Network Aware Adaptive Media Streaming in Mobile Cloud

Network Aware Adaptive Media Streaming in Mobile Cloud Authors Name/s per 1st Affiliation (Author) line 1 (of Affiliation): dept. name of organization line 2-name of organization, acronyms acceptable line 3-City, Country line 4-e-mail address if desired Authors Name/s per 2nd Affiliation (Author) line 1 (of Affiliation): dept. name of organization line 2-name of organization, acronyms acceptable line 3-City, Country line 4-e-mail address if desired Abstract—This electronic document is a â€Å"live† template and already defines the components of your paper [title, text, heads, etc.] in its style sheet. *CRITICAL: Do Not Use Symbols, Special Characters, or Math in Paper Title or Abstract. (Abstract) Keywords—multimedia streaming; mobile cloud; ranking; QoS I. Introduction In recent years, mobility of computing devices has caught the fascination and attention of many users all over the world. This has led to rapid advancement in mobile technology and now users can easily stream high quality multimedia content like audio and video on the go. A huge limitation to this, however, is the loss of quality that is incurred while transferring the data. Due to the mobile nature of the devices, varying signal strength can lead to packet loss which ultimately leads to the reduction in the quality of service (QoS). In addition, the memory available in mobile devices is relatively low. To overcome these constraints, data is stored and retrieved from a cloud. Cloud computing addresses the QoS related issues and reliability problems. The cloud has a large amount of storage space and computation power. Harnessing the power of the cloud, it will be possible to service the needs of multiple mobile clients simultaneously. Using the cloud, it is possible to allocate resources on demand and reallocate them dynamically. In order to stream data from a cloud to a mobile device, a coding and decoding architecture like H264/SVC is necessary. This architecture is an extension of the H.264/AVC. It ensures that the same quality of video that can be obtained using H.264/MPEG-4 AVC design on the mobile device. It employs spatial scalability and temporal scalability. According to spatial scalability samples of high quality data can be predicted from their decoded low quality counterparts. Using temporal scalability, the entire video is modelled in such a way that the motion is encoded as dependencies so that the picture for subsequent frames need not be encoded directly. In order to boost QoS, a technique called Bayesian-Gaussian method is used to predict the bandwidth available to the mobile user. Once the bandwidth has been predicted, the data is encoded using xuggler transcoding algorithm. To finally stream the video, multipath routing protocols are used and ranks are provided to each node to ensure that none of them have to wait indefinitely to be serviced. Following this, a comparison is made to the existing Bayesian technique proposed by Keshav[1]. II. Related Work A. Mobile Cloud Compuing A mobile cloud computing setup is one in which mobile devices outsource the computational power of the cloud. Data storage and processing are both performed outside the mobile device. B. Streaming Content C. Role of Cache in Real Time Streaming The role of the cache has been outlined by Wu et al[2]. When a Real Time Streaming Protocol (RTSP) request is sent by a machine, the cache memory is initially searched. In case a cache miss occurs, the original server services the request. D. Improving Quality of Service A number of different approaches have been proposed in order to ensure that the quality of service is maximized. One such method presented by Wang and Dey[3] uses a technique that varies the complexity of the content depending on the network. Non-essential data in a scene are omitted to accomplish this. Lai et al[4] have also put forth an approach to data streaming that depends on the network. Prediction of the bandwidth is done based on measured historical data. This will help prevent the wastage of bandwidth. It is also noted that the video format to be used is to be chosen. This is performed by a Bayesian prediction module. A third approach is detailed by Thuy An et al[5]. Enhancements are made to the Remote Desktop Protocol (RDP) in order to provide an overall better experience. The data separated into two categories and compressed. Lossless techniques are used to provide the best possible output. E. Ranking The various approaches mentioned in the previous section discuss improving QoS with respect to one user. But in reality, the cloud is simultaneously accessed by more than just one user. For this reason, it is important to ensure that there is some scheduling mechanism in place that will monitor the incoming requests so that no client request is forced to wait for too long without being serviced. Zhou et al[6] have proposed a novel approach in which all the competing mobile devices work together to minimize congestion. This approach aims to strike a balance between reducing the distortion in data and increasing the performance of the network as a whole. III. Proposed Work The proposed model has two major components: the mobile device and the cloud. The mobile device simply issues the request while the cloud provides a rank, predicts the bandwidth and then streams the video accordingly. The architecture has been outlined in Fig.1. A. Mobile Device The implementation of the mobile portion of the architecture is fairly straight forward. The user is provided with the option to specify the location of the video in the cloud server. Then, the cache is check to see if the requested data is available. If it is, the data is transferred directly from the cache. This type of cached data will be accessible offline as well. In the case where a cache miss occurs, the server is accessed to retrieve the data. B. The Cloud The videos that are to be stream are stored in a separate database. When a request is made, the video is streamed using the cloud. In order to do this, three major modules are implemented in the cloud. In the cloud, the users are ranked and then the bandwidth available is estimated. Finally, xuggler transcoding is used to encode the data and the encoded data is transferred to the mobile device for viewing. Each operation is handled by a different module as show in Fig.2. C. Bandwidth Prediction D. Xuggler Transcoding E. Ranking The ranking module is used to ensure that QoS is improved while transmitting the data. Once the bandwidth has been determined, the data has to be sent in such a way that the congestion in the network is as low as possible. Ranking is done based on the user profile. The user profile contains a history of the user’s downloads as well as the bandwidth measured. Poorly performing nodes in the system are identified using this ranking system and they can be enhanced to improve the overall functioning of the network as a whole. F. Channel Assignment Once ranking is done, multipath routing algorithms are used to transfer the data. Link states are determined and the several feasible paths are selected. Since several paths are selected, the chances of congestion and packet loss are reduced. The most suitable channel for transmission of the data can be determined by solving the linear programming equation Min (1) The process of selecting the best channel is shown in Fig. 3. IV. Output The proposed system has been implemented and its results have been compared with that of the Keshav’s Bayesian technique. It can be seen that the proposed system works better than Keshav’s system consistently. Comparative studies have been undertaken on the basis of bandwidth and peak signal to noise ratio (PSNR). A. Bandwidth The bandwidth predicted by the proposed system is a lot closer to the actual measured bandwidth than that predicted by Keshav’s system. The graph in Fig.4 clearly shows the deviation of both techniques from the actual measured bandwidth. B. PSNR and Bit Rate The quality of the video streamed can be determined based on the bit rate as well as the PSNR. The proposed system performs better than Keshav’s system on both counts. This is shown in the graph in Fig.5. C. Video Quality The comparative study only shows us how the system works in comparison to Keshav’s existing system. To determine the effectiveness of this system, a detailed study of the video quality was performed and has been summarized in Table 1. Conclusion It is clear from the studies undertaken that the proposed Bayesian-Gaussian technique works well at predicting the bandwidth available. The xuggler transcoding also ensures that quality is preserved. Thus, using a mobile cloud it is possible to stream videos without a loss in quality and also without forcing the user to wait for the video to load. Acknowledgment (Heading 5) The preferred spelling of the word â€Å"acknowledgment† in America is without an â€Å"e† after the â€Å"g.† Avoid the stilted expression â€Å"one of us (R. B. G.) thanks †. Instead, try â€Å"R. B. G. thanks†. Put sponsor acknowledgments in the unnumbered footnote on the first page. References The template will number citations consecutively within brackets [1]. The sentence punctuation follows the bracket [2]. Refer simply to the reference number, as in [3]—do not use â€Å"Ref. [3]† or â€Å"reference [3]† except at the beginning of a sentence: â€Å"Reference [3] was the first † Number footnotes separately in superscripts. Place the actual footnote at the bottom of the column in which it was cited. Do not put footnotes in the reference list. Use letters for table footnotes. Unless there are six authors or more give all authors’ names; do not use â€Å"et al.†. Papers that have not been published, even if they have been submitted for publication, should be cited as â€Å"unpublished† [4]. Papers that have been accepted for publication should be cited as â€Å"in press† [5]. Capitalize only the first word in a paper title, except for proper nouns and element symbols. For papers published in translation journals, please give the English citation first, followed by the original foreign-language citation [6]. G. Eason, B. Noble, and I.N. Sneddon, â€Å"On certain integrals of Lipschitz-Hankel type involving products of Bessel functions,† Phil. Trans. Roy. Soc. London, vol. A247, pp. 529-551, April 1955. (references) J. Clerk Maxwell, A Treatise on Electricity and Magnetism, 3rd ed., vol. 2. Oxford: Clarendon, 1892, pp.68-73. I.S. Jacobs and C.P. Bean, â€Å"Fine particles, thin films and exchange anisotropy,† in Magnetism, vol. III, G.T. Rado and H. Suhl, Eds. New York: Academic, 1963, pp. 271-350. K. Elissa, â€Å"Title of paper if known,† unpublished. R. Nicole, â€Å"Title of paper with only first word capitalized,† J. Name Stand. Abbrev., in press. Y. Yorozu, M. Hirano, K. Oka, and Y. Tagawa, â€Å"Electron spectroscopy studies on magneto-optical media and plastic substrate interface,† IEEE Transl. J. Magn. Japan, vol. 2, pp. 740-741, August 1987 [Digests 9th Annual Conf. Magnetics Japan, p. 301, 1982]. M. Young, The Technical Writer’s Handbook. Mill Valley, CA: University Science, 1989.