Figure 1 - uploaded by Ismat Aldmour
Content may be subject to copyright.
Contexts in source publication
Context 1
... is a key technique to achieve peak data rates in LTE networks ( Ji et al., 2012;Fan et al., 2011) thereby increasing their throughput and improving SE (Chen et al., 2011). The AMC is responsible to provide the user's channel condition to the packet scheduler at the eNodeB, which in turn selects the most proper MCS level to that user. Each user performs signal to interference and noise ratio (SINR) measurement for its channel, then maps the obtained SINR value to the corresponding CQI value. After that, the user sends the CQI as a feedback message to its eNodeB, which will adapt the transmission rate according to the CQI feedback ( Ji et al., 2012;3GPP, 2013). LTE network defines 15 levels of CQI depending on the channel condition which start with the worst level (CQI = 1) and end with the highest level (CQI = 15) (see Table 1). Moreover, there are different types of modulation in LTE network such as quadrature phase shift keying (QPSK), 16 quadrature amplitude modulation (QAM), and 64QAM. Each modulation type has several coding rates, which composes the 15 different MCS levels. As shown in Figure 1, users who are located near the eNodeB receive the transmitted data at high bit rate using 64QAM, while cell edge users receive the transmitted data at low bit. This is in order to maintain the connection quality and link stability without the need to increase the transmission power (Fantacci et al., 2009;Cohen et al., ...
Context 2
... AMC can be implemented easily and efficiently in point to point (PtP) transmission mode because MCS for any user can be accurately chosen according to user's CQI feedback report. However, in point to multi-point (PtM) transmission mode (such as eMBMS), a group of users listen to the same broadcast channel, and each user suffers from different channel propagation conditions. The eNodeB should adopt a proper MCS level which is suitable for all users in terms of throughput and robustness (Kim and Cho, 2005;Liang et al., 2011). The eMBMS's SE and throughput directly depend on the selected MCS. According to 3GPP (2006), in OFDMA-based multicast system, the selection of MCS is very important to improve the SE. A user who is located close to its eNodeB can receive the signal at high data rate (e.g., 64QAM modulation level), whereas those who are located far from the eNodeB and suffer bad channel conditions can only reliably receive the signal at low data rates; using, say QPSK modulation. This is depicted in Figure 1. In the literature, there are several MCS selection mechanisms that have been proposed. We next review three of these mechanisms, namely; mechanisms based on fixed user's channel, average user's channel, and worst users channel conditions. a Fixed predetermined MCS selection A fixed MCS level will be selected to cover most of the eMBMS area (around 95%). This mechanism promises to provide QoS with a high robustness and suppresses errors at the expense of throughput and SE (Kim and Cho, 2005). This is adequate if there are few users with steady transmission configurations and interference conditions (3GPP, 2006), otherwise it will waste the allocated bandwidth. The most advantage of fixed mechanism is that it does not need a feedback from eMBMS ...
Context 3
... third scenario was conducted to investigate how the number of users in a multicast group affects the performance of all mechanisms under study. As shown in Figure 14, the throughput and delay of WCG are degraded as the number of users is increased. The reason behind this is the probability of finding a cell edge user increases as the number of users increases. Other mechanisms' performance does not change when the number of users increases. Compared to the conventional WCG and the average mechanisms, the proposed mechanisms show better performance in view of throughput BLER and FI. However, the average MCS introduces less delay than all other mechanisms at all video rates with comparable or better throughputs at high video rates. The high BLER of the average mechanism (more than 3%) makes it unreliable to be used in a real system without combining other techniques such as opportunistic, multirate, multilayer multicast networks. Combining with any one of these techniques would improve the proposed mechanisms. This idea can be a future work. ...
Context 4
... experiments similar to those described in the previous section were applied in this section as well with more performance measures recorded. The M-WCG mechanism has been compared with the conventional WCG mechanism in terms of PLR, throughput, FI, and delay. Moreover, the M-AVG was compared with the conventional average mechanism. Only eMBMS has been activated in the simulation, so no unicast services have been used during the eMBMS session. The whole bandwidth was assigned to the eMBMS and shared between all sessions. We simulated all mechanisms through three scenarios. In first scenario we experimented using different number of multicast groups, each group has the same number of users (N = 60 and video rate equal to 440 kbps). In the second scenario, we experiment using different video rates with single group of 60 users. Finally, a single group with different number of users has been evaluated using a video rate equal to 440 kbps in the third scenario. Moreover, the network traffic was generated from a real video file with 440 kbps. In order to scale the 440 kbps trace file to generate new trace files with different rates, we followed the two steps method as described in Seeling and Reisslein (2005). Since the video application is sensitive to delay, the maximum delay allowed was set to 0.06 second. The scheduler will drop any packet exceeding this. For more accurate results, each scenario was repeated ten times, and then the average performance values were calculated. The simulation parameters are listed in Table 2. Figure 5 shows the PLR vs. number of eMBMS groups. The M-WCG mechanism provides lower PLR compared to the WCG. This is because WCG selects the lowest MCS level which results in low bit rate transmission, and this, in turn, forces the packets to wait in the MAC queue until they are transmitted or dropped once a packet delay exceeds the maximum allowable delay period. The same reason makes the delay of WCG bigger than the delay of the M-WCG, as shown in Figure 7. In fact, the WCG provides lower packet error rate when comparing to the M-WCG. However, when the number of sessions was increased, each session will get a lower bandwidth which may not be enough to transmit the data with low MCS level and then causes in high PLR at the transmitter. As a consequence of the low PLR and high bit rate, M-WCG make its throughput better than the throughput of the WCG mechanism, as plotted in Figure 6. These enhancements of the M-WCG are attained on account of FI degradation. This is shown in Figure 8 where the WCG provides a higher FI compare to the M-WCG. This is because M-WCG transmits the data at a little bit higher rates which cannot be decoded by all UEs. Figure 9 shows that the M-AVG MCS selection mechanism introduces higher throughputs than the conventional Average MCS selection mechanism. The M-AVG MCS selection mechanism provides low PLR compared to the conventional MCS average selection as exhibited in Figure 10. When the number of mobile groups increases, here exceeding four groups, the number of radio resources assigned to each group will not be enough to transmit the data using the selected MCS. Thus, the dropped packets at the MAC layer will increase, as well as, the waiting time of each packet to get scheduled and transmitted through the physical layer, as illustrated in Figure 11. Finally, the FI of all mechanisms are plotted in Figure 12. The best FI was introduced by M-WCG, WCG, and then M-AVG. whereas, the average mechanism introduces the lowest FI when comparing to other mechanisms. ...
Context 5
... experiments similar to those described in the previous section were applied in this section as well with more performance measures recorded. The M-WCG mechanism has been compared with the conventional WCG mechanism in terms of PLR, throughput, FI, and delay. Moreover, the M-AVG was compared with the conventional average mechanism. Only eMBMS has been activated in the simulation, so no unicast services have been used during the eMBMS session. The whole bandwidth was assigned to the eMBMS and shared between all sessions. We simulated all mechanisms through three scenarios. In first scenario we experimented using different number of multicast groups, each group has the same number of users (N = 60 and video rate equal to 440 kbps). In the second scenario, we experiment using different video rates with single group of 60 users. Finally, a single group with different number of users has been evaluated using a video rate equal to 440 kbps in the third scenario. Moreover, the network traffic was generated from a real video file with 440 kbps. In order to scale the 440 kbps trace file to generate new trace files with different rates, we followed the two steps method as described in Seeling and Reisslein (2005). Since the video application is sensitive to delay, the maximum delay allowed was set to 0.06 second. The scheduler will drop any packet exceeding this. For more accurate results, each scenario was repeated ten times, and then the average performance values were calculated. The simulation parameters are listed in Table 2. Figure 5 shows the PLR vs. number of eMBMS groups. The M-WCG mechanism provides lower PLR compared to the WCG. This is because WCG selects the lowest MCS level which results in low bit rate transmission, and this, in turn, forces the packets to wait in the MAC queue until they are transmitted or dropped once a packet delay exceeds the maximum allowable delay period. The same reason makes the delay of WCG bigger than the delay of the M-WCG, as shown in Figure 7. In fact, the WCG provides lower packet error rate when comparing to the M-WCG. However, when the number of sessions was increased, each session will get a lower bandwidth which may not be enough to transmit the data with low MCS level and then causes in high PLR at the transmitter. As a consequence of the low PLR and high bit rate, M-WCG make its throughput better than the throughput of the WCG mechanism, as plotted in Figure 6. These enhancements of the M-WCG are attained on account of FI degradation. This is shown in Figure 8 where the WCG provides a higher FI compare to the M-WCG. This is because M-WCG transmits the data at a little bit higher rates which cannot be decoded by all UEs. Figure 9 shows that the M-AVG MCS selection mechanism introduces higher throughputs than the conventional Average MCS selection mechanism. The M-AVG MCS selection mechanism provides low PLR compared to the conventional MCS average selection as exhibited in Figure 10. When the number of mobile groups increases, here exceeding four groups, the number of radio resources assigned to each group will not be enough to transmit the data using the selected MCS. Thus, the dropped packets at the MAC layer will increase, as well as, the waiting time of each packet to get scheduled and transmitted through the physical layer, as illustrated in Figure 11. Finally, the FI of all mechanisms are plotted in Figure 12. The best FI was introduced by M-WCG, WCG, and then M-AVG. whereas, the average mechanism introduces the lowest FI when comparing to other mechanisms. ...
Context 6
... experiments similar to those described in the previous section were applied in this section as well with more performance measures recorded. The M-WCG mechanism has been compared with the conventional WCG mechanism in terms of PLR, throughput, FI, and delay. Moreover, the M-AVG was compared with the conventional average mechanism. Only eMBMS has been activated in the simulation, so no unicast services have been used during the eMBMS session. The whole bandwidth was assigned to the eMBMS and shared between all sessions. We simulated all mechanisms through three scenarios. In first scenario we experimented using different number of multicast groups, each group has the same number of users (N = 60 and video rate equal to 440 kbps). In the second scenario, we experiment using different video rates with single group of 60 users. Finally, a single group with different number of users has been evaluated using a video rate equal to 440 kbps in the third scenario. Moreover, the network traffic was generated from a real video file with 440 kbps. In order to scale the 440 kbps trace file to generate new trace files with different rates, we followed the two steps method as described in Seeling and Reisslein (2005). Since the video application is sensitive to delay, the maximum delay allowed was set to 0.06 second. The scheduler will drop any packet exceeding this. For more accurate results, each scenario was repeated ten times, and then the average performance values were calculated. The simulation parameters are listed in Table 2. Figure 5 shows the PLR vs. number of eMBMS groups. The M-WCG mechanism provides lower PLR compared to the WCG. This is because WCG selects the lowest MCS level which results in low bit rate transmission, and this, in turn, forces the packets to wait in the MAC queue until they are transmitted or dropped once a packet delay exceeds the maximum allowable delay period. The same reason makes the delay of WCG bigger than the delay of the M-WCG, as shown in Figure 7. In fact, the WCG provides lower packet error rate when comparing to the M-WCG. However, when the number of sessions was increased, each session will get a lower bandwidth which may not be enough to transmit the data with low MCS level and then causes in high PLR at the transmitter. As a consequence of the low PLR and high bit rate, M-WCG make its throughput better than the throughput of the WCG mechanism, as plotted in Figure 6. These enhancements of the M-WCG are attained on account of FI degradation. This is shown in Figure 8 where the WCG provides a higher FI compare to the M-WCG. This is because M-WCG transmits the data at a little bit higher rates which cannot be decoded by all UEs. Figure 9 shows that the M-AVG MCS selection mechanism introduces higher throughputs than the conventional Average MCS selection mechanism. The M-AVG MCS selection mechanism provides low PLR compared to the conventional MCS average selection as exhibited in Figure 10. When the number of mobile groups increases, here exceeding four groups, the number of radio resources assigned to each group will not be enough to transmit the data using the selected MCS. Thus, the dropped packets at the MAC layer will increase, as well as, the waiting time of each packet to get scheduled and transmitted through the physical layer, as illustrated in Figure 11. Finally, the FI of all mechanisms are plotted in Figure 12. The best FI was introduced by M-WCG, WCG, and then M-AVG. whereas, the average mechanism introduces the lowest FI when comparing to other mechanisms. ...
Context 7
... this scenario, we evaluated both the proposed M-WCG and M-AVG mechanisms and the existing WCG and AVG mechanisms. As shown in Figure 13, throughput and delay of all mechanisms increase as the video rate expands, whereas, the block error ratio (BLER) and FI are slightly affected by expanding the video rate. However, the increase in throughput will continue until the transmitted video rate exceeds the available network resources. As shown in Figure 13(a) at the low-rate video (such as 440 kbps), the proposed M-WCG mechanism produces better throughput than the other mechanisms, whereas, the Average mechanism provides the lowest throughput. Once the video rate exceeds the 1,320 kbps, M-AVG provides the best throughput. Figure 13(b) depicts the delay of each mechanism when various video rates are used. The WCG mechanism provides the largest delay among other mechanisms. This is due to its low transmission rate, which necessitates more time to transmit a frame. On the contrary, the lowest delay is obtained with the Average mechanism. Figure 13(c) and 13(d) illustrate the performance of all mechanisms under study in terms of BLER and FI respectively. It is worthy to notice that the conventional WCG provides the lowest BLER and the highest FI. The proposed M-WCG introduces a close result to the WCG. Moreover, the M-AVG mechanism provides a moderate performance, whereas, the Average mechanism provides the worst BLER and FI among all other mechanisms. Overall, the M-AVG mechanism performance is significantly better than the Average mechanism. However, the performance of each mechanism does not much change with all video rates. This is because the BLER and FI are directly affected by the users' channels quality. The high video rates slightly affects the BLER, because the video frame size becomes large and requires several TTIs to be transmitted. During a long frame transmission time, and due to fast fading, some users may fail to receive part of the frame, which results in dropping the frame. Thus, the probability of BLER increases as the frame size increases. ...
Context 8
... this scenario, we evaluated both the proposed M-WCG and M-AVG mechanisms and the existing WCG and AVG mechanisms. As shown in Figure 13, throughput and delay of all mechanisms increase as the video rate expands, whereas, the block error ratio (BLER) and FI are slightly affected by expanding the video rate. However, the increase in throughput will continue until the transmitted video rate exceeds the available network resources. As shown in Figure 13(a) at the low-rate video (such as 440 kbps), the proposed M-WCG mechanism produces better throughput than the other mechanisms, whereas, the Average mechanism provides the lowest throughput. Once the video rate exceeds the 1,320 kbps, M-AVG provides the best throughput. Figure 13(b) depicts the delay of each mechanism when various video rates are used. The WCG mechanism provides the largest delay among other mechanisms. This is due to its low transmission rate, which necessitates more time to transmit a frame. On the contrary, the lowest delay is obtained with the Average mechanism. Figure 13(c) and 13(d) illustrate the performance of all mechanisms under study in terms of BLER and FI respectively. It is worthy to notice that the conventional WCG provides the lowest BLER and the highest FI. The proposed M-WCG introduces a close result to the WCG. Moreover, the M-AVG mechanism provides a moderate performance, whereas, the Average mechanism provides the worst BLER and FI among all other mechanisms. Overall, the M-AVG mechanism performance is significantly better than the Average mechanism. However, the performance of each mechanism does not much change with all video rates. This is because the BLER and FI are directly affected by the users' channels quality. The high video rates slightly affects the BLER, because the video frame size becomes large and requires several TTIs to be transmitted. During a long frame transmission time, and due to fast fading, some users may fail to receive part of the frame, which results in dropping the frame. Thus, the probability of BLER increases as the frame size increases. ...
Context 9
... this scenario, we evaluated both the proposed M-WCG and M-AVG mechanisms and the existing WCG and AVG mechanisms. As shown in Figure 13, throughput and delay of all mechanisms increase as the video rate expands, whereas, the block error ratio (BLER) and FI are slightly affected by expanding the video rate. However, the increase in throughput will continue until the transmitted video rate exceeds the available network resources. As shown in Figure 13(a) at the low-rate video (such as 440 kbps), the proposed M-WCG mechanism produces better throughput than the other mechanisms, whereas, the Average mechanism provides the lowest throughput. Once the video rate exceeds the 1,320 kbps, M-AVG provides the best throughput. Figure 13(b) depicts the delay of each mechanism when various video rates are used. The WCG mechanism provides the largest delay among other mechanisms. This is due to its low transmission rate, which necessitates more time to transmit a frame. On the contrary, the lowest delay is obtained with the Average mechanism. Figure 13(c) and 13(d) illustrate the performance of all mechanisms under study in terms of BLER and FI respectively. It is worthy to notice that the conventional WCG provides the lowest BLER and the highest FI. The proposed M-WCG introduces a close result to the WCG. Moreover, the M-AVG mechanism provides a moderate performance, whereas, the Average mechanism provides the worst BLER and FI among all other mechanisms. Overall, the M-AVG mechanism performance is significantly better than the Average mechanism. However, the performance of each mechanism does not much change with all video rates. This is because the BLER and FI are directly affected by the users' channels quality. The high video rates slightly affects the BLER, because the video frame size becomes large and requires several TTIs to be transmitted. During a long frame transmission time, and due to fast fading, some users may fail to receive part of the frame, which results in dropping the frame. Thus, the probability of BLER increases as the frame size increases. ...
Context 10
... this scenario, we evaluated both the proposed M-WCG and M-AVG mechanisms and the existing WCG and AVG mechanisms. As shown in Figure 13, throughput and delay of all mechanisms increase as the video rate expands, whereas, the block error ratio (BLER) and FI are slightly affected by expanding the video rate. However, the increase in throughput will continue until the transmitted video rate exceeds the available network resources. As shown in Figure 13(a) at the low-rate video (such as 440 kbps), the proposed M-WCG mechanism produces better throughput than the other mechanisms, whereas, the Average mechanism provides the lowest throughput. Once the video rate exceeds the 1,320 kbps, M-AVG provides the best throughput. Figure 13(b) depicts the delay of each mechanism when various video rates are used. The WCG mechanism provides the largest delay among other mechanisms. This is due to its low transmission rate, which necessitates more time to transmit a frame. On the contrary, the lowest delay is obtained with the Average mechanism. Figure 13(c) and 13(d) illustrate the performance of all mechanisms under study in terms of BLER and FI respectively. It is worthy to notice that the conventional WCG provides the lowest BLER and the highest FI. The proposed M-WCG introduces a close result to the WCG. Moreover, the M-AVG mechanism provides a moderate performance, whereas, the Average mechanism provides the worst BLER and FI among all other mechanisms. Overall, the M-AVG mechanism performance is significantly better than the Average mechanism. However, the performance of each mechanism does not much change with all video rates. This is because the BLER and FI are directly affected by the users' channels quality. The high video rates slightly affects the BLER, because the video frame size becomes large and requires several TTIs to be transmitted. During a long frame transmission time, and due to fast fading, some users may fail to receive part of the frame, which results in dropping the frame. Thus, the probability of BLER increases as the frame size increases. ...
Context 11
... modulation type has several coding rates, which composes the 15 different MCS levels. As shown in Figure 1, users who are located near the eNodeB receive the transmitted data at high bit rate using 64QAM, while cell edge users receive the transmitted data at low bit. This is in order to maintain the connection quality and link stability without the need to increase the transmission power (Fantacci et al., 2009;Cohen et al., 2010). ...
Context 12
... user who is located close to its eNodeB can receive the signal at high data rate (e.g., 64QAM modulation level), whereas those who are located far from the eNodeB and suffer bad channel conditions can only reliably receive the signal at low data rates; using, say QPSK modulation. This is depicted in Figure 1. In the literature, there are several MCS selection mechanisms that have been proposed. ...
Context 13
... is because M-WCG transmits the data at a little bit higher rates which cannot be decoded by all UEs. Figure 9 shows that the M-AVG MCS selection mechanism introduces higher throughputs than the conventional Average MCS selection mechanism. The M-AVG MCS selection mechanism provides low PLR compared to the conventional MCS average selection as exhibited in Figure 10. When the number of mobile groups increases, here exceeding four groups, the number of radio resources assigned to each group will not be enough to transmit the data using the selected MCS. ...
Context 14
... the number of mobile groups increases, here exceeding four groups, the number of radio resources assigned to each group will not be enough to transmit the data using the selected MCS. Thus, the dropped packets at the MAC layer will increase, as well as, the waiting time of each packet to get scheduled and transmitted through the physical layer, as illustrated in Figure 11. Finally, the FI of all mechanisms are plotted in Figure 12. ...
Context 15
... the dropped packets at the MAC layer will increase, as well as, the waiting time of each packet to get scheduled and transmitted through the physical layer, as illustrated in Figure 11. Finally, the FI of all mechanisms are plotted in Figure 12. The best FI was introduced by M-WCG, WCG, and then M-AVG. ...
Context 16
... this scenario, we evaluated both the proposed M-WCG and M-AVG mechanisms and the existing WCG and AVG mechanisms. As shown in Figure 13, throughput and delay of all mechanisms increase as the video rate expands, whereas, the block error ratio (BLER) and FI are slightly affected by expanding the video rate. However, the increase in throughput will continue until the transmitted video rate exceeds the available network resources. ...
Context 17
... the increase in throughput will continue until the transmitted video rate exceeds the available network resources. As shown in Figure 13(a) at the low-rate video (such as 440 kbps), the proposed M-WCG mechanism produces better throughput than the other mechanisms, whereas, the Average mechanism provides the lowest throughput. Once the video rate exceeds the 1,320 kbps, M-AVG provides the best throughput. ...
Context 18
... the video rate exceeds the 1,320 kbps, M-AVG provides the best throughput. Figure 13(b) depicts the delay of each mechanism when various video rates are used. The WCG mechanism provides the largest delay among other mechanisms. ...
Context 19
... the contrary, the lowest delay is obtained with the Average mechanism. Figure 13(c) and 13(d) illustrate the performance of all mechanisms under study in terms of BLER and FI respectively. It is worthy to notice that the conventional WCG provides the lowest BLER and the highest FI. ...
Context 20
... third scenario was conducted to investigate how the number of users in a multicast group affects the performance of all mechanisms under study. As shown in Figure 14, the throughput and delay of WCG are degraded as the number of users is increased. The reason behind this is the probability of finding a cell edge user increases as the number of users increases. ...
