ch7 Flashcards
Evolution of mobile and wireless networks
Get the global picture of technological developments
Slides 2,3,4 of Lecture 7
Evolution:
- Higher bandwidth, lower latencies (Latencies for wireless networks)
- More efficient use of radio spectrum (Mobile networks)
- Higher speeds and better mobility (wireless networks)
- Change in focus, Internet of things (5G)
General Packet Radio Service (GPRS)
Understand basics and concepts (compared to GSM)
8 Channels connected to end device
5 channels are used for transferring data
1 for up, and 4 for down.
GSM uses circuit switching
Hence, inefficient use of radio resources
GPRS uses packet switching
Hence, can handle much more traffic.
GSM/GPRS performance model
Understand ideas of the 2D Markov chain and mapping to
performance metrics
Slide 14, 15
GPRS consists of:
- Voice traffic
- Data traffic
Each voice call uses single channel
Number of channels used by data traffic changes over time (Max 4 channels per data call. Min 1/8 of a channel)
2 areas:
- Shared (priority voice + data)
- data-only
Finite access queue for pending transfers when there is no space for the call.
Markov chain analysis:
2D markov chain. (Number of voice calls (V) and data calls (D) as the 2 dimensions)
V(t), D(t)
Handling of Data calls:
Access queue: Data calls waiting to get access (Can hold up to Qa data calls)
Transfer queue: data calls being processed (Can hold up to Qt, 8 * total capacity)
Newly incoming data is blocked iff. number of data calls is Qt + Qa (when both queues are full)
Transition rates in Markov Chain: - Arrival rates (lambda) Departure rates (Don't have to know in detail)
WLAN: infrastructure vs. ad-hoc mode
Understand the difference
WLAN: \+ cheap hardware \+ easy to deploy - bad security - best effort performance Ad hoc: - connections can go through different devices each time - So the topology may change Infrastructure: - Fixed topology - Specific static access points
Distributed Coordination Function (DCF)
Understand how medium access in WLAN roughly works
Calculates the throughput over in a cell, depending on the number of users in a system
Protocol to avoid collisions between devices
Each station has a timer/counter. The counter goes down. When the counter is 0, you try to send.
Starting values of counters are random. (within range of contention window) (Counter may also be known as back-off counter)
If the medium has been empty for some time, the counter starts decrementing
If the counter was the same, you may still have collisions.
In case of collision: Double the contention window
Drawback:
- If a collision occurs while sending a long packet, you have to resend the entire packet. This may be inefficient
Hidden node problem
Understand what it is
A —- B —- C
A wants to send to B
C wants to send to B
A and C don’t see eachother, so they don’t see traffic on the medium
The signals collide at B, but collision not detected
Basic Access Mode vs. RTS/CTS mode
Understand difference with basic mode, and solution to
hidden node problem
Solution to hidden node problem.
Still checks for empty medium. If this is found,
Basic Access Mode immediately starts sending everything (Naive)
RTS/CTS mode send a small message asking if they can send data (handshake)
When this the receiver allows this, all data is sent to the receiver.
RTS = Request To Send (Sent from sender to receiver)
CTS = Clear To Send (Sent from receiver to sender if the sender should be allowed to send)
RTS/CTS:
- Reduced duration of a collision for long messages
- Lower collision rate
- Much better throughput for larger number of data flows
Back-off counter and retransmission counters
Understand the concepts
r is the value of the counter/timer (initially randomly chosen within the range of the contention window (which depends on the backoff counter))
b is how many times the device was not allowed to send or did not manage to
Bianchi model (with jump chain), fixed-point equation Understand idea of the model (fixed-point iteration was not discussed and you do not need to know that)
Markov chain keeps track of r (retransmission counter) and b (backoff counter) Slide 30 (best to have a look at this part of the lecture) Idea is basically the same as before but now modelled as a Markov chain.
Probability of a collision is dependent on the number of users in the system
Throughput analysis
Understand the line of thinking (no formulas by head)
We use the probability that a user wants to transmit data, but adding up the equilibrium states of all states in the Markov chain where the retransmission counter is 0
We calculate the number of transmission in a given time slot
We use the probability that there is at least one transmission in a given time slot
We use the probability that a transmission is successful (exactly 1 transmission)
Throughput = The expected amount of data transmitted in a “slot time” / length of a “slot time”
Where a slot time is the time between moments in which the medium is empty