Flashcards in Quiz 2 Content Deck (45):
What is the difference between 1-persistent CSMA and non-persistent CSMA?
- 1-persistent will continuously check a channel until it becomes idle and then sends with probability 1
- non persistent checks a channel and if it is not free will wait (delay) a random amount of time before repeating algorithm
Briefly contrast classic Ethernet and 802.11 on the issue of acknowledgments.
Classic Ethernet - if a collision doesn't occur assumes success, no acks
802.11 - needs acks (because of the wireless environment with all of its challenges)
Other than performance, what other benefit does switched Ethernet provide?
What is done in 802.11n for larger performance increases over a/b/g variants?
- reduce overhead (due to larger frames)
- multiple antennas
What is the fundamental mode of operation for the network layer?
Store-and-forward packet switching
For a Go-back-n implementation, circular buffer space for a total of 12 frames is allocated. What is the maximum size of the sliding window?
12 - 1 = 11
Identify two key differences between PAR and Go-back-N data-link protocols
- PAR is unidirectional and Go-back-N is bidirectional
- PAR sends packets one at time, Go-back-N sends multiple
- PAR has explicit acks (consumes more bandwitch) vs Go-back-N piggy-backed acks (more efficient)
What is the key contrast between MANs and LANs in terms of medium access?
MANs: point-to-point (longer distances with larger amounts of traffic)
LANs: broadcast/ multiple-access
For pure ALOHA, slotted ALOHA, and 1-persistent CSMA, briefly describe similarities/ differences in the throughput S as the offered load G increases/
- they all increase from G = 0
- (in 1-persistent) carrier-sensing means higher max vs ALOHA
- for high G they all perform poorly
How are reliability challenges stemming from interference and other issues addressed in wireless 802.11 LAN?
- dynamically adjust rate
- fragmentation of a larger frame
Why is the maximum throughput of slotted ALOHA higher than pure ALOHA?
The window of vulnerability for collisions is twice as large in pure ALOHA than in slotted ALOHA. Therefore, at higher offered loads, collisions are more likely in pure ALOHA. (So slotted ALOHA has fewer collisions, hence higher throughput)
Why is the max. throughoutput of 1 persistent CSMA higher than slotted ALOHA?
CSMA has carrier sensing, whereas slotted ALOHA does not. Therefore, CSMA has fewer collisions because it checks forin-progress transmissions before starting a new transmission, hence higher throughput than slotted ALOHA for higher offered load. (Not all collisions are avoided, however)
In the context of computer networks as discussed in this course, what does a Poisson distributionwith mean G represent? (i.e. what is it used to describe?)
It represents the number of events (e.g., frame transmission attemts from different sources) within a specified interval of time (e.g., duration of a frame or length of a time slot)
Describe the bases for the analysis of efficienct in classic Ethernet, i.i., just identify the three factors that are considered (no math is required)
- duration of a conteution slot (normally smaller than frame duration)
- average number of conteution slots (due to collisions) before successful frame xmit.
length of transmitted frame
Compare classis Ethernet and 802.11 on the issue of collisions
Ethernet allows collisions and handles then afterwared with binary exponential back off to reduce the re-occurence of collision between competing sources.
802.11 seeks to avoid collisions (can't entirely eliminate them) with randow backoff before initiating frame transmission
Just identify two ways that 802.11 handles wireless reliability challenges arising from interference and other aspects.
- Auto adjustment of transmission rate to reduce errors when signals are weak or in presence of interference
- Divide frames into smaller fragments so that error occurence affects one fragment and not the entire frame (retransmit only the fragment)
How does 802.11 help to ensure that senders are not likely to wait indefinitely for acknoqledgements?
Between frames, there are 5 defined intervals for different types of prioritized traffic. Trnasmission of ack frames can be initiated in one of these earlier intervals to have priority.
What is the purpose of the network allocation vector in 802.11?
It indicates expected duration of actual use of the shared wireless channel as an additional virual means of "channel sensing" beyond physical sensing. It further aids in collision avoidance by delaying senders from using the channel until presumed idle, ever when physical sensing is less reliable (e.g., interference)
What is the fundamental mode of operation for the network layer?
store-and-forward pack switching (with routers)
Explain the end-to-end argument
Hosts connected to overall network must already implement flow control and order-enforcement mechanisms. Avoid complexity from introducting redundant features within the network (i.i., routers) for the same purpose.
Deciding on the overall path for packet/ frame through network from source to destination through multiple hops/ transmissions
Local decision made within a router to choose appropriate outgoing link for a given packet/ frame as part of overall path
Identify congestion-control techniques for the network layer
provisioning, traffic-aware routing, admission control, traffic throttling, load shedding (as last resort), random early detection (more selective load shedding)
What is the purpose of the TTL field in tn IP packet?
Time-to-Live --> originally a countdown in seconds, now a countdown in hops during routing.
Used to prevent a packed from circulating indefinitely through a network in case of errors in routing tables or other problems
What additional field is included in the frame for a go-back-n sliding window protocol? Why? (in terms of a key difference betweengo-back-n and PAR...)
Frame includes field for piggyback acknowledgement (as seq. number). Go-back-n differs from PAR in that it is a bidirectional protocol for data frames, so data traffic in reverse direction carries with it acknowledgement information related to prior forward traffic.
In a go-back-n sliding-window protocol implementation, write an if statement that shows how the limit on the window size is enforced. Assume the existence of functions that enable or disable the network layer from providing a packet.
if(num_buffered_frames < MAX_SEQ)
enable_network_layer ( ) ;
disable_network_layer ( ) ;
PAR stands for
positive acknowledgment with retransmission
What is pure ALOHA?
lets user transmit whenever it has data to be sent
S=G(e^-2G), S=throuput per frame time, G=attempts per packet time
Whenever two frames try to occupy the channel at the same time, collisions happen and frames are garbled, the sender waits a random amount of time and resends
What is the Poisson distribution used for?
in pure aloha - the probability that k frames are generated during a given frame time, in which G frames are expected, is given by the Poisson distribution
What is slotted ALOHA?
Divide time into discrete intervals called slots, each interval corresponds to one frame/ Stations required to wait for the befinning of the next slot before transmitting
In ALOHA, after each station has sent its frame to the central computer, ...? what happens
each computer rebradcasts the frame to all of the stations. A sending station can thus listen for the broadcast from the hub to see if its frame has gotten through
Briefly describe CSMA with collision detection
Used to prevent collisions from simultaneous transmissions
Worst case a station cannot be sure that it has seized the channel until it has transmitted for 2t, where t = time for signal to propogate between the two farthest stations
We can think of CSMA/ CD contention as a slotted ALOHA system with a slot width of 2t
What is the classic version of Ethernet
uses 1-persistent CSMA/CD with binary exponential backoff on a shared cable :
1st time collision, both wait for either 0 or 1 slot time, 2nd collision, picks either of 0,1,2,3 at random and wait for that number of slot times. nth collision, pick from 0 to (2^n-1)
How does switched ethernet work?
Switches only output frames to the ports for which those frames are destined (indicated by ethernet address)
Each port has its own independent collision domain
How does the switch improve performance over a hub (used by classic ethernet)?
1. no collisions, capacity is used more efficiently
2. frames can be sent simultaneously by different stations
What are the two modes for wireless LANs 802.11?
Infrastructure mode: each client is associated with an AP (access point) that is in turn connected to the other network
Ad hoc network: collecting of computers that are associated so that they can directly send frames to each other
What are the two unique features that 1 Gigabit Ethernet has?
Carrier extension: tells the hardware to add its own padding after the normal frame extend the frame to 512 bytes
Frame bursting: allows a sender to transmit a concatenated sequence of multiple frames in a single transmission
What is the Rate adaptation?
If the wireless signal is weak, a low rate can be used. If the signal is clear, the highest rate can be used.
What are the different physical layer protocols? (802.11b, 802.11a, 802.11g, 802.11n)
802.11b: spread-spectrum method, supports rates of 1, 2, 5.5, and 11Mbps
802.11a: Based on OFDM, which used the spectrum efficiently and resists wireless signal degradations such as multipath. Bits are sent over 52 subcarriers in parallel, up to 54Mbps, but only 1/7 of the range 802.11b
802.11g: Based on OFDM, offers the same rates as 802.11a plus compatibility with 802.11b devices
802.11n: Reduced overheads, used multiple antennas
What is the MAC sublayer protocol?
Radios are nearly always half duplex, recieved signals are too weak, collisions detection is not feasible.
How do you use CSMA/CD
Starts wtih random backoff of 0-15 slots (OFDM), pausing countdown when other frames are sent. Send the frame when countdown reaches 0, if it doesn't get acknowledgement back, continuing with exponential backoff as in Ethernet until successfully sent or maximum num. of retrains. has been reached/
What is a NAV for?
Each frame carries a NAV field that says how the sequence of which this frame is part will take to complete
The 802.11 standard pays attention to the issue of...
power management so that clients need not waste power when they have neither information to send nor recieve
What are different power managements?
Periodic AP beacon frames for client wake-up:
- beacons are periodic broadcasts by the AP, the frames advertise the presence of the AP to clients and identifier to the AP, the time, how long until the next beacon, and security settings
AP buffering until client transmission:
- Client can doze and the AP will buffer traffic intended for it
- AP buffers frames and sends them to a client just after the client sends frames to the AP. The client can then go to sleep until it has more traffic to send/ recieve