1. A Generic Quantitative Approach to the Scheduling of Synchronous Packets in a Shared Uplink Wireless Channel Authors: Authors: Reuven Cohen, Liran Katzir Published: IEEE/ACM Transactions on Networking, August 2007

2. Outline Introduction Introduction Quantitative-Based Frameworks Quantitative-Based Frameworks Computing the Probabilities Computing the Probabilities Proposed Algorithms Proposed Algorithms Simulation Simulation Conclusion Conclusion

3. Introduction In a single channel with single-carrier PHY by the OFDM PHY, describe a centralization scheduling scheme for uplink wireless networks. In a single channel with single-carrier PHY by the OFDM PHY, describe a centralization scheduling scheme for uplink wireless networks. The proposed scheme follows five scheduling considerations. The proposed scheme follows five scheduling considerations.

4. Scheduling Considerations SC1: The specific QoS requirements of each call: the grants should meet the negotiated grant size, grant periodicity, and tolerated grant jitter. SC2: The specific conditions of each uplink channel: basically, if a channel experiences bad SNR, the scheduler will try to delay the grant as much as possible.

5. Scheduling Considerations SC3: The specific application layer loss recovery mechanism employed by each synchronous call codec. The quality of a synchronous call can therefore be improved by assigning a higher drop priority to the more important packets.

6. Scheduling Considerations SC4: SC4: The specific MAC layer loss recovery mechanisms employed by the network, and in particular, whether automatic repeat request (ARQ) is employed. SC5: Adaptive modulation and coding (AMC), along with power control.

7. Quantitative-Based Frameworks Model 1: The base station maintains a profit matrix Φ. Entry Φ[c, t] in this matrix indicates the profit if the first pending packet of call c is transmitted starting from slot t and is correctly received by the base station. Φ[c, t] = the priority of the packet, if success Φ[c, t] = 0, if nonsuccess Satisfied SC1 and SC3. Satisfied SC1 and SC3.

8. Quantitative-Based Frameworks Model 2: A schedule σ is a transmission vector that indicates which packet should start being transmitted in which slot. If σ(t) = c, then at time slot t, the transmission of the current packet of call c should start. The overall profit gained from a schedule σ is Profit(σ)=, where [1…T] is the scheduling interval, μ[c, t] = Φ[c, t] *λ(c, t), and λ(c, t) is the success probability of the transmission for call c in slot t.

9. Quantitative-Based Frameworks Model 3: Entry μ[c, t, m] in this matrix is set to Φ[c, t, m] *λ(c, t, m) Where λ(c, t, m) is the probability that the packet of call c will be transmitted correctly starting from slot t using PHY profile m.

10. Quantitative-Based Frameworks Model 4: The schedule takes into account MAC layer retransmissions. Where R is the maximum number of transmissions.

11. Computing the Probabilities S(n)  {0, 1}, S(n)  {0, 1}, where 0 represents a good channel where 0 represents a good channel and 1 represents a bad channel and 1 represents a bad channel p = Prob[S(n+1)=0 | S(n) = 0] p = Prob[S(n+1)=0 | S(n) = 0] q = Prob[S(n+1)=1 | S(n) = 1] q = Prob[S(n+1)=1 | S(n) = 1] T(n) is the probability that the channel is in error state at time n. T(n) is the probability that the channel is in error state at time n.

12. Computing the Probabilities T(n+1)=q T(n) +(1-p)(1- T(n) ) T(n+1)=q T(n) +(1-p)(1- T(n) ) T(n+1)=(q+p-1) T(n) +(1-p) T(n+1)=(q+p-1) T(n) +(1-p) When T(0)=C When T(0)=C Assuming that a≠1, for n > 0 we get Assuming that a≠1, for n > 0 we get And therefore And therefore

13. Computing the Probabilities C=1 C=1 C=0 C=0 And And

14. Proposed Algorithms

16. Algorithm 1: Bigger Profit First Algorithm 1: Bigger Profit First Algorithm 2: Algorithm 2: Earliest Deadline First Earliest Deadline First Mack sure the schedule is feasible Mack sure the schedule is feasible

17. Simulation

18. Simulation

19. Simulation

20. Simulation

21. Simulation

22. Simulation

23. Simulation

24. Simulation

25. Simulation

26. Simulation

27. Conclusion Presented a generic quantitative-based scheme for scheduling. Presented a generic quantitative-based scheme for scheduling. Select the most important packets for transmission Select the most important packets for transmission Increases the number of synchronous packets that transmitted on time Increases the number of synchronous packets that transmitted on time Decreases the number of packets that are transmitted when the channel is noisy Decreases the number of packets that are transmitted when the channel is noisy