Layer-Assisted Adaptive Video Streaming

Size: px
Start display at page:

Download "Layer-Assisted Adaptive Video Streaming"

Transcription

1 Layer-Assisted Adaptive Video Streaming Afshin Taghavi Nasrabadi, Ravi Prakash The University of Texas at Dallas, U.S.A. ABSTRACT HTTP Adaptive Streaming (HAS) is the widely adopted solution for video streaming over the Internet. When network throughput is highly variable, designing an optimal HAS solution that maximizes Quality of Experience (QoE) becomes challenging. Each chunk should be prefetched at highest possible quality while rebufferings and quality switches are minimized. Scalable Video Coding (SVC), with its layered encoding of video, provides more flexibility for HAS clients. It can reduce the occurrence of rebufferings under variable network conditions. However, SVC introduces at least 10% overhead on video bitrate per layer and increases the number of HTTP requests to fetch video chunks. So streaming SVC video at high qualities is more expensive. We propose a solutions that employs both SVC and non-svc video to improve user s QoE while avoiding the increased bandwidth overhead and HTTP signaling of SVC. Our experiments using real-world bandwidth traces show that this method improves QoE compared to the state-of-the-art adaptation methods under various network conditions. KEYWORDS HTTP Adaptive Video Streaming, QoE, Rebuffering 1 INTRODUCTION HTTP Adaptive Streaming (HAS) is widely accepted as the solution for video streaming over the Internet while being compatible with existing infrastructure and protocols. HAS uses commodity HTTP servers to host the video content and content distribution networks (CDNs) to scale to millions of users. HAS divides the video stream into chunks, each of a fixed duration (few seconds), and makes each chunk available at a set of quality levels on the server. Higher quality chunks are encoded at higher bitrate, requiring higher network throughput to be streamed. Video streaming has to meet strict timing requirements while dynamically adapting to variations in the underlying network conditions. Quality of Experience (QoE) of video streaming is affected by three main metrics: (1) the average quality of video playback, (2) quality changes during playback, and (3) the number and duration of playback stalls, also known as rebuffering [16]. Rebuffering adversely affects QoE [12]. The client decides the quality at which it fetches each chunk. Figure 1 shows the building blocks of a HAS Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than ACM must be honored. Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Request permissions from permissions@acm.org. NOSSDAV 18, June 12 15, 2018, Amsterdam, Netherlands 2018 Association for Computing Machinery. ACM ISBN /18/06... $ client. When the client sends a request for a chunk, it waits for that chunk to be downloaded completely, and then puts it in its buffer. Client s buffer holds a number of chunks to provide smooth video playback in spite of network throughput variations. Client should avoid video rebuffering while maintaining the highest possible playback quality with fewest quality switches. If the client chooses a high quality level for a chunk that cannot be delivered before its playback deadline, rebuffering happens. On the other hand, streaming low quality video when the network throughput is high results in bandwidth under-utilization and user dissatisfaction. Moreover, clients should not react to throughput changes too quickly as frequent video quality fluctuation degrades QoE. When transient bandwidth drop happens, the client can use its buffer to provide smooth high quality playback. Many adaptation methods are designed for single-layer encoded video, e.g., H.265/HEVC, since it is the mainstream encoding method, and many devices support its decoding. However, there is some research on the application of Scalable Video Coding (SVC) in HAS [6][4]. SVC layered structure provides more flexibility for clients because they can adapt at finer granularity, i.e, layers instead of chunks, and react better to network conditions. However, encoding overhead of SVC has prevented it from being deployed broadly. An SVC-encoded video has at least 10% higher bitrate per layer compared to non-svc-encoded video at the same quality level [2]. So required throughput for streaming SVC video increases as the number of layers increases. We propose to utilize the advantages of both SVC- and non-svcencoded video for better adaptation. In addition to regular representation set used in non-svc HAS, a set of enhancement layers are made available for each quality level. So, the client should make decision about the base quality for each chunk, similar to regular HAS with non-svc video. When estimated network throughput is low, the client can be conservative and prefetch chunks at lower bitrates to avoid rebuffering. When network conditions improve drastically, the client has the luxury of enhancing the quality of buffered chunks by fetching enhancement layers. Using this method, unlike pure SVC methods, the client does not always start with lowest quality base layer. Thus, the overhead of SVC is not incurred when it is unnecessary. We call this method Layered-Assisted Adaptive Video Streaming (LAAVS). Our experimental results using real-world network throughput traces show that LAAVS outperforms the state-ofthe-art adaptation methods, such as Pensieve [11], under various network conditions. LAAVS is a low complexity algorithm and does not need training, but Pensieve uses machine learning to optimize QoE and needs training. Pensieve performance degrades drastically when it encounters unseen and different scenarios. The rest of this paper is organized as follows. In Section 2, related work is discussed. The proposed method is explained in Section 3. Section 4 presents the experimental results. Finally, Section 5 concludes the paper.

2 Figure 1: Internal components of a HAS client 2 RELATED WORK As depicted in Figure 1, the client uses the following information on its side as inputs to the adaptation logic: (1) Video specs: the number of representations, their quality level, bitrate, and chunk duration, (2) Buffer status: video duration already fetched and stored in the buffer, and (3) Network conditions: perceived network throughput. Existing solutions can be categorized based on how they utilize these information to make pre-fetching decisions, with the goal of optimizing QoE. Throughput-based (TB) methods rely mostly on network throughput estimation. They choose the highest quality that can be streamed, given the estimated throughput [8]. Network bandwidth is dynamic, and bandwidth estimation is done by smoothing recent bandwidth samples to avoid unnecessary quality switches. FESTIVE [7] is one of these solutions that provides a trade-off among fairness, efficiency, and stability where multiple HAS clients are streaming video. However, it is very challenging to predict the future throughput because of interactions between the application layer adaptation and TCP congestion control [1], and highly variable nature of networks, especially mobile wireless networks. Sun et al. [18] have shown that similar video streaming sessions share similar throughput patterns. Based on this observation, they have improved throughput prediction using the Hidden Markov Model. However, exploiting sessions pattern requires large amount of data about the previous sessions. Another challenge for TB methods is Variable Bit-Rate encoded video. The client adapts based on average bitrate, while the size of each segment varies from the average. Buffer-based (BB) methods are proposed to reduce the dependency on bandwidth estimation. Control-theoretic approach in [19] adjusts throughput estimation based on buffer occupancy as a feedback signal. However, Huang et al. [5] observed that this adjustment is very challenging. They proposed a BB approach which maps the instantaneous buffer occupancy to a quality level. But BB methods require longer buffers to have adequate space between buffer thresholds corresponding to quality levels. Otherwise, the frequency of quality switches increases. BOLA [17] is another BB method that tries to maximize QoE by using Lyapunov optimization technique. BOLA also uses download abandonment technique to reduce rebufferings, i.e., when download duration of a chunk is longer than buffer duration. Model Predictive Control (MPC) utilizes buffer occupancy and network throughput estimation to improve QoE under variable conditions [20]. Performance depends on the accuracy of throughput estimation. Machine learning techniques are used in [3][11] for quality adaptation. Pensieve [11] is based on reinforcement learning. It has been shown to outperform other methods by minimizing rebuffering events and maximizing average playback quality. Achievable QoE with Pensieve is approximately within 10% of the offline optimal. However, this method needs training. SVC is a hierarchical encoding method that encodes a video into a base layer and several enhancement layers [2]. Each layer can provide quality/resolution/frame-rate enhancement over lower layers and depends on them. The effectiveness of using SVC in HAS has been studied in [4]. Results show that under stable network conditions, average quality of SVC is lower than non-svc video due to the higher bitrate of SVC. However, for variable network conditions, SVC can perform better since client divides the request per chunk into requests per layer. Using SVC, client has two choices: (1) prefetching a new chunk (base layer), or (2) backfilling: fetching an enhancement layer for an already fetched chunk. Giving priority to prefetching reduces rebufferings. Buffer-based methods are proposed in [14] and [6]. The client starts to fill the buffer with the base layer, and if lower layers for all segments in the buffer were downloaded, it fetches the next enhancement layer for them. Experimental results show that these methods successfully reduce rebuffering events. 3 PROPOSED METHOD 3.1 Video Encoding In HAS, each video chunk is available in a set, S, of representations. Let S = {(q 1, r 1 ), (q 2, r 2 ),..., (q n, r n )}. Each tuple (q i, r i ), 1 i n, is a representation with quality level q i and average bitrate r i. Given j greater than i, we have q j > q i and r j > r i. For each representation i, LAAVS needs enhancement layers to improve the quality of prefetched chunks at quality q i to any quality q j such that i < j n. So for each representation i in S, we will have a set of layers L i = {(bq i,br i ), (eq i,1, er i,1 ),..., (eq i,n i, er i,n i )}, in which (bq i,br i ) is the base layer with quality bq i = q i. Any (eq i, j, er i, j ) is the j th enhancement layer with bitrate of er i, j that can provide quality level equivalent to q i+j, if added on top of all lower layers. So quality level q i can be achieved from all base layers {bq j 1 j i} according to equation 1. i j q i = bq j + eq j,k (1) k=1 i j r i = br j + er j,k (2) k=1 = [1 + (i j)δ] (3) Equation 2 shows that the bitrate of SVC video is times greater than non-svc one. As shown in equation 3, each enhancement layer adds at least δ 10% overhead to the bitrate compared to non-svc video. Figure 2 shows how a video with six quality levels can be represented in LAAVS using a combination of HAS base layer and SVC enhancement layers.

3 from 0.5T e to T e. So, for bitrate selection only in the startup state, T a is used instead of T e. Figure 2: Set of representations for LAAVS 3.2 Adaptation algorithm At each step, the client can take one of these two actions: (1) Prefetching: A new base layer chunk of quality i is requested (bq i ); (2) Backfilling: An enhancement layer is requested for a quality i base layer chunk already in the buffer (eq i, j ). For prefetching, quality of the base layer should be determined. For backfilling, the client should decide which chunk in the buffer should be chosen for quality enhancement, and then the suitable enhancement layer should be requested for that chunk. LAAVS considers both throughput estimation and buffer occupancy level (in seconds) for adaptation. The algorithm has two main states: startup and Prefetching+Backfilling (PB). Assume that the client has a buffer with duration of B seconds, we refer to the current buffer occupancy with B c. We define multiple thresholds on the buffer to switch between states: panic (B p ), backfilling (B b ), and startup (B s ), such that B p < B b < B s. LAAVS starts from initialization and then enters the start-up state. The goal of start-up state is to build up buffer quickly, possibly at a lower quality, to a certain duration (B s ). Then, LAAVS enters the PB state and tries to stream highest possible quality with lowest rebuffering and switches. In this state, both prefetching and backfilling can be done Throughput estimation and bitrate selection. Client estimates throughput from previous chunk downloads: dividing chunk size by download duration. The average of last three throughput samples is taken as the throughput estimate, T e. We use a function that takes T e as input and returns the highest video bitrate lower than T e Initialization. In the beginning of a session, the client does not have sufficient history about network conditions. Although throughput can be estimated from downloading the meta-data at the beginning of a session, we decide to start from a fixed quality level for the first chunk Startup state. Every time buffer occupancy level becomes lower than B p (including in the beginning of a session), the client panics and enters the startup state. It tries to accumulate chunks in the buffer up to B s quickly. Having more chunks in the buffer improves robustness against rebuffering; however, fetching chunks at very low quality reduces QoE. So, proportional to the accumulated chunks and network throughput, we prefetch chunks at lower bitrate. For this purpose, we adjust T e using a linear function according to equation 4, and call it T a. When buffer is less than B p, the client halves T e. From B p to B s, the client linearly increases T a 0.5T e if B c < B p T a = 0.5(B c + B s ) B p (4) T e if B p < B c < B s B s B p Note that, backfilling is not allowed in this state as it does not add any new chunk to the buffer PB state. When buffer occupancy reaches B s, LAAVS enters the PB state, where it tries to get highest possible quality and do backfilling opportunistically. We consider a set of conditions for backfilling. If all conditions are satisfied then an enhancement layer for a specific chunk is fetched; otherwise, prefetching is done. Backfilling does not add any new chunk to the buffer, so if the buffer is shallow, it can increase the risk of rebuffering. So the first condition is that buffer occupancy should be more than a pre-defined threshold B b. To allow backfillings happen from the moment that the client enters PB state, we set B b lower than B s. Then, the client searches the buffer, starting from the most recently downloaded chunk, to find the lowest quality chunk. Let the quality of the found chunk be Q f, while the last prefetched chunk s quality is Q l. Client backfills if Q f < Q l. In other words, the client backfills only up to Q l in order to reduce quality switches and fill the quality gaps between prefetched chunks. Increasing quality to a level higher than Q l is done only during prefetching. Finally, the client checks if the estimated throughput can support streaming of quality level Q f + 1. Assume that the base layer quality of the found chunk is Q b, it checks the following condition: Q f Q b +1 T e > br Qb + er Qb,i. (5) i=1 Note that we consider the cumulative bitrate instead of only that enhancement layer s bitrate; otherwise, download duration of an enhancement layer can be up to that of a complete chunk which results in buffer length reduction. If the conditions for backfilling do not hold, the client determines the quality level for the next chunk to prefetch. In PB state, we make the client more aggressive for getting higher qualities and staying at those quality levels.t e will be given to the bitrate selection function. If chosen quality is higher than current quality level Q l, the client always switches to the higher quality. But, if chosen quality is lower than Q l, the client tries to resist against quality downgrade. The client checks buffer occupancy level B c. If B c is greater than B b, the client tries to be aggressive and resists against quality downgrade. It stays on the current quality level unless br Ql > 1.5T e. If B c is less than B b, the client switches to lower quality level. If the client switched to a lower quality level, it stays at that quality level for the next chunk and does not do any backfilling. In other words, the client backs off from higher quality levels. If the drop is transient, those lower quality chunks will be backfilled later. Whenever the buffer occupancy goes lower than B p, the client switches to the startup state. Considering B = 60s, using experimental analysis, we set the buffer thresholds as follows: B p = 10s, B b = 15s, B s = 20s. Lower

4 Algorithm 1: LAAVS adaptation algorithm This algorithm runs after downloading each chunk or layer if B c < B p then state = star tup ; else if B c > B s then state = P B ; if state is startup then if B < B p then T a = 0.5T e 0.5(Bc +Bs ) Bp else T a = B s B p T e ; i = дet Bitr ate(t a ) //get index of highest bitrate lower than T a return bq i else Q l quality of the most recently prefetched chunk if B c > B b then //find the most recently downloaded chunk with quality < Q l //and return its index, base quality(q b ) and its overall quality(q f ) index,q b, Q f f indchunk(q l ) if index!=null and T e > br Qb + Q f Q b +1 i=1 return eq Qb,f +1 ; end i = дet Bitr ate(t e ) if i > Q l then return bq i else if B c < B b or r Ql > 1.5T e then return bq i else return bq Ql ; end end er Qb,i then values will increase rebuffering probability, and higher values delay switching to higher quality chunks. 4 PERFORMANCE EVALUATION We compare LAAVS with different adaptation algorithms, considering various network conditions and two different video sequences. 4.1 Experimental setup Network Conditions. In our experiments, a client connects to the video source, which is an Apache HTTP server hosting video files, through wireless cellular network. We chose Mahimahi LinkShell [13] for emulating cellular link throughput. Mahimahi emulates cellular link by replaying pre-recorded traces of network packet transmissions. We use 3G and 4G network bandwidth traces. 55 3G/HSDPA traces are taken from the dataset provided by Riiser et al. [15]. It contains bandwidth traces sampled every second from different public transportation mobility scenarios in Norway. We also used 20 4G traces from available traces in Mahimahi emulator. These traces were sampled every millisecond and recorded from major US wireless carriers in car driving scenarios. We considered a round-trip time of 80ms between the client and server Videos. Available bitrates of a video, i.e., bitrate ladder, can affect the performance of adaptation algorithms. Bitrate ladder depends on encoding configuration and the content of the video. If bitrates are close to each other, small bandwidth fluctuations could trigger frequent bitrate switching. An aggressive adaptation algorithm can take advantage of such videos, and opportunistically switch to higher resolutions, while network bandwidth is low and Figure 3: Average bitrate per resolution for YouTube videos ± one standard deviation Table 1: Bitrate ladder for videos (Envivio, TOS) in Mbps Cumulative bitrate up to: Rep. Res. r rep er 1 er 2 er 3 er 4 er p (0.30, 0.25) (0.87, 0.92) (1.56, 1.95) (2.68, 3.48) (4.56, 6.40) (7.53, 17.50) 2 360p (0.75, 0.80) (1.38, 1.73) (2.41, 3.12) (4.13, 5.80) (6.88, 16.00) p (1.20, 1.50) (2.13, 2.76) (3.70, 5.20) (6.23, 14.50) p (1.85, 2.40) (3.28, 4.60) (5.60, 13.00) p (2.85, 4.00) (4.95, 11.50) p (4.30, 10.00) buffer occupancy is high. However, this strategy may not work for a video with wider gaps between bitrates. We consider both scenarios to evaluate how the aggressiveness of different methods can affect QoE. Since YouTube uses an efficient title-based encoding method [10], we studied bitrate ladder of over 200 videos from different categories on YouTube. As shown in Figure 3, from each representation to the next one the bitrate roughly doubles. However, the video sequences used in research papers usually have tighter gaps [11][20]. So we use two video sequences, the first one is EnvivioDASH3 1 which is used in [11] and encoded at lower bitrates with narrower gaps. The second is Tears of Steal (ToS) 2 [9], and we used the bitrates according to its YouTube version 3, encoded at higher bitrate with wider gaps. Each video is available in 6 different resolutions:{240p, 360p, 480p, 720p, 1080p, 1440p}. We selected the first 200 seconds of each video, with chunk duration of 4 seconds, for a total of 50 chunks per video. The bitrates for each video are provided in Table 1. Bitrate of each representation, r rep, is in the third column, and it is the same as the bitrate for the base layer of SVC representations. All other columns to the right are the cumulative bitrates up to a specific enhancement layer for SVC version. SVC encoded videos are generated from the non-svc videos with the assumption of 15% more bitrate per layer compared to non-svc one at the same quality level. With the combination of 3G and 4G traces and two videos, we have 4 different scenarios. We refer to these scenarios with format of (3g/4G-Envivio/ToS) Adaptation algorithms. We have chosen the following adaptation algorithms for comparison: Pensieve: This method uses reinforcement learning to train an algorithm that maximizes QoE [11]. We used the source code and trained model provided by the authors

5 MPC: This method tries to make QoE-optimal decisions by considering throughput measurements of past 5 downloaded chunks and buffer occupancy as feedback signals [20]. We used robustmpc implementation with error prediction in throughput estimation 4. BOLA: We used BOLA [17] implementation available in dash.js version Dash.js: We used the default throughput-based adaptation algorithm of dash.js reference client as a throughput-based method. We used dash.js client version 2.6.2s 5. Buffer-based: The method introduced by Huang et al. [5] is used as a pure buffer-based adaptation algorithm. We used the implementation provided with Pensieve authors 4. The parameters for reservoir and cushion were set to 0.1B and 0.8B, respectively. WQUAD: This is an adaptation method designed for SVC encoded video [6]. We implemented this method in our experimental framework. For all the above methods except WQUAD, the Pensieve automated experimental setup is used to record information about QoE. We implemented LAAVS and WQUAD in Java. The client sends requests to the server through an HTTP connection, and places the downloaded chunks inside its buffer. Playback is emulated, and logs containing all information about downloading chunks, playback quality, and rebuffering events are recorded. All adaptation methods were configured to have a buffer with duration of B = 60s QoE metric. To measure achieved QoE from different adaptation algorithms, we use the QoE metric introduced in [20]. This metric is calculated by assigning weights to overall quality, rebuffering duration, and switches count. In this metric, instead of considering a subjective or objective quality level for each chunk, the average bitrate of each chunk is considered for quality. Let R t denote the bitrate of t th chunk, and RB t denote the duration of rebuffering during downloading that chunk. Equation (6) shows the QoE calculation for a video with N chunks. Rebuffering is penalized by weight of µ which is considered equal to the highest bitrate of a video. Quality switches are penalized by weight of -1. N N N 1 QoE = R n µ RB n R n+1 R n (6) n=1 n=1 n=1 Since the bitrate of SVC video is higher than non-svc one, we consider R t of an SVC video equal to the bitrate of the non-svc videe at the same quality level. For each video streaming session, we calculate the average QoE by dividing the output of equation (6) by the number of chunks. 4.2 Results Figure 4 shows the results for all adaptation algorithms in each scenario. QoE is normalized according to the maximum value. In all scenarios, LAAVS outperforms other methods in terms of QoE by at least 4% and up to 64%. This performance is achieved by streaming highest possible bitrate with less rebuffering events and low switches. In some cases, such as 3G-ToS, although LAAVS has lower bitrate than BB and MPC, it has less rebuffering and switches. 5 Figure 4: QoE and its components for different scenarios In most cases, LAAVS is followed by Pensieve. It has been shown in [11] that Pensieve s QoE is around 10% lower than the offline optimal in 3G-Envivio scenario. LAAVS performs 4% better than Pensieve in this scenario. However, Pensieve has the lowest QoE in ToS-4G scenario. Pensieve is trained with 3G network bandwidth traces that have lower bandwidth capacity compared to 4G traces. Moreover, Pensieve was trained for videos with narrow gaps between bitrates. Considering six quality levels, most of the time Pensieve uses quality levels 1,2,4, and 6. It fills the buffer with low quality chunks, and then jumps to a quality level with bitrate higher than network throughput. Since gaps between bitrates is not wide, it can stay at those quality levels for a while and then switch back to a lower quality level. This approach increases QoE for 3G-Envivio scenario, but fails for videos with higher bitrate gaps. Therefore, although it performs well in the trained scenarios, under other conditions with another video with different bitrate ladder the performance degrades drastically. One of the main reasons behind designing LAAVS is reducing rebuffering. Since client can enhance quality later, when buffer occupancy is low, it switches to lower qualities and stays at those levels. If network conditions improve, client switches to higher quality and backfills lower quality chunks. Otherwise, it has switched to a lower quality earlier, and prevented rebuffering to happen. BB and WQUAD have lower rebuffering than LAAVS in most scenarios, but at the cost of lower bitrate and QoE. LAAVS is able to stream higher bitrate video with low rebuffering. WQUAD streams pure SVC video. As it always starts from the lowest base quality and then backfills, it reduces rebuffering significantly. However, it has the lowest bitrate score in all scenarios. The reason is SVC encoding overhead. For example if the client wants

6 5 CONCLUSION We presented a new technique for HAS, called LAAVS. It has the advantages of SVC video without incurring associated overhead all the time. LAAVS uses both throughput estimation and buffer occupancy information to prefetch and backfill chunks. So it can switch and stay at low quality chunks during bandwidth drops and backfill later. This method reduces rebufferings significantly without compromising quality. Our experimental results using realworld bandwidth traces show that LAAVS can outperform existing solutions such as Pensieve. Figure 5: Heatmap of QoE for 3G-Envivio Scenario Table 2: % of downloaded enh. layers per scenario % of chunks backfilled with Trace-Video enh. layer: G-Envivio G-Envivio G-ToS G-ToS to stream highest resolution, then it has to download the video with 75% higher bitrate. Therefore, it has lower QoE score in general. In terms of switching, whenever LAAVS backfills, not only it enhances the quality, it also reduces the quality variations. So, although LAAVS does not control quality switches in start-up state, and has some unnecessary down switches in the PB state, backfilling enables it to have relatively low switching frequency in all scenarios. Figure 5 presents the heatmap of QoE results for 3G-Envivio scenario. Each heatmap demonstrates how QoE for an adaptation method changes according to average throughput and throughput variation of a bandwidth trace. X axis is average throughput, and Y axis is throughput standard deviation. Each point is the achieved QoE for a single bandwidth trace. QoE values are demonstrated with colors, and we filled the gaps between points by interpolation. Higher network throughput results in higher QoE, and lower throughput with high variation results in lower QoE because of rebufferings. Although Pensieve and MPC have achieved high QoE for high throughput scenarios, they have poor performance in low throughput and high variance traces. BB and WQUAD have higher QoE for low throughput and high variance scenarios. LAAVS is able to combine the benefits of SVC and non-svc together. In high throughput scenarios it has high QoE (unlike BB and WQUAD), and in low throughput and high variation scenarios it has comparable performance to BB and WQUAD Overhead and Complexity. LAAVS tries to avoid using all layers (low quality base layer) when it is unnecessary. Table 2 shows the percentage of backfilled chunks with enhancement layers from different levels for all scenarios. It can be seen that most of the time enhancement layers up to level 3 are required. Thus, using LAAVS, up to 3 layers can be used which can decrease encoding, storage, and decoding overhead and complexity. 6 ACKNOWLEDGMENTS This work was supported in part by a gift from Cisco Systems, Inc. REFERENCES [1] S. Akhshabi, A. C Begen, and C. Dovrolis An experimental evaluation of rate-adaptation algorithms in adaptive streaming over HTTP. In Proc. ACM MMsys. [2] J. M Boyce, Y. Ye, J. Chen, and A. K Ramasubramonian Overview of SHVC: scalable extensions of the high efficiency video coding standard. IEEE Transactions on Circuits and Systems for Video Technology 26, 1 (2016), [3] F. Chiariotti, S. D Aronco, L. Toni, and P. Frossard Online learning adaptation strategy for DASH clients. In Proc. ACM MMSys. [4] J. Famaey, S. Latré, N. Bouten, W. Van de Meerssche, B. De Vleeschauwer, W. Van Leekwijck, and F. De Turck On the merits of SVC-based HTTP adaptive streaming. In 2013 IFIP/IEEE International Symposium on Integrated Network Management (IM 2013). IEEE, [5] T. Huang, R. Johari, N. McKeown, M. Trunnell, and M. Watson A bufferbased approach to rate adaptation: Evidence from a large video streaming service. ACM SIGCOMM 44, 4 (2015), [6] J. Hwang, J. Lee, and C. Yoo Eliminating bandwidth estimation from adaptive video streaming in wireless networks. Signal Processing: Image Communication 47 (2016), [7] J. Jiang, V. Sekar, and H. Zhang Improving fairness, efficiency, and stability in http-based adaptive video streaming with festive. In Proceedings of the 8th international conference on Emerging networking experiments and technologies. ACM, [8] J. Kua, G. Armitage, and P. Branch A Survey of Rate Adaptation Techniques for Dynamic Adaptive Streaming over HTTP. IEEE Communications Surveys & Tutorials (2017). [9] S. Lederer, C. Müller, and C. Timmerer Dynamic adaptive streaming over HTTP dataset. In Proc. ACM MMSys. ACM, [10] Y. Lin, H. Denman, and A. Kokaram Multipass encoding for reducing pulsing artifacts in cloud based video transcoding. In Proc. IEEE ICIP [11] H. Mao, R. Netravali, and M. Alizadeh Neural Adaptive Video Streaming with Pensieve. In Proceedings of the Conference of the ACM Special Interest Group on Data Communication. ACM, [12] R. Mok, E. Chan, and R. Chang Measuring the quality of experience of HTTP video streaming. In 12th IFIP/IEEE International Symposium on Integrated Network Management (IM 2011) and Workshops. IEEE, [13] R. Netravali et al Mahimahi: a lightweight toolkit for reproducible web measurement. ACM SIGCOMM 44, 4 (2015), [14] S. Petrangeli, N. Bouten, M. Claeys, and F. De Turck Towards SVC-based adaptive streaming in information centric networks. In Proc. IEEE ICMEW [15] H. Riiser, P. Vigmostad, C. Griwodz, and P. Halvorsen Commute path bandwidth traces from 3G networks: analysis and applications. In Proc. ACM MMSys [16] M. Seufert, S. Egger, M. Slanina, T. Zinner, T. Hoßfeld, and P. Tran-Gia A survey on quality of experience of http adaptive streaming. IEEE Communications Surveys & Tutorials 17, 1 (2015), [17] K. Spiteri, R. Urgaonkar, and R. Sitaraman BOLA: near-optimal bitrate adaptation for online videos. In Proc. IEEE INFOCOM. IEEE, 1 9. [18] Y. Sun, X. Yin, J. Jiang, V. Sekar, F. Lin, N. Wang, T. Liu, and B. Sinopoli Cs2p: Improving video bitrate selection and adaptation with data-driven throughput prediction. In Proc. ACM SIGCOMM [19] G. Tian and Y. Liu Towards agile and smooth video adaptation in dynamic HTTP streaming. In Proceedings of the 8th international conference on Emerging networking experiments and technologies. ACM, [20] X. Yin, A. Jindal, V. Sekar, and B. Sinopoli A control-theoretic approach for dynamic adaptive video streaming over HTTP. ACM SIGCOMM 45, 4 (2015),

A Comparative Case Study of HTTP Adaptive Streaming Algorithms in Mobile Networks

A Comparative Case Study of HTTP Adaptive Streaming Algorithms in Mobile Networks Theodoros Karagkioules, Cyril Concolato, Dimitrios Tsilimantos and Stefan Valentin LTCI Telecom ParisTech Universite Paris-Saclay Mathematical and Algorithmic Sciences Lab France Research Center Huawei

More information

Lecture 18: Video Streaming

Lecture 18: Video Streaming MIT 6.829: Computer Networks Fall 2017 Lecture 18: Video Streaming Scribe: Zhihong Luo, Francesco Tonolini 1 Overview This lecture is on a specific networking application: video streaming. In particular,

More information

Oboe: Auto-tuning Video ABR Algorithms to Network Conditions

Oboe: Auto-tuning Video ABR Algorithms to Network Conditions Oboe: Auto-tuning Video ABR Algorithms to Network Conditions Zahaib Akhtar, Yun Seong Nam, Ramesh Govindan, Sanjay Rao, Jessica Chen, Ethan Katz-Bassett, Bruno Ribeiro, Jibin Zhan, Hui Zhang : Co-primary

More information

Qoe-aware adaptive bitrate video streaming over mobile networks with caching proxy

Qoe-aware adaptive bitrate video streaming over mobile networks with caching proxy University of New Brunswick From the SelectedWorks of Jun He March 30, 2015 Qoe-aware adaptive bitrate video streaming over mobile s with caching proxy Jun He, University of New Brunswick Available at:

More information

SARA: Segment Aware Rate Adaptation for DASH Video Services

SARA: Segment Aware Rate Adaptation for DASH Video Services SARA: Segment Aware Rate Adaptation for DASH Video Services, Venkatesh Tamarapalli*, Deep Medhi University of Missouri Kansas City * Indian Institute of Technology-Guwahati, India Overview Introduction

More information

Downton Abbey Without the Hiccups: Buffer-Based Rate Adaptation for HTTP Video Streaming

Downton Abbey Without the Hiccups: Buffer-Based Rate Adaptation for HTTP Video Streaming Downton Abbey Without the Hiccups: Buffer-Based Rate Adaptation for HTTP Video Streaming Te-Yuan Huang Ramesh Johari Nick McKeown Stanford University {huangty,ramesh.johari, nickm}@stanford.edu ABSTRACT

More information

Adaptive 360-Degree Video Streaming using Scalable Video Coding

Adaptive 360-Degree Video Streaming using Scalable Video Coding Adaptive 360-Degree Video Streaming using Scalable Video Coding Afshin Taghavi Nasrabadi, Anahita Mahzari, Joseph D. Beshay, Ravi Prakash The University of Texas at Dallas 800 W Campbell Rd. Richardson,

More information

arxiv: v1 [cs.mm] 4 May 2017

arxiv: v1 [cs.mm] 4 May 2017 A Comparative Case Study of HTTP Adaptive Streaming Algorithms in Mobile Networks arxiv:175.1762v1 [cs.mm] 4 May 217 Theodoros Karagkioules, Cyril Concolato LTCI, Télécom ParisTech, Université Paris-Saclay

More information

Neural Adaptive Video Streaming with Pensieve

Neural Adaptive Video Streaming with Pensieve Neural Adaptive Video Streaming with Hongzi Mao, Ravi Netravali, Mohammad Alizadeh MIT Computer Science and Artificial Intelligence Laboratory {hongzi,ravinet,alizadeh}@mit.edu ABSTRACT Client-side video

More information

An HTTP/2 Push-Based Approach for SVC Adaptive Streaming

An HTTP/2 Push-Based Approach for SVC Adaptive Streaming An HTTP/2 Push-Based Approach for SVC Adaptive Streaming Jeroen van der Hooft, Stefano Petrangeli, Niels Bouten, Tim Wauters, Rafael Huysegems, Tom Bostoen and Filip De Turck Ghent University - iminds,

More information

Minimizing the Impact of Delay on Live SVC-based HTTP Adaptive Streaming Services

Minimizing the Impact of Delay on Live SVC-based HTTP Adaptive Streaming Services Minimizing the Impact of Delay on Live SVC-based HTTP Adaptive Streaming Services Niels Bouten, Steven Latré, Jeroen Famaey, Filip De Turck Department of Information Technology, Ghent University - iminds

More information

Automated Performance Evaluation of Adaptive HTML5 Player Deployments

Automated Performance Evaluation of Adaptive HTML5 Player Deployments Automated Performance Evaluation of Adaptive HTML5 Player Deployments Anatoliy Zabrovskiy, Evgeny Petrov, Evgeny Kuzmin Petrozavodsk State University Petrozavodsk, Russia {z anatoliy, johnp, kuzmin}@petrsu.ru

More information

GroupCast: Preference-Aware Cooperative Video Streaming with Scalable Video Coding

GroupCast: Preference-Aware Cooperative Video Streaming with Scalable Video Coding 1 GroupCast: Preference-Aware Cooperative Video Streaming with Scalable Video Coding Anis Elgabli and Vaneet Aggarwal Purdue University, West Lafayette, IN 47906 Email: {aelgabli,vaneet}@purdue.edu Abstract

More information

Streamloading: Low cost high quality video streaming. Shivendra S. Panwar (joint work with Amir Hosseini and Fraida Fund)

Streamloading: Low cost high quality video streaming. Shivendra S. Panwar (joint work with Amir Hosseini and Fraida Fund) Streamloading: Low cost high quality video streaming Shivendra S. Panwar (joint work with Amir Hosseini and Fraida Fund) Modern Cellular Networks Modern Cellular Networks Evolution of modern cellular networks:

More information

A QoE Friendly Rate Adaptation Method for DASH

A QoE Friendly Rate Adaptation Method for DASH A QoE Friendly Rate Adaptation Method for DASH Yuming Cao 1,3, Xiaoquan You 2,3, Jia Wang 1,3, Li Song 1,3 1 Institute of Image Communication and Network Engineering, Shanghai Jiao Tong University 2 Communication

More information

A Control Theoretic Approach to ABR Video Streaming: A Fresh Look at PID-based Rate Adaptation

A Control Theoretic Approach to ABR Video Streaming: A Fresh Look at PID-based Rate Adaptation A Control Theoretic Approach to ABR Video Streaming: A Fresh Look at PID-based Rate Adaptation Yanyuan Qin, Ruofan Jin, Shuai Hao, Krishna R. Pattipati, Feng Qian, Subhabrata Sen, Bing Wang, Chaoqun Yue

More information

An Experimental Evaluation of Rate Adaptation Algorithms in Adaptive Streaming over HTTP

An Experimental Evaluation of Rate Adaptation Algorithms in Adaptive Streaming over HTTP An Experimental Evaluation of Rate Adaptation Algorithms in Adaptive Streaming over HTTP Saamer Akhshabi, Constantine Dovrolis Georgia Institute of Technology Ali C. Begen Cisco Systems February 24, 2011

More information

QoE-aware Traffic Shaping for HTTP Adaptive Streaming

QoE-aware Traffic Shaping for HTTP Adaptive Streaming , pp.33-44 http://dx.doi.org/10.14257/ijmue.2014.9.2.04 QoE-aware Traffic Shaping for HTTP Adaptive Streaming Xinying Liu 1 and Aidong Men 2 1,2 Beijing University of Posts and Telecommunications No.10

More information

What Happens When HTTP Adaptive Streaming Players Compete for Bandwidth?

What Happens When HTTP Adaptive Streaming Players Compete for Bandwidth? What Happens When HTTP Adaptive Streaming Players Compete for Bandwidth? Saamer Akhshabi, Ali C. Begen Lakshmi Anantakrishnan, Video and Content Platforms Research and Constantine Dovrolis, Advanced Development

More information

arxiv: v1 [cs.mm] 29 Aug 2016

arxiv: v1 [cs.mm] 29 Aug 2016 On the Efficiency and Fairness of Multiplayer HTTP-based Adaptive Video Streaming Xiaoqi Yin, Mihovil Bartulović, Vyas Sekar, Bruno Sinopoli 1 arxiv:1608.08469v1 [cs.mm] 29 Aug 2016 Abstract User-perceived

More information

Evaluating and Improving Push based Video Streaming with HTTP/2

Evaluating and Improving Push based Video Streaming with HTTP/2 Evaluating and Improving Push based Video Streaming with HTTP/2 Mengbai Xiao 1 Viswanathan Swaminathan 2 Sheng Wei 2,3 Songqing Chen 1 1 Dept. of CS George Mason University {mxiao3, sqchen}@gmu.edu 2 Adobe

More information

QoE-Driven Video Streaming and Video Content Caching

QoE-Driven Video Streaming and Video Content Caching CommNet2 & IcoreJoint Workshop on Content Caching & Distributed Storage for Future Communication Networks QoE-Driven Video Streaming and Video Content Caching Xiaohong Peng Adaptive Communications Networks

More information

Empirical study for Dynamic Adaptive Video Streaming Service based on Google Transport QUIC protocol

Empirical study for Dynamic Adaptive Video Streaming Service based on Google Transport QUIC protocol Empirical study for Dynamic Adaptive Video Streaming Service based on Google Transport QUIC protocol Van TONG, Hai Anh TRAN, Sami SOUIHI and Abdelhamid MELLOUK LISSI-TincNET Research Team University of

More information

Using Bandwidth Aggregation to Improve the Performance of Video Quality- Adaptive Streaming Over Multiple Wireless Access Networks

Using Bandwidth Aggregation to Improve the Performance of Video Quality- Adaptive Streaming Over Multiple Wireless Access Networks Using Bandwidth Aggregation to Improve the Performance of Video Quality- Adaptive Streaming Over Multiple Wireless Access Networks 1 J.P. Satheesh, 2 R. Karthikeyan 1 PG Student, 2 Assistant Professor,

More information

Future Buffer based Adaptation for VBR Video Streaming over HTTP

Future Buffer based Adaptation for VBR Video Streaming over HTTP Future Buffer based Adaptation for VBR Video Streaming over HTTP Tuan Vu 1, Hung T. Le 2, Duc V. Nguyen 2, Nam Pham Ngoc 1, Truong Cong Thang 2 1 Hanoi University of Science and Technology, Hanoi, Vietnam

More information

Adaptation Algorithm for Adaptive Streaming over HTTP

Adaptation Algorithm for Adaptive Streaming over HTTP Adaptation Algorithm for Adaptive Streaming over HTTP Konstantin Miller, Emanuele Quacchio, Gianluca Gennari and Adam Wolisz Technische Universität Berlin, Germany Email: {konstantin.miller, adam.wolisz}@tu-berlin.de

More information

Neural Adaptive Content-aware Internet Video Delivery. Hyunho Yeo, Youngmok Jung, Jaehong Kim, Jinwoo Shin, Dongsu Han

Neural Adaptive Content-aware Internet Video Delivery. Hyunho Yeo, Youngmok Jung, Jaehong Kim, Jinwoo Shin, Dongsu Han Neural Adaptive Content-aware Internet Video Delivery Hyunho Yeo, Youngmok Jung, Jaehong Kim, Jinwoo Shin, Dongsu Han Observation on Current Video Ecosystem 2 Adaptive streaming has been widely deployed

More information

Queueing Theoretic Approach to Playout Buffer Model for HTTP Adaptive Streaming

Queueing Theoretic Approach to Playout Buffer Model for HTTP Adaptive Streaming KSII TRANSACTIONS ON INTERNET AND INFORMATION SYSTEMS VOL. 12, NO. 8, Aug. 218 3856 Copyright c 218 KSII Queueing Theoretic Approach to Playout Buffer Model for HTTP Adaptive Streaming Jiwoo Park and Kwangsue

More information

IMPROVING LIVE PERFORMANCE IN HTTP ADAPTIVE STREAMING SYSTEMS

IMPROVING LIVE PERFORMANCE IN HTTP ADAPTIVE STREAMING SYSTEMS IMPROVING LIVE PERFORMANCE IN HTTP ADAPTIVE STREAMING SYSTEMS Kevin Streeter Adobe Systems, USA ABSTRACT While HTTP adaptive streaming (HAS) technology has been very successful, it also generally introduces

More information

NETWORK-ASSISTED STRATEGY FOR DASH OVER CCN. Rihab Jmal*, Gwendal Simon** and Lamia Chaari* *University of Sfax/ CRNS Tunisia **IMT Atlantique, France

NETWORK-ASSISTED STRATEGY FOR DASH OVER CCN. Rihab Jmal*, Gwendal Simon** and Lamia Chaari* *University of Sfax/ CRNS Tunisia **IMT Atlantique, France NETWORK-ASSISTED STRATEGY FOR DASH OVER CCN Rihab Jmal*, Gwendal Simon** and Lamia Chaari* *University of Sfax/ CRNS Tunisia **IMT Atlantique, France ABSTRACT MPEG Dynamic Adaptive Streaming over HTTP

More information

Survey on Concurrent Multipath Scheduling for Real Time Video Streaming in Wireless Network

Survey on Concurrent Multipath Scheduling for Real Time Video Streaming in Wireless Network RESEARCH ARTICLE Survey on Concurrent Multipath Scheduling for Real Time Video Streaming in Wireless Network Rohit Salkute 1, Prof. D.G. Vyawahare 2 1(Computer Science and Engineering, SGBAU, Amravati

More information

Quality of Experience-Centric Management of Adaptive Video Streaming Services: Status and Challenges

Quality of Experience-Centric Management of Adaptive Video Streaming Services: Status and Challenges Quality of Experience-Centric Management of Adaptive Video Streaming Services: Status and Challenges STEFANO PETRANGELI, JEROEN VAN DER HOOFT, TIM WAUTERS, and FILIP DE TURCK, Ghent University - imec,

More information

Video Splicing Techniques for P2P Video Streaming

Video Splicing Techniques for P2P Video Streaming Video Splicing Techniques for P2P Video Streaming MD Iftakharul Islam and Javed I Khan Kent State University Email: mislam4@kent.edu, javed@cs.kent.edu Abstract In HTTP live streaming (HLS), a video is

More information

MULTI-BUFFER BASED CONGESTION CONTROL FOR MULTICAST STREAMING OF SCALABLE VIDEO

MULTI-BUFFER BASED CONGESTION CONTROL FOR MULTICAST STREAMING OF SCALABLE VIDEO MULTI-BUFFER BASED CONGESTION CONTROL FOR MULTICAST STREAMING OF SCALABLE VIDEO Chenghao Liu 1, Imed Bouazizi 2 and Moncef Gabbouj 1 1 Department of Signal Processing, Tampere University of Technology,

More information

Improving the Expected Quality of Experience in Cloud-Enabled Wireless Access Networks

Improving the Expected Quality of Experience in Cloud-Enabled Wireless Access Networks Improving the Expected Quality of Experience in Cloud-Enabled Wireless Access Networks Dr. Hang Liu & Kristofer Smith Department of Electrical Engineering and Computer Science The Catholic University of

More information

360 Degree Video Streaming

360 Degree Video Streaming 360 Degree Video Streaming Yao Wang Dept. of Electrical and Computer Engineering Tandon School of Engineering New York University http://vision.poly.edu 360 Video Streaming https://www.youtube.com/watch?v=wsmjbmxputc

More information

On the Feasibility of DASH Streaming in the Cloud

On the Feasibility of DASH Streaming in the Cloud On the Feasibility of DASH Streaming in the Cloud Cong Wang and Michael Zink Department of Electrical and Computer Engineering University of Massachusetts, Amherst, MA 13 {cwang, zink}@ecs.umass.edu ABSTRACT

More information

On the merits of SVC-based HTTP Adaptive Streaming

On the merits of SVC-based HTTP Adaptive Streaming On the merits of SVC-based HTTP Adaptive Streaming Jeroen Famaey, Steven Latré, Niels Bouten, Wim Van de Meerssche, Bart De Vleeschauwer, Werner Van Leekwijck, and Filip De Turck Ghent University IBBT

More information

QoE-Aware Dynamic Video Rate Adaptation

QoE-Aware Dynamic Video Rate Adaptation QoE-Aware Dynamic Video Rate Adaptation Yanjiao Chen, Fan Zhang, Fan Zhang, Kaishun Wu, Qian Zhang Department of Computer Science & Engineering, Hong Kong University of Science and Technology Department

More information

Dynamic Adaptive Streaming over HTTP (DASH) using feedback linearization: a comparison with a leading Italian TV operator

Dynamic Adaptive Streaming over HTTP (DASH) using feedback linearization: a comparison with a leading Italian TV operator Dynamic Adaptive Streaming over HTTP (DASH) using feedback linearization: a comparison with a leading Italian TV operator Vito Caldaralo, Luca De Cicco, Saverio Mascolo, and Vittorio Palmisano Abstract

More information

A Markov Model for Evaluating Resource Sharing Policies for DASH Assisting Network Elements

A Markov Model for Evaluating Resource Sharing Policies for DASH Assisting Network Elements 2016 28th International Teletraffic Congress A Markov Model for Evaluating Resource Sharing Policies for DASH Assisting Network Elements Jan Willem Kleinrouweler, Sergio Cabrero, Rob van der Mei, Pablo

More information

Receiver-initiated Sending-rate Control based on Data Receive Rate for Ad Hoc Networks connected to Internet

Receiver-initiated Sending-rate Control based on Data Receive Rate for Ad Hoc Networks connected to Internet Receiver-initiated Sending-rate Control based on Data Receive Rate for Ad Hoc Networks connected to Internet Akihisa Kojima and Susumu Ishihara Graduate School of Engineering, Shizuoka University Graduate

More information

Mobile Video Streaming with Video Quality and Streaming Performance Guarantees

Mobile Video Streaming with Video Quality and Streaming Performance Guarantees Mobile Video Streaming with Video Quality and Streaming Performance Guarantees Victor K. C. Wu, Yan Liu, and Jack Y. B. Lee Department of Information Engineering The Chinese University of Hong Kong Hong

More information

On the Efficacy of the Dynamic Adaptive Streaming Over HTTP (DASH) Protocol

On the Efficacy of the Dynamic Adaptive Streaming Over HTTP (DASH) Protocol On the Efficacy of the Dynamic Adaptive Streaming Over HTTP (DASH) Protocol James Martin, Yunhui Fu, Gongbing Hong School of Computing Clemson University Clemson, USA Jim.martin@cs.clemson.edu Abstract

More information

MaVIS: Media-aware Video Streaming Mechanism

MaVIS: Media-aware Video Streaming Mechanism MaVIS: Media-aware Video Streaming Mechanism Sunhun Lee and Kwangsue Chung School of Electronics Engineering, Kwangwoon University, Korea sunlee@adamskwackr and kchung@kwackr Abstract Existing streaming

More information

Video Through a Crystal Ball: Effect of Bandwidth Prediction Quality on Adaptive Streaming in Mobile Environments

Video Through a Crystal Ball: Effect of Bandwidth Prediction Quality on Adaptive Streaming in Mobile Environments Video Through a Crystal Ball: Effect of Bandwidth Prediction Quality on Adaptive Streaming in Mobile Environments Tarun Mangla, Nawanol Theera-Ampornpunt, Mostafa Ammar, Ellen Zegura, Saurabh Bagchi Georgia

More information

Why Is HTTP Adaptive Streaming So Hard?

Why Is HTTP Adaptive Streaming So Hard? Why Is HTTP Adaptive Streaming So Hard? Sangwook Bae Dahyun Jang KyoungSoo Park Department of Electrical Engineering, KAIST Daejeon, South Korea hoops@ndsl.kaist.edu jangdh93@kaist.ac.kr kyoungsoo@ee.kaist.ac.kr

More information

Page 1. Outline / Computer Networking : 1 st Generation Commercial PC/Packet Video Technologies

Page 1. Outline / Computer Networking : 1 st Generation Commercial PC/Packet Video Technologies Outline 15-441/15-641 Computer Networking Lecture 18 Internet Video Delivery Peter Steenkiste Slides by Professor Hui Zhang Background Technologies: - HTTP download - Real-time streaming - HTTP streaming

More information

Skype Video Responsiveness to Bandwidth Variations

Skype Video Responsiveness to Bandwidth Variations Skype Video Responsiveness to Bandwidth Variations L. De Cicco,, V. Palmisano Dipartimento di Elettronica ed Elettrotecnica Politecnico di Bari Italy -1- Motivation 1/2 Multimedia real-time applications

More information

Content-Aware Personalised Rate Adaptation for Adaptive Streaming via Deep Video Analysis

Content-Aware Personalised Rate Adaptation for Adaptive Streaming via Deep Video Analysis Content-Aware Personalised Rate Adaptation for Adaptive Streaming via Deep Video Analysis Guanyu Gao, Linsen Dong, Huaizheng Zhang, Yonggang Wen, and Wenjun Zeng 2 Nanyang Technological University, Singapore

More information

Quality Selection for Dynamic Adaptive Streaming over HTTP with Scalable Video Coding

Quality Selection for Dynamic Adaptive Streaming over HTTP with Scalable Video Coding Quality Selection for Dynamic Adaptive Streaming over HTTP with Scalable Video Coding Travis Andelin travis.andelin@gmail.com Sean Warnick sean@cs.byu.edu Vasu Chetty chettyv@byu.edu Daniel Zappala zappala@cs.byu.edu

More information

Delivery of Adaptive Bit Rate Video: Balancing Fairness, Efficiency and Quality

Delivery of Adaptive Bit Rate Video: Balancing Fairness, Efficiency and Quality Delivery of Adaptive Bit Rate Video: Balancing Fairness, Efficiency and Quality Jason J. Quinlan, Ahmed H. Zahran, K.K. Ramakrishnan, Cormac J. Sreenan Dept. of Computer Science, University College Cork,

More information

Internet Video Delivery. Professor Hui Zhang

Internet Video Delivery. Professor Hui Zhang 18-345 Internet Video Delivery Professor Hui Zhang 1 1990 2004: 1 st Generation Commercial PC/Packet Video Technologies Simple video playback, no support for rich app Not well integrated with Web browser

More information

Machine Learning Concepts for Adaptive Video Streaming Systems (SS17)

Machine Learning Concepts for Adaptive Video Streaming Systems (SS17) Machine Learning Concepts for Adaptive Video Streaming Systems (SS17) Kickoff Denny Stohr, M.Sc. Jeremias Blendin Prof. Dr.-Ing. Wolfgang Effelsberg www.dms.informatik.tu-darmstadt.de 21.04.17 Department

More information

A Case Study on Cloud Based Hybrid Adaptive Mobile Streaming: Performance Evaluation

A Case Study on Cloud Based Hybrid Adaptive Mobile Streaming: Performance Evaluation A Case Study on Cloud Based Hybrid Adaptive Mobile Streaming: Performance Evaluation T. Mahesh kumar 1, Dr. k. Santhisree 2, M. Bharat 3, V. Pruthvi Chaithanya Varshu 4 Student member of IEEE, M. tech

More information

SVC-based Multi-user Streamloading for Wireless Networks

SVC-based Multi-user Streamloading for Wireless Networks 1 SVC-based Multi-user Streamloading for Wireless Networks S. Amir Hosseini 1, Zheng Lu 2, Gustavo de Veciana 2, and Shivendra S. Panwar 1 1 Department of Electrical and Computer Engineering, NYU Tandon

More information

Reduction of Periodic Broadcast Resource Requirements with Proxy Caching

Reduction of Periodic Broadcast Resource Requirements with Proxy Caching Reduction of Periodic Broadcast Resource Requirements with Proxy Caching Ewa Kusmierek and David H.C. Du Digital Technology Center and Department of Computer Science and Engineering University of Minnesota

More information

On the Impact of Redirection on HTTP Adaptive Streaming Services in Federated CDNs

On the Impact of Redirection on HTTP Adaptive Streaming Services in Federated CDNs On the Impact of Redirection on HTTP Adaptive Streaming Services in Federated CDNs Jeroen Famaey 1,StevenLatré 1, Ray van Brandenburg 2, M. Oskar van Deventer 2, and Filip De Turck 1 1 Ghent University

More information

Markov Decision Processes for Bitrate Harmony in Adaptive Video Streaming

Markov Decision Processes for Bitrate Harmony in Adaptive Video Streaming Markov Decision Processes for Bitrate Harmony in Adaptive Video Streaming Koffka Khan, Wayne Goodridge The University of the West Indies Email: koffka.khan@gmail.com, wayne.goodridge@sta.uwi.edu Abstract

More information

Adaptive Video Acceleration. White Paper. 1 P a g e

Adaptive Video Acceleration. White Paper. 1 P a g e Adaptive Video Acceleration White Paper 1 P a g e Version 1.0 Veronique Phan Dir. Technical Sales July 16 th 2014 2 P a g e 1. Preface Giraffic is the enabler of Next Generation Internet TV broadcast technology

More information

Delivering Stable High-Quality Video: An SDN Architecture with DASH Assisting Network Elements

Delivering Stable High-Quality Video: An SDN Architecture with DASH Assisting Network Elements Delivering Stable High-Quality Video: An SDN Architecture with DASH Assisting Network Elements Jan Willem Kleinrouweler 1, Sergio Cabrero 1, Pablo Cesar 1, 1 CWI: Centrum Wiskunde & Informatica Delft University

More information

Performance Characterization of a Commercial Video Streaming Service

Performance Characterization of a Commercial Video Streaming Service Performance Characterization of a Commercial Video Streaming Service Mojgan Ghasemi, Princeton University P. Kanuparthy, 1 A. Mansy, 1 T. Benson, 2 J. Rexford 3 1 Yahoo, 2 Duke University, 3 Princeton

More information

irtc: Live Broadcasting

irtc: Live Broadcasting 1 irtc: Live Broadcasting Delivering ultra-low-latency media at massive scale with LiveSwitch and WebRTC Introduction In the early days of the Internet and personal computing, it wasn t uncommon to wait

More information

Lecture 27 DASH (Dynamic Adaptive Streaming over HTTP)

Lecture 27 DASH (Dynamic Adaptive Streaming over HTTP) CS 414 Multimedia Systems Design Lecture 27 DASH (Dynamic Adaptive Streaming over HTTP) Klara Nahrstedt Spring 2012 Administrative MP2 posted MP2 Deadline April 7, Saturday, 5pm. APPLICATION Internet Multimedia

More information

MOOC-DASH: A DASH System for Delivering High-Quality MOOCs Videos

MOOC-DASH: A DASH System for Delivering High-Quality MOOCs Videos MOOC-DASH: A DASH System for Delivering High-Quality MOOCs Videos Yi Wang, Wenjun Wu, Yihua Lou National Laboratory of Software Development Environment School of Computer Science and Engineering, Beihang

More information

Synthesizing Adaptive Protocols by Selective Enumeration (SYNAPSE)

Synthesizing Adaptive Protocols by Selective Enumeration (SYNAPSE) Synthesizing Adaptive Protocols by Selective Enumeration (SYNAPSE) Problem Definition Solution Approach Benefits to End User Talk Overview Metrics Summary of Results to Date Lessons Learned & Future Work

More information

ABR Streaming of VBR-encoded Videos: Characterization, Challenges, and Solutions

ABR Streaming of VBR-encoded Videos: Characterization, Challenges, and Solutions ABR Streaming of VBR-encoded Videos: Characterization, Challenges, and Solutions Yanyuan Qin, Shuai Hao 2, K.R. Pattipati, Feng Qian 3, Subhabrata Sen 2, Bing Wang and Chaoqun Yue University of Connecticut

More information

arxiv: v1 [cs.ni] 27 Sep 2018

arxiv: v1 [cs.ni] 27 Sep 2018 Cross-Layer Effects on Training Neural Algorithms for Video Streaming Pablo Gil Pereira Saarland Informatics Campus Saarbrücken 66123 gilpereira@cs.uni-saarland.de Andreas Schmidt Saarland Informatics

More information

Neural Adaptive Content-aware Internet Video Delivery

Neural Adaptive Content-aware Internet Video Delivery Neural Adaptive Content-aware Internet Video Delivery Hyunho Yeo, Youngmok Jung, Jaehong Kim, Jinwoo Shin, and Dongsu Han, KAIST https://www.usenix.org/conference/osdi8/presentation/yeo This paper is included

More information

QoE-driven Dynamic Adaptive Video Streaming Strategy with Future Information

QoE-driven Dynamic Adaptive Video Streaming Strategy with Future Information 1 QoE-driven Dynamic Adaptive Video Streaming Strategy with Future Information Li YU, Student Member, IEEE, Tammam TILLO, Senior Member, IEEE, and Jimin XIAO, Member, IEEE Abstract Dynamic Adaptive Video

More information

Improving Mobile Video Quality Through Predictive Channel Quality Based Buffering

Improving Mobile Video Quality Through Predictive Channel Quality Based Buffering Improving Mobile Video Quality Through Predictive Channel Quality Based Buffering Jan Willem Kleinrouweler Centrum Wiskunde & Informatica j.w.m.kleinrouweler@cwi.nl Hans van den Berg TNO, Centrum Wiskunde

More information

Achieving Low-Latency Streaming At Scale

Achieving Low-Latency Streaming At Scale Achieving Low-Latency Streaming At Scale Founded in 2005, Wowza offers a complete portfolio to power today s video streaming ecosystem from encoding to delivery. Wowza provides both software and managed

More information

Cache Replacement Strategies for Scalable Video Streaming in CCN

Cache Replacement Strategies for Scalable Video Streaming in CCN Cache Replacement Strategies for Scalable Video Streaming in CCN Junghwan Lee, Kyubo Lim, and Chuck Yoo Dept. Computer Science and Engineering Korea University Seoul, Korea {jhlee, kblim, chuck}@os.korea.ac.kr

More information

Performance of UMTS Radio Link Control

Performance of UMTS Radio Link Control Performance of UMTS Radio Link Control Qinqing Zhang, Hsuan-Jung Su Bell Laboratories, Lucent Technologies Holmdel, NJ 77 Abstract- The Radio Link Control (RLC) protocol in Universal Mobile Telecommunication

More information

GreenBag: Energy-efficient Bandwidth Aggregation For Real-time Streaming in Heterogeneous Mobile Wireless Networks

GreenBag: Energy-efficient Bandwidth Aggregation For Real-time Streaming in Heterogeneous Mobile Wireless Networks GreenBag: Energy-efficient Bandwidth Aggregation For Real-time Streaming in Heterogeneous Mobile Wireless Networks Duc Hoang Bui, Kilho Lee, Sangeun Oh, Insik Shin Dept. of Computer Science KAIST, South

More information

A Proxy-assisted DASH Live Streaming Scheme

A Proxy-assisted DASH Live Streaming Scheme A Proxy-assisted DASH Live Streaming Scheme Cheng Zhao 1, Li Song 1,2, Da Huo 1, Rong Xie 1,2, Nam Ling 3 1 Institute of Image Communication and Network Engineering, Shanghai Jiao Tong University 2 Cooperative

More information

CS2P: Improving Video Bitrate Selection and Adaptation with Data-Driven Throughput Prediction

CS2P: Improving Video Bitrate Selection and Adaptation with Data-Driven Throughput Prediction CS2P: Improving Video Bitrate Selection and Adaptation with Data-Driven Throughput Prediction Yi Sun, Xiaoqi Yin, Junchen Jiang, Vyas Sekar Fuyuan Lin, Nanshu Wang, Tao Liu, Bruno Sinopoli ICT/CAS, CMU,

More information

SINGLE PASS DEPENDENT BIT ALLOCATION FOR SPATIAL SCALABILITY CODING OF H.264/SVC

SINGLE PASS DEPENDENT BIT ALLOCATION FOR SPATIAL SCALABILITY CODING OF H.264/SVC SINGLE PASS DEPENDENT BIT ALLOCATION FOR SPATIAL SCALABILITY CODING OF H.264/SVC Randa Atta, Rehab F. Abdel-Kader, and Amera Abd-AlRahem Electrical Engineering Department, Faculty of Engineering, Port

More information

Combining DASH with MPTCP for video streaming

Combining DASH with MPTCP for video streaming Combining DASH with MPTCP for video streaming Dissertation presented by Lena Victoria PESCHKE for obtaining the Master s degree in Computer Science and Engineering Option: Networking and security Supervisor

More information

A Web-Based Framework for Fast Synchronization of Live Video Players

A Web-Based Framework for Fast Synchronization of Live Video Players A Web-Based Framework for Fast Synchronization of Live Video Players Dries Pauwels, Jeroen van der Hooft, Stefano Petrangeli, Tim Wauters, Danny De Vleeschauwer, and Filip De Turck Ghent University - iminds,

More information

An HTTP/2 Push-Based Approach for Low-Latency Live Streaming with Super-Short Segments

An HTTP/2 Push-Based Approach for Low-Latency Live Streaming with Super-Short Segments Journal on Network and Systems Management manuscript No. (will be inserted by the editor) An HTTP/2 Push-Based Approach for Low-Latency Live Streaming with Super-Short Segments Jeroen van der Hooft Stefano

More information

MASH: A Rate Adaptation Algorithm for Multiview Video Streaming over HTTP

MASH: A Rate Adaptation Algorithm for Multiview Video Streaming over HTTP MASH: A Rate Adaptation Algorithm for Multiview Video Streaming over HTTP Khaled Diab School of Computing Science Simon Fraser University Burnaby, BC, Canada Mohamed Hefeeda School of Computing Science

More information

NOWADAYS, video is the most dominant application in

NOWADAYS, video is the most dominant application in IEEE COMMUNICATION SURVEYS & TUTORIALS, VOL. 17, NO. 1, FIRST QUARTER 2015 469 A Survey on Quality of Experience of HTTP Adaptive Streaming Michael Seufert, Sebastian Egger, Martin Slanina, Thomas Zinner,

More information

Dynamic Adaptive Streaming over HTTP (DASH) Application Protocol : Modeling and Analysis

Dynamic Adaptive Streaming over HTTP (DASH) Application Protocol : Modeling and Analysis Dynamic Adaptive Streaming over HTTP (DASH) Application Protocol : Modeling and Analysis Dr. Jim Martin Associate Professor School of Computing Clemson University jim.martin@cs.clemson.edu http://www.cs.clemson.edu/~jmarty

More information

ADAPTIVE STREAMING. Improve Retention for Live Content. Copyright (415)

ADAPTIVE STREAMING. Improve Retention for Live Content. Copyright (415) ADAPTIVE STREAMING Improve Retention for Live Content A daptive streaming technologies make multiple video streams available to the end viewer. True adaptive bitrate dynamically switches between qualities

More information

Performance Characterization of a Commercial Video Streaming Service. Mojgan Ghasemi, Akamai Technologies - Princeton University

Performance Characterization of a Commercial Video Streaming Service. Mojgan Ghasemi, Akamai Technologies - Princeton University Performance Characterization of a Commercial Video Streaming Service Mojgan Ghasemi, Akamai Technologies - Princeton University MGhasemi,PKanuparthy,AMansy,TBenson,andJRexford ACM IMC 2016 1 2 First study

More information

TCP Protocol Optimization for HTTP Adaptive Streaming

TCP Protocol Optimization for HTTP Adaptive Streaming TCP Protocol Optimization for HTTP Adaptive Streaming Soutenance de thèse Chiheb Ben Ameur Salle Aurigny, IRISA, Rennes 17 décembre 2015 Institut de Recherche en Informatique et Systèmes Aléatoires HTTP

More information

SERVICE PROVISIONING AND PROFIT MAXIMIZATION IN NETWORK-ASSISTED ADAPTIVE HTTP STREAMING

SERVICE PROVISIONING AND PROFIT MAXIMIZATION IN NETWORK-ASSISTED ADAPTIVE HTTP STREAMING SERVICE PROVISIONING AND PROFIT MAXIMIZATION IN NETWORK-ASSISTED ADAPTIVE HTTP STREAMING Zhisheng Yan 1, Cedric Westphal 2,3, Xin Wang 2, and Chang Wen Chen 1 1 State University of New York at Buffalo,

More information

A Framework for Video Streaming to Resource- Constrained Terminals

A Framework for Video Streaming to Resource- Constrained Terminals A Framework for Video Streaming to Resource- Constrained Terminals Dmitri Jarnikov 1, Johan Lukkien 1, Peter van der Stok 1 Dept. of Mathematics and Computer Science, Eindhoven University of Technology

More information

One-pass bitrate control for MPEG-4 Scalable Video Coding using ρ-domain

One-pass bitrate control for MPEG-4 Scalable Video Coding using ρ-domain Author manuscript, published in "International Symposium on Broadband Multimedia Systems and Broadcasting, Bilbao : Spain (2009)" One-pass bitrate control for MPEG-4 Scalable Video Coding using ρ-domain

More information

Dynamic Deferred Acknowledgment Mechanism for Improving the Performance of TCP in Multi-Hop Wireless Networks

Dynamic Deferred Acknowledgment Mechanism for Improving the Performance of TCP in Multi-Hop Wireless Networks Dynamic Deferred Acknowledgment Mechanism for Improving the Performance of TCP in Multi-Hop Wireless Networks Dodda Sunitha Dr.A.Nagaraju Dr. G.Narsimha Assistant Professor of IT Dept. Central University

More information

Quality-Assured Energy Balancing for Multi-hop Wireless Multimedia Networks via 2-D Channel Coding Rate Allocation

Quality-Assured Energy Balancing for Multi-hop Wireless Multimedia Networks via 2-D Channel Coding Rate Allocation Quality-Assured Energy Balancing for Multi-hop Wireless Multimedia Networks via 2-D Channel Coding Rate Allocation Lin Xing, Wei Wang, Gensheng Zhang Electrical Engineering and Computer Science, South

More information

Analysis of Optimum Substream Lengths for Dual TCP/UDP Streaming of HD H.264 Video Over Congested WLANs

Analysis of Optimum Substream Lengths for Dual TCP/UDP Streaming of HD H.264 Video Over Congested WLANs Analysis of Optimum Substream Lengths for Dual TCP/UDP Streaming of HD H.264 Video Over Congested WLANs Arul Dhamodaran, Kevin Gatimu, and Ben Lee School of Electrical Engineering and Computer Science

More information

Transmission algorithm for video streaming over cellular networks

Transmission algorithm for video streaming over cellular networks Transmission algorithm for video streaming over cellular networks Y. Falik 1, A. Averbuch 1, U. Yechiali 2 1 School of Computer Science, Tel Aviv University Tel Aviv 69978, Israel 2 Department of Statistics

More information

Bandwidth Optimization On Design Of Visual Display Information System Based Networking At Politeknik Negeri Bali

Bandwidth Optimization On Design Of Visual Display Information System Based Networking At Politeknik Negeri Bali Journal of Physics: Conference Series PAPER OPEN ACCESS Bandwidth Optimization On Design Of Visual Display Information System Based Networking At Politeknik Negeri Bali To cite this article: IKG Sudiartha

More information

EXPLORING THE DELAY VERSUS QUALITY TRADEOFF IN REAL-TIME STREAMING OF SCALABLE VIDEO FROM MOBILE DEVICES

EXPLORING THE DELAY VERSUS QUALITY TRADEOFF IN REAL-TIME STREAMING OF SCALABLE VIDEO FROM MOBILE DEVICES EXPLORING THE DELAY VERSUS QUALITY TRADEOFF IN REAL-TIME STREAMING OF SCALABLE VIDEO FROM MOBILE DEVICES Matti Siekkinen, Alberto Barraja,, Jukka K. Nurminen, Enrico Masala Department of Computer Science,

More information

Multimedia Systems Project 3

Multimedia Systems Project 3 Effectiveness of TCP for Video Transport 1. Introduction In this project, we evaluate the effectiveness of TCP for transferring video applications that require real-time guarantees. Today s video applications

More information

Data Driven Networks

Data Driven Networks Data Driven Networks Is it possible for to Learn the control planes of networks and applications? Operators specify what they want, and the system learns how to deliver CAN WE LEARN THE CONTROL PLANE OF

More information

Insights into PPLive: A Measurement Study of a Large-Scale P2P IPTV System

Insights into PPLive: A Measurement Study of a Large-Scale P2P IPTV System Insights into PPLive: A Measurement Study of a Large-Scale P2P IPTV System Xiaojun Hei, Chao Liang, Jian Liang, Yong Liu and Keith W. Ross Department of Computer and Information Science Department of Electrical

More information

Quest for High-Performance Bufferless NoCs with Single-Cycle Express Paths and Self-Learning Throttling

Quest for High-Performance Bufferless NoCs with Single-Cycle Express Paths and Self-Learning Throttling Quest for High-Performance Bufferless NoCs with Single-Cycle Express Paths and Self-Learning Throttling Bhavya K. Daya, Li-Shiuan Peh, Anantha P. Chandrakasan Dept. of Electrical Engineering and Computer

More information