Networks and Grids Technology and Theory Solutions Manual By Thomas G. Robertazzi

A Tour through Networking and Grids

1. There is no need to dig trenches for cables (though tower construction must be taken into account). 2. Fiber optics has the largest information carrying capacity (easily into the Terabits per second).
3. The electronic bottleneck refers to the fact that with current fiber speeds nodal electronics is often slower than fiber speeds so that the information throughput bottleneck is in the nodes.
4. An advantage of geosynchronous satellites is that they are in one spot in the sky all the time so a fixed directional antenna on earth can be used. An advantage of Low Earth Orbit Satellites (LEOS) is that because of their much lower orbits than those of geosynchronous satellites, they have much smaller channel propagation delay.
5. The use of infrared light is technologically possible and systems have been built, but the use of infrared is not that popular.
6. Ad hoc transmission is usually over much smaller distances than satellite transmission (as small as tens of feet for ad hoc networks compared to hundreds or thousands of kilometers for satellite transmission).
7. Wireless sensor networks may be used for such applications as environmental monitoring, interconnecting engine components and connecting computer components (i.e. Bluetooth technology).
8. As a packet moves down a layered protocol stack (is transmitted) each layer may append information to the packet so that it is at its largest at the bottom of the stack (physical layer). At the receiver, as the packet moves up the stack, information may be extracted at each layer so the packet is at its smallest at the top (application) layer.
9. Packets carried on a virtual circuit arrive in the same order they were originally transmitted as they follow a single sequential path.
10. A communication between peer network layer entities moves down the stack at the transmitter to the physical layer, across the network to the destination and up the destination stack to the appropriate network layer entity.
11. One function of the data link layer is to manage communication over a single link between a pair of nodes. Also, encryption is sometimes done at the data link layer.
12. The network layer manages communication over a multiplicity of links and nodes whereas the data link manages communication over a single link between a pair of nodes. \
13. The transport layer is responsible for providing end to end communication over possibly unreliable sub-networks.
14. Thruput decreases in Ethernet under heavy loads because of the time wasted by collisions (in the CSMA versions of Ethernet).
15. With longer frames, as opposed to shorter frames, a bigger fraction of time is spent in useful transmission, as opposed to time spent in seizing the channel and collisions, so that utilization is higher.
16. A total of 3 × 3 = 9 symbols can be sent at once which is equivalent to 3 bits (23 = 8).
17. The use of unshielded twisted pair is a good thing because it is lightweight and relatively inexpensive.
18. If node A does not receive a Clear to Send message from node B it might try again to send a Request to Send or move to a different location and try again to send a RTS.
19. The use of a Clear to Send solves the hidden station problem in distributed wireless networks (some node over the radio horizon may try to transmit to node B at the same time as node A, causing a collision).
20. The amount of radio spectrum available for specific purposes is more limited than the equivalent bandwidth available on wired fiber.
21. If a network had to be constructed and taken down frequently, one would be better off using some combination of 802.11 WiFi and 802.15 Bluetooth technology as it is wireless, unlike standard Ethernet versions.
22. Using two 64 kps channels to carry 80 kps means that 128–80 or 48 kps of capacity is wasted.
23. In an ATM NNI connection 16 header bits are used for virtual channel identification (VCI) so that there are 216 or approximately 64 thousand virtual channels (per virtual path).
24. ATM assumes that data is transported over low error rate fiber optics so only the header is protected by error coding. If errors in the data field cannot be tolerated, then a higher level protocol can provide error coding.
25. The VPI and VCI fields in the ATM header, like the rest of the header, are protected by error coding which makes misrouting extremely rare.
26. The use of ATM technology leads to service class independent switches. If one needed different switches for each class of traffic one would have an intractable traffic prediction problem. Since it is almost impossible to accurately predict demand by service class into the future one might install too few or too many of each class of switch, leading to problems in either providing inadequate capacity or over-investing in network facilities.
27. One application where it is critical to deliver data quickly is the transmission of stock prices. Also real time applications such as voice or video require packets be delivered within certain time limits (bounds).
28. A contract for an ATM session is difficult to define because there are so many quality of service parameters that could be used. Thus there are many possible contracts with many possible pricing options.
29. A T1 line has a data rate of approximately 1.5 Mbps. A SONET OC-3 channel has a data rate of approximately 155Mbps.
30. Each byte entry in a SONET OC-1 frame has an equivalent data rate of 64 kbps. Thus to carry 1 Mbps one needs 1 Mbps/64 kbps or 16 entries of the 87 × 9 = 783 entries in the frame. Note that if the question is how many frames are needed to carry 1 Mbyte, not 1 Mbps, the answer is 1,000,000/783 or 1278 frames (neglecting path overhead).
31. A SONET add/drop multiplexer is a device that allows signals to be tapped off of and inserted onto a fiber.
32. In SONET protection fibers are backup fibers that can be brought into use if the service fibers that normally carry traffic do not function.
33. The line protocol layer in SONET is most similar to the data link layer in the OSI protocol stack. 34. More virtual paths are allowed on an ATM NNI link than a UNI link as the NNI link is like a trunk that is likely to carry much more traffic than a UNI network access link.
35. Transmitting in one byte of a SONET table has an effective data rate of 64 kbps. This is the standard data rate for an uncompressed digital telephone channel.
36. The approximate data rate of OC-3072 is 3072 times 51 Mbps (the OC-1 rate) or approximately 160 Gbps.
37. Each OC-192 channel has a data rate of approximately 10 Gbps so approximately 76 WDM OC-192 channels are needed to carry 760 Gbps.
38. The technology of WDM allows fiber already put in the ground (at great expense) to be upgraded through a simple replacement of transmitters and receivers to carry a much larger amount of traffic. 39. The basic idea behind grid computing is to allow a user to access a large distributed network of powerful computers and storage devices from anywhere on earth to carry out substantial computations of a scientific, economic or other nature. It is actually an old concept that predates the use of the word “grid”.
40. To partake in grid computing a computer installation needs to be more open to use to outside users and entities, something not that well tolerated under old policies.
41. In the past when the (computer) resource owner was the key person, operations were optimized for high throughput. Making the user the key person necessitates a new set of requirements.

Link Download & Reading online: Free Download