Q and A for Ch 7 13 IS 333 Victor Norman Spring 2016 Circuit switching Q Is circuit switching obsolete If not where is it used p222 A Telephone uses circuit switching still I think or at least virtual circuit switching Circuit switching is still used DSL uses it On my DSL mod ID: 761374
Download Presentation The PPT/PDF document "Q and A for Ch. 7, 13 IS 333 Victor Norm..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Q and A for Ch. 7, 13 IS 333 Victor Norman Spring 2016
Circuit switching? Q: Is circuit switching obsolete? If not, where is it used ? ( p.222) A: Telephone uses circuit switching still (I think), or at least, virtual circuit switching. Circuit switching is still used: DSL uses it! On my DSL modem, the “WAN” side shows the VPI/VCI for my connection: which identifies the virtual path/virtual channel ids used for the circuit. Companies can also buy MPLS service from providers, and this is circuit-switched.
Circuit-switching Q: How does multiplexing for circuit switching work? Specifically , how can multiplexing occur if circuit switching is supposed to guarantee an “isolated” point-to- point communication ? A: A cable can support only a fixed # of channels. The endpoints manage the allocation of the channels. Each channel gets a frequency (FDM) or time slot (TDM). End-to-end channels form a circuit. Delivery guarantees can be made because the circuit is established before any data is sent. See pages 182-188.
Adv/ Disadv of Circuit Switching fixed # of “channels” and only max are assigned – so when data is available on a channel it can be sent without delay. Can offer circuits that get multiple channels – thus, having more bandwidth. If channel is assigned and no data is sent, the bandwidth is unused/wasted. Uses set-up phase, transmission, and tear-down phase.
Shielded vs. Unshielded Q: What is shielding? I don’t understand. A: Shielding is metallic foil wrapped around the pairs of wires. Q : I'm not sure I understand why you would leave a twisted pair unshielded. Can you explain ? A: $ Q: Do you understand the picture on page 116? A: Yes ! Q: Is electrical noise still really a problem in today’s world? A: Yes.
Signal Attenuation Q: What does it mean that “optical cables have less signal attenuation”? A: Attenuation is the degradation of a signal as it flows through a medium – electricity in copper or light in fiber. As a signal degrades the receivers have to get better and better at decoding the signal. Q: Is there any kind of noise when using optical fibers? A: Not that I know of. But, the signal does degrade still, from bouncing off the edges of the glass.
Packet Switching Speed Q: Does the overall speed decrease when multiple transfers from multiple computers are trying to send out packets ? A: Yes: because packets get queued up for processing and forwarding at “intersections” – routers, switches, etc.
Byte Stuffing Q : Can you explain byte stuffing? How does the receiver “decode” it? A: Consider what you learned in CS 108 about putting double quotes in a string. You put \” for ” inside a string. How do you put \ in a string? \\ This is the same as byte stuffing: converting special characters to other characters.
Gigabit Ethernet? Q: If CAT 5 is only rated to handle 100mbps speeds, then what about gigabit ethernet ? A: CAT 5 can actually be used for gigabit ethernet . https:// en.wikipedia.org/wiki/Gigabit_Ethernet
Mesh topology? Q: Why would anyone use a mesh topology network if it's so impractical? Is it more of just a concept ? A: A mesh is still used within a supercomputer or similar architecture.
Comparing Topologies Bus: can easily plug in new machines, as long as they aren’t too close together. Only need one wire. But, if wire is cut, whole network is down. All machines get all transmissions. Only one machine can talk at once. Mesh: very expensive – lots of wire. Lots of interfaces on each machine. But, contend only with the machine on other end. Star: Contend only with one machine on other end. Affords more privacy. Requires a device in the middle to forward packets/signals. Can’t be expanded easily. Can use more cable. Ring: not used much, afaik . Requires 2 interfaces per machine. If any two links go down, a section is isolated.
Calvin’s network topology Q: What kind(s) of topology does Calvin use ? A: Calvin uses a star topology. There is only one main router, and everything else radiates off of that. And, each of those sub-networks is completely switched (not routed). Most companies use some kind of star or web topology – multiple connections for redundancy.
Kinds of MAC addresses Q: Are Unicast, Broadcast and Multicast addresses all types of MAC addresses ? A: Unicast, broadcast, and multicast are all concepts in sending information. But, yes, Ethernet MACs have each type of address. But, I don’t think multicast addressing is used much, if at all, at the Ethernet layer.
MAC addresses Q: Are all MAC addresses the same length? A: Yes: 6 octets (48 bits), written: xx:xx:xx:xx:xx:xx or xx-xx-xx-xx-xx-xx where each x is a hexadecimal digit (representing 4 bits) Note that the broadcast address is all 1s, so is written: ff:ff:ff:ff:ff:ff (12 hex digits, where each digit is 4 bits, all 1s.)
MAC addresses (2) Q: What are the other bits in the “most significant byte” of an IEEE MAC address (besides the multicast and global/local bits), and why are they read this way? Also, why 48 bits/6 bytes? A: I don’t know if the other bits have significance. And, multicasting is rarely used, afaik . I don’t know why they chose 48 bits, but it does make (theoretically) 2 48 addresses, which is a lot.
Ethernet Frame Formats Ethernet contains a header and a payload. The header is just there to help get the data through. Header contains dest MAC address, source MAC address, type of data in the payload (but no payload length?). Data is sent with voltages so certain patterns are reserved to indicate a start of header (SOH), end of transmission (EOT), etc.
Important Point about Ethernet Ethernet was originally a bus topology – plug multiple machines into one single wire. Ethernet protocols still assume it is a bus topology – so sending a packet out on an Ethernet means multiple machines could see it. The protocols do not assume if you send a packet out it goes to one machine at the other end of the wire (star topology).
Old Slides
Router <-> ISP Q: Is the connection between a Router and the ISP a p2p network? A: It is going to be a point-to-point link. The ISP run a fiber/cable to a business and you plug your/their router into it.
Different cable types Q: Why don’t people employ CAT7 cables versus the CAT5 cables that seem so common? Is it cost? A: $
Fiber in America Q: Could you discuss what we are doing (if anything) in America right now to switch over to optical fiber ? A: Fiber is what carries most data on the core of the Internet. Fiber is being pushed out closer and closer to the home. AT&T Uverse gets fiber to the neighborhood (FTTN). Verizon in some places offers Fiber to the Home (FTTH). See Figure 12.8 (page 208).
Coaxial cables Q: Are coaxial cables a type of shielded twisted pair cables? A: See picture in the book. There is one wire down the middle with a wire mess around it.
Copper vs. fiber Q: Which is the most practical/efficient copper wire or fiber optics ? Which is mostly used generally and is preferred, the optical fiber or the copper wire?" A: Copper is much cheaper and much easier to work with – you can splice it and cut it and fix it, etc. You can walk on it. Fiber is more expensive and you have to be careful about kinking it, breaking it, walking on it, etc. But fiber has higher bandwidth, lower signal degradation.
Copper vs. fiber (2) Q: How widely are fiber optic cables used compared to copper wires? It seems to me like most of the time copper wires are used, but I know that fiber optic cables are used but have never heard for what besides going very long distances. A: Fiber is used when higher bandwidth is needed – like between buildings on campus.
Communicating with light Q: How does light carry information across an optical fiber? A: Lasers/LEDs turn the light on and off and the receivers detect it. See 6.16 for an example of encoding.
Q: Can optical fibers end up with mixed up signals, since the reflection alters the speed and energy of the light traveling and eventually received? A: Signals do degrade, just as with copper. But, I think light over fiber degrades much less… There is lots of physics involved.
Network Topologies Q: Can you compare network topologies – their advantages and disadvantages? A: In practical terms for “regular” networks, the star is always used – with some redundant links, often. (For supercomputers, backplane networks, etc. you might use mesh, etc.)
Comparing Topologies Bus: can easily plug in new machines, as long as they aren’t too close together. Only need one wire. But, if wire is cut, whole network is down. All machines get all transmissions. Only one machine can talk at once. Mesh: very expensive – lots of wire. Lots of interfaces on each machine. But, contend only with the machine on other end. Star: Contend only with one machine on other end. Affords more privacy. Requires a device in the middle to forward packets/signals. Can’t be expanded easily. Can use more cable. Ring: not used much, afaik . Requires 2 interfaces per machine. If any two links go down, a section is isolated.
New wiring/energy? Q: What are some other forms of wiring and energy being explored today? A: Most everything seems to be all about wireless – making it faster, traveling farther, or restricting its reach. Also, satellite internet is big, in some areas/applications.
Circuit-switching Q: How does multiplexing for circuit switching work? Specifically , how can multiplexing occur if circuit switching is supposed to guarantee an “isolated” point-to- point communication ? A: A cable can support only a fixed # of channels. The endpoints manage the allocation of the channels. Each channel gets a frequency (FDM) or time slot (TDM). End-to-end channels form a circuit. Delivery guarantees can be made because the circuit is established before any data is sent.
Adv/ Disadv of Circuit Switching fixed # of “channels” and only max are assigned – so when data is available on a channel it can be sent without delay. Can use high-priority circuits that get multiple channels – so more bandwidth, etc. If channel is assigned and no data is sent, the bandwidth is unused/wasted. Uses set-up phase, transmission, and tear-down phase.
Packet Switching Q: What exactly is packet-switching aside from just using packets to send the information? A: Packet switching “chunks” up the data, but doesn’t assign channels to packets. The packets are sent when they get to the front of the queue. So, less bandwidth is wasted, but there are not hard guarantees on delays of packets. (In fact, packets can be dropped if way too many arrive at the same time.)
Statistical Multiplexing Q: What is statistical about “statistical multiplexing”? A: Statistically speaking, if you have a reasonable amount of traffic on a network, your data will get through “pretty quickly.”
Byte and bit stuffing Q : Why would you use bit/byte stuffing rather than the header/footer method? A: You use both. Because a transmission is “couched” in these special patterns (SOH, EOT), you have to “escape” these patterns if they appear in the transmission.
Byte and bit stuffing (2) Q: I am confused with how the byte stuffing works. How is it that by turning SOH into ESC A, EOT into ESC B, and ESC into ESC C making it smaller? It seems like it would be bigger because you are replacing one thing with two now. A: You must have misread the text. It does make it bigger. But, that’s the price you pay for clarity.
Manchester Encoding Chapter 6.16. I’ll explain…