An Efficient Video Streaming Architecture With QoS Control for Virtual Desktop Infrastructure

Abstract and 1 Introduction 2 Related Work 3 Image Compression Module 4 System Architecture 5 Experimental Results 6 Conclusions A Compression Results References Abstract In virtual desktop infrastructure (VDI) environments, the remote display protocol has a big responsibility: to transmit video data from a data center-hosted desktop to the endpoint. The protocol must ensure a high level of client-perceived end-to-end quality of service (QoS) under heavy load conditions. Each remote display protocol works differently depending on the network and which applications are being delivered. In health care applications, doctors and nurses can use mobile devices directly to monitor patients. Moreover, the ability to implement tasks requiring high consumption of CPU and other resources is applicable to a variety of applications including research and cloud gaming. Such computer games and complex processes will run on powerful cloud servers and the screen contents will be transmitted to the client. To enable such applications, remote display technology requires further enhancements to meet more stringent requirements on bandwidth and QoS, and to allow real-time operation. In this paper, we present an architecture including flexible QoS control to improve the users’ quality of experience (QoE). The QoS control includes linear regression modeling of historical network data as a means to consider the network condition. Additionally, the architecture includes a novel compression algorithm of 2D images, designed to guarantee the best image quality and to reduce video delay; this algorithm is based on k-means clustering and can satisfy the requirements of real-time onboard processing. Through simulations with a real-world dataset collected by the MIT Computer Science and Artificial Intelligence Lab, we represent experimental as well as explain the performance of QoS system. Keywords Video streaming · Mobile thin-client · K-means clustering · Quality of service · Remote protocol · Cloud computing Introduction Cloud computing has become a hot topic in todays world, largely due to the flexible and cost savings it can deliver. In the same manner as with visualization, cloud computing implementations started with server infrastructure and has moved to the desktop; the cloud is now ripe for desktop infrastructure. By moving desktops to the cloud rather than an internally deployed and managed VDI solution in a data center, businesses can realise all of the promised benefits of virtual desktops [1]. Furthermore, the Internet has rapidly emerged as a mechanism for users to find and retrieve content, originally for webpages and recently for streaming media. Moreover, mobile phones and tablets or any thin client devices can run powerful applications from anywhere and at any time via the cloud. Therefore, there are a lot of case studies needed to transmit data from the virtual desktop in the data cloud center to the screen endpoints. he doctors tired all the time to check a patient’s status frequently, included their heart rate or other medical signs read by specialized equipment. With VDI solutions, a patient can get convenient health care from the comfort of their own home. This allows doctors to quickly and conveniently make such information available remotely on the mobile devices which are connected to a server by using a remote desktop protocol [1]. Moreover, VDI helps save valuable time and satisfies the demand for personal control, while also reducing the cost of long-term medical care by avoiding waste of human resources on manual processing [2]. Because any mistakes is always leading cause of death, so new developments in health care are of considerable interest. Computer gaming is a very different application where remote display technology will also be useful. Gamers spent a lot of money on computer games, hardware, and accessories. Traditionally, computer games are delivered either in boxes or via Internet downloads. These games have to install the computer games on physical machines to play them. The installation process becomes extremely tedious because the games are too complicated and the computer hardware and system software are very fragmented. Take Blizzard’s Star-craft II as example, it may take more than an hour to install it on an i5 PC, and another hour to apply the online patches. Furthermore, gamers may find their computers are not powerful enough to enable all the visual effects yet achieve high frame rates. So, gamers have to repeatedly upgrade their computers so as to play the latest computer games. Hence, cloud gaming is a better way to deliver high-quality gaming experience and opens new business opportunity [3]. In a cloud gaming system, computer games run on powerful cloud servers and streaming video data to thin clients by using remote protocols [4]. The thin clients are lightweight and can be ported to resource-constrained platforms, such as mobile devices and TV set-top boxes. With cloud gaming, the key benefit is that gamers can play the latest computer games anywhere and anytime, while the game developers can optimize their games for a specific PC configuration. Laboratory research is another excellent potential application for cloud-based remote display technology. By using remote desktop protocols, researchers perform complex tasks requiring intensive computation resources or specialized software. In addition to these example applications, there are other fields in which using remote display protocols would be useful for transmitting screen data from a cloud server to a client. However, in the past there have been several challenges with deploying VDI. Main problems included the network condition, server overload, bandwidth and limitation of mobile device (hardware resource, storage, battery life time). Specifically, a client sends an input event request to a server, the sever receives and deploys several offloading computing tasks, then renders screen image, encodes, and finally transfers the results to the client (as shown in Fig 1). These graphical images will be decoded and displayed on the client side. Therefore, it is so importance to guarantee client-perceived end-to-end quality of service. We can summarize several important factors to measure performance of the system: – Network condition: latency and bandwidth utilization. Latency should be as small as possible. Furthermore, with many different bandwidth quotas, users always want to get the highest acceptable performance. – Hardware resource consumption: CPU has a critical meaning in the processing because the limited hardware resource of the client device and overhead consumption on the server. – User demands: Cross-platform adaptation because it is a great convenience for users who use many kinds of devices such as laptop, smart phone, tablet, etc. with different operating systems. Above factors directly affect QoE (Quality of Experience) that user perceives in each request which is made from local desktop, where claims for more high fidelity of display. Many solutions have been developed and reported for improving QoE in content delivery networks; typically, solutions include hardware and software improvements. The general idea is that whatever graphical elements are painted to the screen on the host and then scraped by the protocol interface, are sent down to the client. This can happen in two ways: – The client can contact the server and pull a new snapshot of the screen from the frame buffer. This is how Virtual network computing (VNC) works [5], [7]. – The server can continuously push its screen activity to the client. This can be at the frame buffer level, the GDI / window manager level, or a combination of both. This is how (Remote Desktop Protocol) RDP [8] and Independent Computing Architecture work [9]. ith flexible QoS control based on network condition analysis to improve users’ QoE. We carefully establish the best compression parameters based on statistical analysis of historical data on network conditions. Moreover, to improve the real-time video streaming quality we develop a novel compression algorithm for 2D images, based on k-means clustering. The contributions of this work can be summarized as follows: 1. We develop a novel image compression algorithm based on k-means clustering which can be applied in real-world case studies. 2. We present an efficient architecture including QoS control based on analysis of historical data to account for network conditions. This system applies multiple linear regression to make optimal decisions for the QoS policy. We evaluate the proposed image compression algorithm by means of simulations in Matlab, using as input a dataset of real images collected by the MIT Computer Science and Artificial Intelligence Lab [36]. We demonstrate the QoS control system in a testbed cloud environment and analyze its performance. The rest of this paper is organized as follows. Related work is reviewed in Section 2. Section 3 gives an image compression algorithm based on the kmeans algorithm. Section 4 suggests an architecture to dynamically modify the compression rate of a video stream based on historical data of the recipients bandwidth. Section 5 describes experiments with the proposed QoS control and evaluates its performance. Finally, Section 6 gives conclusions. This paper is available on arxiv under CC BY 4.0 DEED license. Authors: (1) Huu-Quoc Nguyen, Department of Computer Engineering, Kyung Hee University, 1 Seocheon, Giheung, Yongin, Gyeonggi, South Korea and quoc@khu.ac.kr; (2) Tien-Dung Nguyen, Department of Computer Engineering, Kyung Hee University, 1 Seocheon, Giheung, Yongin, Gyeonggi, South Korea and ntiendung@khu.ac.kr; (3) Van-Nam Pham, Department of Computer Engineering, Kyung Hee University, 1 Seocheon, Giheung, Yongin, Gyeonggi, South Korea and nampv@khu.ac.kr; (4) Xuan-Qui Pham, Department of Computer Engineering, Kyung Hee University, 1 Seocheon, Giheung, Yongin, Gyeonggi, South Korea and pxuanqui@khu.ac.kr; (5) Quang-Thai Ngo, Department of Computer Engineering, Kyung Hee University, 1 Seocheon, Giheung, Yongin, Gyeonggi, South Korea and nqthai@khu.ac.kr; (6) Eui-Nam Huh, Department of Computer Engineering, Kyung Hee University, 1 Seocheon, Giheung, Yongin, Gyeonggi, South Korea and johnhuh@khu.ac.kr. Abstract and 1 Introduction Abstract and 1 Introduction 2 Related Work 2 Related Work 3 Image Compression Module 3 Image Compression Module 4 System Architecture 4 System Architecture 5 Experimental Results 5 Experimental Results 6 Conclusions 6 Conclusions A Compression Results A Compression Results References References Abstract In virtual desktop infrastructure (VDI) environments, the remote display protocol has a big responsibility: to transmit video data from a data center-hosted desktop to the endpoint. The protocol must ensure a high level of client-perceived end-to-end quality of service (QoS) under heavy load conditions. Each remote display protocol works differently depending on the network and which applications are being delivered. In health care applications, doctors and nurses can use mobile devices directly to monitor patients. Moreover, the ability to implement tasks requiring high consumption of CPU and other resources is applicable to a variety of applications including research and cloud gaming. Such computer games and complex processes will run on powerful cloud servers and the screen contents will be transmitted to the client. To enable such applications, remote display technology requires further enhancements to meet more stringent requirements on bandwidth and QoS, and to allow real-time operation. In this paper, we present an architecture including flexible QoS control to improve the users’ quality of experience (QoE). The QoS control includes linear regression modeling of historical network data as a means to consider the network condition. Additionally, the architecture includes a novel compression algorithm of 2D images, designed to guarantee the best image quality and to reduce video delay; this algorithm is based on k-means clustering and can satisfy the requirements of real-time onboard processing. Through simulations with a real-world dataset collected by the MIT Computer Science and Artificial Intelligence Lab, we represent experimental as well as explain the performance of QoS system. Keywords Video streaming · Mobile thin-client · K-means clustering · Quality of service · Remote protocol · Cloud computing Keywords Introduction Cloud computing has become a hot topic in todays world, largely due to the flexible and cost savings it can deliver. In the same manner as with visualization, cloud computing implementations started with server infrastructure and has moved to the desktop; the cloud is now ripe for desktop infrastructure. By moving desktops to the cloud rather than an internally deployed and managed VDI solution in a data center, businesses can realise all of the promised benefits of virtual desktops [1]. Furthermore, the Internet has rapidly emerged as a mechanism for users to find and retrieve content, originally for webpages and recently for streaming media. Moreover, mobile phones and tablets or any thin client devices can run powerful applications from anywhere and at any time via the cloud. Therefore, there are a lot of case studies needed to transmit data from the virtual desktop in the data cloud center to the screen endpoints. he doctors tired all the time to check a patient’s status frequently, included their heart rate or other medical signs read by specialized equipment. With VDI solutions, a patient can get convenient health care from the comfort of their own home. This allows doctors to quickly and conveniently make such information available remotely on the mobile devices which are connected to a server by using a remote desktop protocol [1]. Moreover, VDI helps save valuable time and satisfies the demand for personal control, while also reducing the cost of long-term medical care by avoiding waste of human resources on manual processing [2]. Because any mistakes is always leading cause of death, so new developments in health care are of considerable interest. Computer gaming is a very different application where remote display technology will also be useful. Gamers spent a lot of money on computer games, hardware, and accessories. Traditionally, computer games are delivered either in boxes or via Internet downloads. These games have to install the computer games on physical machines to play them. The installation process becomes extremely tedious because the games are too complicated and the computer hardware and system software are very fragmented. Take Blizzard’s Star-craft II as example, it may take more than an hour to install it on an i5 PC, and another hour to apply the online patches. Furthermore, gamers may find their computers are not powerful enough to enable all the visual effects yet achieve high frame rates. So, gamers have to repeatedly upgrade their computers so as to play the latest computer games. Hence, cloud gaming is a better way to deliver high-quality gaming experience and opens new business opportunity [3]. In a cloud gaming system, computer games run on powerful cloud servers and streaming video data to thin clients by using remote protocols [4]. The thin clients are lightweight and can be ported to resource-constrained platforms, such as mobile devices and TV set-top boxes. With cloud gaming, the key benefit is that gamers can play the latest computer games anywhere and anytime, while the game developers can optimize their games for a specific PC configuration. Laboratory research is another excellent potential application for cloud-based remote display technology. By using remote desktop protocols, researchers perform complex tasks requiring intensive computation resources or specialized software. In addition to these example applications, there are other fields in which using remote display protocols would be useful for transmitting screen data from a cloud server to a client. However, in the past there have been several challenges with deploying VDI. Main problems included the network condition, server overload, bandwidth and limitation of mobile device (hardware resource, storage, battery life time). Specifically, a client sends an input event request to a server, the sever receives and deploys several offloading computing tasks, then renders screen image, encodes, and finally transfers the results to the client (as shown in Fig 1). These graphical images will be decoded and displayed on the client side. Therefore, it is so importance to guarantee client-perceived end-to-end quality of service. We can summarize several important factors to measure performance of the system: – Network condition: latency and bandwidth utilization. Latency should be as small as possible. Furthermore, with many different bandwidth quotas, users always want to get the highest acceptable performance. – Hardware resource consumption: CPU has a critical meaning in the processing because the limited hardware resource of the client device and overhead consumption on the server. – User demands: Cross-platform adaptation because it is a great convenience for users who use many kinds of devices such as laptop, smart phone, tablet, etc. with different operating systems. Above factors directly affect QoE (Quality of Experience) that user perceives in each request which is made from local desktop, where claims for more high fidelity of display. Many solutions have been developed and reported for improving QoE in content delivery networks; typically, solutions include hardware and software improvements. The general idea is that whatever graphical elements are painted to the screen on the host and then scraped by the protocol interface, are sent down to the client. This can happen in two ways: – The client can contact the server and pull a new snapshot of the screen from the frame buffer. This is how Virtual network computing (VNC) works [5], [7]. – The server can continuously push its screen activity to the client. This can be at the frame buffer level, the GDI / window manager level, or a combination of both. This is how (Remote Desktop Protocol) RDP [8] and Independent Computing Architecture work [9]. ith flexible QoS control based on network condition analysis to improve users’ QoE. We carefully establish the best compression parameters based on statistical analysis of historical data on network conditions. Moreover, to improve the real-time video streaming quality we develop a novel compression algorithm for 2D images, based on k-means clustering. The contributions of this work can be summarized as follows: 1. We develop a novel image compression algorithm based on k-means clustering which can be applied in real-world case studies. 2. We present an efficient architecture including QoS control based on analysis of historical data to account for network conditions. This system applies multiple linear regression to make optimal decisions for the QoS policy. We evaluate the proposed image compression algorithm by means of simulations in Matlab, using as input a dataset of real images collected by the MIT Computer Science and Artificial Intelligence Lab [36]. We demonstrate the QoS control system in a testbed cloud environment and analyze its performance. The rest of this paper is organized as follows. Related work is reviewed in Section 2. Section 3 gives an image compression algorithm based on the kmeans algorithm. Section 4 suggests an architecture to dynamically modify the compression rate of a video stream based on historical data of the recipients bandwidth. Section 5 describes experiments with the proposed QoS control and evaluates its performance. Finally, Section 6 gives conclusions. This paper is available on arxiv under CC BY 4.0 DEED license. This paper is available on arxiv under CC BY 4.0 DEED license. available on arxiv Authors: (1) Huu-Quoc Nguyen, Department of Computer Engineering, Kyung Hee University, 1 Seocheon, Giheung, Yongin, Gyeonggi, South Korea and quoc@khu.ac.kr; (2) Tien-Dung Nguyen, Department of Computer Engineering, Kyung Hee University, 1 Seocheon, Giheung, Yongin, Gyeonggi, South Korea and ntiendung@khu.ac.kr; (3) Van-Nam Pham, Department of Computer Engineering, Kyung Hee University, 1 Seocheon, Giheung, Yongin, Gyeonggi, South Korea and nampv@khu.ac.kr; (4) Xuan-Qui Pham, Department of Computer Engineering, Kyung Hee University, 1 Seocheon, Giheung, Yongin, Gyeonggi, South Korea and pxuanqui@khu.ac.kr; (5) Quang-Thai Ngo, Department of Computer Engineering, Kyung Hee University, 1 Seocheon, Giheung, Yongin, Gyeonggi, South Korea and nqthai@khu.ac.kr; (6) Eui-Nam Huh, Department of Computer Engineering, Kyung Hee University, 1 Seocheon, Giheung, Yongin, Gyeonggi, South Korea and johnhuh@khu.ac.kr. Authors: Authors: (1) Huu-Quoc Nguyen, Department of Computer Engineering, Kyung Hee University, 1 Seocheon, Giheung, Yongin, Gyeonggi, South Korea and quoc@khu.ac.kr; (2) Tien-Dung Nguyen, Department of Computer Engineering, Kyung Hee University, 1 Seocheon, Giheung, Yongin, Gyeonggi, South Korea and ntiendung@khu.ac.kr; (3) Van-Nam Pham, Department of Computer Engineering, Kyung Hee University, 1 Seocheon, Giheung, Yongin, Gyeonggi, South Korea and nampv@khu.ac.kr; (4) Xuan-Qui Pham, Department of Computer Engineering, Kyung Hee University, 1 Seocheon, Giheung, Yongin, Gyeonggi, South Korea and pxuanqui@khu.ac.kr; (5) Quang-Thai Ngo, Department of Computer Engineering, Kyung Hee University, 1 Seocheon, Giheung, Yongin, Gyeonggi, South Korea and nqthai@khu.ac.kr; (6) Eui-Nam Huh, Department of Computer Engineering, Kyung Hee University, 1 Seocheon, Giheung, Yongin, Gyeonggi, South Korea and johnhuh@khu.ac.kr.