**2.3.2 The QoS requirements for jitter**

In IEEE802.16e broadband wireless access networks, acrucial component of delay is the buffered packet delay between BS and MSS. Due to varying delays in transmission, the delays of scheduling from packet to packet may cause buffered packet delay. This phenomenon is called jitter (Wu & Chen, 2004).

As shown in Fig. 5, we denote *Packeti* as the *ith* packet of certain connections, with the QoS requirement of delay having 7 time slots and 2 jitters. Assume *Packeti*-1 was scheduled in the first time slot, and the delay of *Packeti*-1 is 0. *Packeti* may schedule into the time slots of the 2nd time slot to the 8th time slot if we only consider the delay constraint of the QoS requirement. However, it is more realistic to consider the jitter constraint of the QoS requirement. Because the delay of *Packeti*-1 and *Packeti* cause jitter, we need to consider the delay of *Packeti* to satisfy the jitter constraint. Assume we schedule *Packeti* in the 5th time slot, the delay of *Packeti* is 3 and the jitter will also 3, and this violates the jitter constraint. Thus, under the jitter

A QoS Guaranteed Energy-Efficient Scheduling for IEEE 802.16e 41

requirements first is to not violate their QoS requirements, as we mentioned previously. The second part of our algorithm is composed of three steps we described in the Section 2.3.1. In addition to these, the scheduler examines the difference in the delay between the present packet and the previous packet when scheduling each packet in each step. The difference in the delay between the present packet and the previous packet can be viewed as jitter. The scheduler schedules the packets to be earlier or later and into the proper time slots in order

An example of our algorithm is represented in Fig. 7. The first packet is scheduled into the 4th frame, which is *FIU* within *D*1 for *C*1,1. Thus, the delay of *C*1,1 is 17. *C*1,2 is scheduled into the 4th frame, which is *FIU* and with a delay of *C*1,2 being 8. Thus, the jitter between *C*1,1 and *C*1,2 is 9, which satisfies the jitter constraint. *C*1,3 is scheduled into the 5th frame according to our algorithm of successive scheduling scheme and with a delay 4. The jitter between *C*1,2 and *C*1,3 is 4, which is smaller than a jitter constraint of 9. *C*1,4 is scheduled into the 7th frame, with a delay of zero time slots and satisfies the jitter constraint of 9 between *C*1,3 and *C*1,4. *C*1,5 is scheduled into the 9th frame with a delay of 4 and the jitter between *C*1,4 and *C*1,5 being 4. Therefore, in order to provide the QoS guarantees of packets scheduling, we need to satisfy

This section evaluated the power consumption of an MSS in terms of the number of listen time slots and status transitions. The QoS requirements of *A*, *B*, *C*, and *D* are listed in Table 2. Both connection types *A* and *B* are VoIP connections. Both connection types *C* and *D* are video streaming connections. The first four connection types on the top half of the list are real-time connections that do not consider the tolerated jitter, and the last four connection types are the same as the first four connection types, but with constrained tolerated jitter. The total energy of an MSS is 1,000,000 units. We compare our proposed SSS algorithm with the Naïve approach without optimizations and the AS approach (Tsao & Chen, 2008). The Naïve approach implies that each connection associates with its preferred type of powersaving class and parameters, and minimizs that packet delay and power consumption for

Fig. 8 shows the operation time and energy usage of an MSS by applying three different scheduling schemes in the different connection types with a varied number of connections without the jitter constraints. In the Naïve approach, the energy usage increases faster than the other two approaches. Because the Naïve approach does not consider the optimization of packet scheduling, it results in the enormous energy consumption in status transitions. The energy usage in the AS approach performs the same as our SSS approach when there is

to satisfy the jitter constraints.

the delay and the jitter constraints.

**3. Simulation results** 

that single connection.

Fig. 7. Example of our SSS algorithm with jitter constraint.

Fig. 5. An example of jitter.

constraint, *Packeti* may only schedule into the time slots of the 2nd time slot to the 4th time slot. Assume we schedule *Packeti* into the 2nd time slot, *Packeti*+1 may only schedule into the time slots of the 5th time slot to the 7th time slot under the jitter constraint. Thus, the previous approaches to power-saving scheduling with QoS may cause the transmission failure when the jiter constraint is not considered.

QoS requirements include the delay and jitter constraints in scheduling packets. However, previous studies focused on delay constraint without considering the effect of the jitter. Therefore, we take the jitter constraint into account in the scheduling algorithm. In Fig. 6, the first packet was scheduled into the 4th frame which is *FIU* (Tsao & Chen, 2008). Thus, the first packet's delay is 17 and satisfies the delay constraint. The second packet is scheduled into the 4th frame, which is *FIU*. The delay of the second packet is 8 and the jitter between the first and second packet is 9, which satisfies the jitter constraint. The third packet is scheduled into the 9th frame, according to the priorities of the frames. If there is no *FIU*, the first priority will be the frame which has the maximum delay. Therefore, the delay of the third packet would be 20 and the jitter between the second and third packet would be 18, which violates the jitter constraint. Once the scheduling violates the jitter constraint, the QoS is no longer guaranteed.

Fig. 6. Example of the scheduling approach (Tsao & Chen, 2008) without considering jitter.

The algorithm of our proposed successive scheduling, which considers jitter constraints, is described in the following two parts. In the first part of our algorithm, the scheduler sorts all connections on an MSS by their delay constraints, and schedules these connections with tight delay requirements. The reason for scheduling connections with tight delay 40 Mobile Networks

constraint, *Packeti* may only schedule into the time slots of the 2nd time slot to the 4th time slot. Assume we schedule *Packeti* into the 2nd time slot, *Packeti*+1 may only schedule into the time slots of the 5th time slot to the 7th time slot under the jitter constraint. Thus, the previous approaches to power-saving scheduling with QoS may cause the transmission failure when

QoS requirements include the delay and jitter constraints in scheduling packets. However, previous studies focused on delay constraint without considering the effect of the jitter. Therefore, we take the jitter constraint into account in the scheduling algorithm. In Fig. 6, the first packet was scheduled into the 4th frame which is *FIU* (Tsao & Chen, 2008). Thus, the first packet's delay is 17 and satisfies the delay constraint. The second packet is scheduled into the 4th frame, which is *FIU*. The delay of the second packet is 8 and the jitter between the first and second packet is 9, which satisfies the jitter constraint. The third packet is scheduled into the 9th frame, according to the priorities of the frames. If there is no *FIU*, the first priority will be the frame which has the maximum delay. Therefore, the delay of the third packet would be 20 and the jitter between the second and third packet would be 18, which violates the jitter constraint. Once the scheduling violates the jitter constraint, the QoS

Fig. 6. Example of the scheduling approach (Tsao & Chen, 2008) without considering jitter.

The algorithm of our proposed successive scheduling, which considers jitter constraints, is described in the following two parts. In the first part of our algorithm, the scheduler sorts all connections on an MSS by their delay constraints, and schedules these connections with tight delay requirements. The reason for scheduling connections with tight delay

Fig. 5. An example of jitter.

is no longer guaranteed.

the jiter constraint is not considered.

requirements first is to not violate their QoS requirements, as we mentioned previously. The second part of our algorithm is composed of three steps we described in the Section 2.3.1. In addition to these, the scheduler examines the difference in the delay between the present packet and the previous packet when scheduling each packet in each step. The difference in the delay between the present packet and the previous packet can be viewed as jitter. The scheduler schedules the packets to be earlier or later and into the proper time slots in order to satisfy the jitter constraints.

An example of our algorithm is represented in Fig. 7. The first packet is scheduled into the 4th frame, which is *FIU* within *D*1 for *C*1,1. Thus, the delay of *C*1,1 is 17. *C*1,2 is scheduled into the 4th frame, which is *FIU* and with a delay of *C*1,2 being 8. Thus, the jitter between *C*1,1 and *C*1,2 is 9, which satisfies the jitter constraint. *C*1,3 is scheduled into the 5th frame according to our algorithm of successive scheduling scheme and with a delay 4. The jitter between *C*1,2 and *C*1,3 is 4, which is smaller than a jitter constraint of 9. *C*1,4 is scheduled into the 7th frame, with a delay of zero time slots and satisfies the jitter constraint of 9 between *C*1,3 and *C*1,4. *C*1,5 is scheduled into the 9th frame with a delay of 4 and the jitter between *C*1,4 and *C*1,5 being 4. Therefore, in order to provide the QoS guarantees of packets scheduling, we need to satisfy the delay and the jitter constraints.

Fig. 7. Example of our SSS algorithm with jitter constraint.
