.edu on the way from the lab machine to SDSC, for instance?but that?s to be expected. Universities are likely to have much, much more internal traffic than cross-country routers are, so the ?best? route inside each university is likely to change more often than a cross-country path. Also, if the point of entry was different from the point of exit, than it seems reasonable that the internal path might be slightly different. For the two other sites, I chose ElCat in Kyrgyzstan, and NetTools in Latvia, and the results were comparable. The routers were not all exactly the same, but they passed through the same systems: Bishkek.gldn.net to Moscow.gldn.net to stk3.alter.net and lattelekom.lv. There were some discrepancies, though?from ElCat to NetTools went through customer.alter.net, whereas NetTools to ElCat went through another stk3.alter.net. It is possible that some of the discrepancy arose from the fact that they were using different IP addresses for each other, but I don?t think that would matter much: ElCat started at 212.42.96.24, whereas NetTools ended at 212.42.96.2; and NetTools began at 80.232.169.253, while ElCat ended at 80.232.169.191. Such a minor difference in addresses, though, shouldn?t really amount to much. Problem 3 (a) (Jeremy Porath) In this analogy, a segment of road is a ?memory storage device,? and the cars on the road are bits in a packet (or multiple packets). It should be obvious why it is impossible for bits to stop along a wire when a router is busy and its buffer full: unlike a car, a ?bit? is not an entity capable of self-propulsion. A car is capable of stopping itself; but the router itself, or the wire, or something else entirely, would have to be responsible for stopping the bits. Furthermore, a ?bit? is not as easily stopped as a car. Depending on the medium of transportation, a bit can be represented as an electron, a photon, or an electromagnetic wave. I am not sure whether it is even physically possible to stop some of these. But even assuming that it were possible, the mechanisms needed to do so, and then to put them in motion again, would be astronomically expensive to put on every router and would undoubtedly add a great deal of overhead, increasing delay. (b) (TA) The speed of an electron in copper wires depends on the types of wire. Most of you might find that it's roughly 2/3 of SOL. The latency introduced by a medium is called 'propagation delay'. The propagation delay introduced by a copper wire would be roughly L/(0.66 * SOL), where L is the length of the wire; on the other hand, that of a fiber is L/SOL. Assuming X(bps) is very large so that the transmission delay becomes nearly as small as the propagation delay, the gain of replacing copper wires with fibers would have some value near 1.5(depending on X, note that the overall delay is 'propagation delay' + 'transmission delay'(n/X, where n is the number of bits we transfer)). However, if we send n bits consecutively, the propagation delay is only introduced at the first bit(electron) while other bits(electrons) will come to the destination back-to-back. In this sense, the propagation delay is not a big portion of the overall delay; not worth to replace copper wires with fibers.