# #Slotted ALOHA **Topic: 5G Random Access Procedure** Bariq Sufi Firmansyah Institut Teknologi Bandung bariqsufif@gmail.com ###### tags: `NTUST-RESEARCH` --- ## Background ::: spoiler Click to view the background knowledge of **Pure ALOHA**, you can skip this if you already understand --- The network is constructed such that each station is coupled to a single passive channel, for example a coaxial cable In normal operation packets are transmitted without error from one station to another and each packet is acknowledged If **two stations** try to transmit at the same time, then the signals interfere and an error is detected by the receiving station and **no acknowledgement** is sent ![](https://i.imgur.com/VR7l1dT.png) After an appropriate ’timeout’, which is at least as long as the maximum two-way propagation time of the cable, the original transmitting stations conclude that a collision has occurred and schedule retransmission of the packets at later times. To avoid repeated collisions of packets, the retransmission times are usually chosen at random by the stations involved #### Throughput Analysis - Let $\tau$ seconds be the transmission time for a packet (propagation time + transmission time) - Let $$Pr[k\ arrivals\ in\ \tau \ seconds]={(\lambda\tau)^k\over k!}{e^{- \lambda\tau}}$$ be the Poisson distribution describing all transmissions over the channel where $\lambda$ = arrival rate in Pkts/Sec . This expression assumes an **infinite population** of infrequent users - Let $S$ be the average number of successful transmissions per packet transmission time $\tau$ - Let $G$ be the average number of attempted packet transmissions per packet transmission time $\tau$ ![](https://i.imgur.com/rREdQOJ.png) *Fig.2 Vulnerable Period for ALOHA* - Figure 2 shows some packet being transmitted. The lower diagram shows the positioning of other packets that just avoid colliding with the original packet. The **vunerable period** can be seen to be of length $2\tau$ - The probability of a successfull transmission is the probability of only one transmissions in the vulnerable period of length $2\tau$. (note that $\lambda={G\over\tau}$) $$Pr[1\ arrival\ in\ 2\tau \ seconds]={(\lambda2\tau)^1\over 1!}{e^{- \lambda2\tau}}={\lambda2\tau}{e^{- \lambda2\tau}}=2Ge^{-2G}$$ - because $S$ is the average number of successful transmissions **per packet transmission time τ seconds,** then $$S={Pr[1\ arrival\ in\ 2\tau \ seconds]\over 2}=Ge^{-2G}$$ ![](https://i.imgur.com/43XiwLM.png) Note that $S$ has a maximum value of ${1\over 2e} ≈ 0.184$ for $G = 0.5$ ::: --- The **Slotted ALOHA** procedure segments the time into slots of a fixed length $\tau$. Every packet transmitted must fit into one of these slots. A packet arriving to be transmitted at any given station must be delayed until the beginning of the next slot. The station can be in either of three states: -- idle (no ready packet to be transmitted) -- backlogged (the packet transmitted in the previous slot was failed, and waiting for the next slot to retransmit) -- transmitting (transmit new packet if the station was in idle state, or retransmit the failed packet if the station was in backlogged state) #### Throughput Analysis An expression for $S$ in terms of $G$ is required. This will be derived for an **infinite** population and a **finite** population of stations. #### Infinite population case ![](https://i.imgur.com/zWVp6O5.png) The vulnerable period for Slotted ALOHA is reduced from $2\tau$ to $\tau$ seconds. $$S=Pr[1\ arrival\ in\ \tau \ seconds]={(\lambda\tau)^1\over 1!}{e^{- \lambda\tau}}={\lambda\tau}{e^{- \lambda\tau}}=Ge^{-G}$$ ![](https://i.imgur.com/xBTwt5h.png) *Fig.5: $S$ vs $G$ for ALOHA and Slotted ALOHA* Compare this with the Pure Aloha case: $S = Ge^{−2G}$. The maximum throughput for Slotted ALOHA is ${1\over e} ≈ 0.368$ and the maximum occurs for $G = 1$. #### Collision Probability Analysis $$Pr[collision]=Pr[>1\ arrival\ in\ \tau \ seconds]$$ $$=1-Pr[0\ arrival\ in\ \tau \ seconds]-Pr[1\ arrival\ in\ \tau \ seconds]$$ $$=1-{e^{-G}}-Ge^{-G}$$ #### Finite population case In the above we have assumed the number of users is infinite. Here we will revise the model for a finite population and determine under what conditions the infinite population model gives a reasonable approximation. Consider a network with $M$ independent stations. The input from each station is modelled as a sequence of *independent Bernoulli Trials* - Let $S_i$ be the probability of station $i$ successfully transmitting a packet in any slot. $1 −S_i$ is the probability of not successfully transmitting a packet. - Let $G_i$ and $1−G_i$ be the probabilities for attempting and not attempting transmissions respectively The probability that station $j$ has a successful transmission is the probability that station $j$ is the only station to attempt a transmission in a particular slot. Therefore: $$S_j=G_j\prod_{i=1,\ i\ne j}^M (1-G_i) $$ Assuming all stations share the load equally then $S = MS_j$ and $G = MG_j$, therefore: $$S = G(1 − {G\over M})^{M−1}$$ Note that: $$\lim_{n\to\infty}[1+{X\over n}]^n=e^X$$ as $M\to\infty$, $S=Ge^{-G}$ (as before) For $M$ stations the maximum throughput $S_{max}$ is calculated by differentiating and equating to zero: $$0 = (1−{G\over M})− G({M− 1\over M})$$ which results in $G = 1$ (as for the infinite case) $$S_{max} = (1 − {1\over M})^{M−1}$$ Collision Probability is the probability that at least two stations attempt transmission in a particular slot $$Pr[collision]=1-\prod_{i=1}^M (1-G_i)-G_j\prod_{i=1,\ i\ne j}^M (1-G_i)$$ $$=1-(1 − {G\over M})^{M}-G(1 − {G\over M})^{M−1}$$ ## System Model #### Scenario 1: Infinite-Station | I/P Parameters | Value | | ----------------- | ------| |$M$: # of STAs | $M\to\infty$ | |$T$: # of access cycles| $10^5$ | |$\lambda$: packet rate| $[0:0.2:18]$ | #### Variable Data structure | Variable | Meaning | Data Type| range | | -------- | -------- | -------- |-------- | | $R$ | random number generated from poisson distribution with mean $\lambda$, represents the number of packets being sent in that timeslot | Integer | $0$~$\infty$ | | $t$ | Access Cycle/ Timeslot ID | Integer | $1$~$T$ | |$success\_pckt$ | the number of successful transmission | Integer |$0$ ~ $\infty$ | |$collision\_pckt$ | the number of failed transmission | Integer |$0$ ~ $\infty$ | #### Output Data structure | O/P | Meaning | Data Type| range | | -------- | -------- | -------- |-------- | | $throughput$ | $success\_pckt\over T$ for each $\lambda$| Vector of Float | $vector\ size=(1, length(\lambda)$, each element has range $0$~$1$ | | $collision\_prob$ | $collision\_pckt\over T$ for each $\lambda$| Vector of Float | $vector\ size=(1, length(\lambda)$, each element has range $0$~$1$ | #### Flow Chart ```flow st=>start: Start e=>end: End op=>operation: lambda=0 op1=>operation: success_pckt=0, collision_pckt=0 op2=>operation: t=1 op3=>operation: Generate random number from poisson distribution with mean=lambda, assign to R op3_5=>operation: Collision detected, increment collision_pckt op4=>operation: packet successfully transmitted, increment success_pckt op5=>operation: increment t op6=>operation: calculate throughput and collision_prob for this lambda op7=>operation: increment lambda by 0.2 op8=>operation: plot Y:throughput, X:lambda cond=>condition: is R=1? cond1=>condition: is R>1? cond2=>condition: is t=T? cond3=>condition: is lambda=18? st->op->op1->op2->op3->cond cond(yes)->op4->cond2 cond(no)->cond1 cond1(yes)->op3_5->cond2 cond1(no)->cond2 cond2(yes)->op6->cond3 cond2(no)->op5->op3 cond3(no)->op7->op1 cond3(yes)->op8->e ``` #### Scenario 2: Finite-Station | I/P Parameters | Value | | ----------------- | ------| |$M$: # of STAs | $10, 50$ | |$T$: # of access cycles| $10^5$ | |$\lambda$: packet rate| $[0:0.2:8]$ | #### Variable Data structure | Variable | Meaning | Data Type| range | | -------- | -------- | -------- |-------- | | $G_i$ | probability of station $i$ attempting transmission | Float | 0~1 | | $P_r$ | Random number generated from uniform distribution | Float | 0~1 | | $m$ | Station ID | Integer | 1~M | | $t$ | Access Cycle/ Timeslot ID | Integer | 1~T | |$pcktThisSlot$ | the number of packets transmitted in slot $t$ | Integer |$0$ ~ $M$ | |$success\_pckt$ | the number of successful transmission | Integer |$0$ ~ $\infty$ | |$collision\_pckt$ | the number of failed transmission | Integer |$0$ ~ $\infty$ | #### Output Data structure | O/P | Meaning | Data Type| range | | -------- | -------- | -------- |-------- | | $throughput$ | $success\_pckt\over T$ for each $\lambda$| Vector of Float | $vector\ size=(1, length(\lambda)$, each element has range $0$~$1$ | | $collision\_prob$ | $collision\_pckt\over T$ for each $\lambda$| Vector of Float | $vector\ size=(1, length(\lambda)$, each element has range $0$~$1$ | #### Timing Diagram ![](https://i.imgur.com/2qugSJD.jpg) *Fig. 6 Timing Diagram* - During each timeslot each station will be transmitting packet with probability $G_i={G\over M}$ - The successful transmission happens when there is only one station transmitting the packet - The input from each station is modelled as a sequence of *independent Bernoulli Trials* - There is no backoff time, the collided packet will be retransmitted again in the future timeslot, based on probability calculation of each station and *bernoulli trial*. (this is called $thinking\ state$) #### Flow Chart ```flow st=>start: Start e=>end: End op=>operation: lambda=0 op1=>operation: success_pckt=0, collision_pckt=0 op2=>operation: t=1 op2_5=>operation: pcktThisSlot=0 op3=>operation: m=1 op4=>operation: Generate random number, assign to Pr op5=>operation: STA is transmitting packet op6=>operation: STA is not transmitting packet op7=>operation: increment pcktThisSlot op8=>operation: increment m op9=>operation: Collision detected, increment collision_pckt op10=>operation: packet successfully transmitted, increment success_pckt op11=>operation: increment t op12=>operation: calculate throughput and collision_prob for this lambda op13=>operation: plot Y:throughput, X:lambda op14=>operation: increment lambda by 0.2 cond=>condition: is Pr<Gi? cond2=>condition: is m=M? cond3=>condition: is pcktThisSlot=1? cond4=>condition: is pcktThisSlot>1? cond5=>condition: is t=T? cond6=>condition: is lambda=8? st->op->op1->op2->op2_5->op3->op4->cond cond(yes)->op5->op7->cond2 cond(no)->op6->cond2 cond2(no)->op8->op4 cond2(yes)->cond3 cond3(yes)->op10->cond5 cond3(no)->cond4 cond4(yes)->op9->cond5 cond4(no)->cond5 cond5(no)->op11->op2_5 cond5(yes)->op12->cond6 cond6(yes)->op13->e cond6(no)->op14->op1 ``` ## Numerical Results #### Scenario 1: Infinite-Station ![](https://i.imgur.com/En8AMPj.png) *Fig. 7 Average Throughput of infinite-station slotted ALOHA* ![](https://i.imgur.com/e3vcxYm.png) *Fig. 8 Collision probability of infinite-station slotted ALOHA* Figures 7, and 8 illustrated simulation results and the analytical result of the throughput and the collision probability, of infinite-station slotted ALOHA obtained from Equations $$S=Ge^{-G}$$$$Pr[Collision]=1-{e^{-G}}-Ge^{-G}$$ for $G=0$ to $18$ and $T=10^5$. #### Scenario 2: Finite-Station ![](https://i.imgur.com/BQdroks.png) *Fig. 9 Average Throughput of finite-station slotted ALOHA* ![](https://i.imgur.com/kp7YO9N.png) *Fig. 10 Collision Probability of finite-station slotted ALOHA* Figures 9, and 10 illustrated simulation results and the analytical result of the throughput and the collision probability, of finite-station slotted ALOHA for $M = 10$ and $50$, for $G$ = $0$ to $8$ and $T=10^5$. from figure 9 we know that lower number of stations will make the maximum throughput higher, but it will decay to zero faster. And also from figure 10 we know that lower number of stations will make the collision probability approaches 1 faster.