Elsevier

Computer Networks

Volume 93, Part 1, 24 December 2015, Pages 199-212
Computer Networks

Maximizing real-time streaming services based on a multi-servers networking framework

https://doi.org/10.1016/j.comnet.2015.10.019Get rights and content

Abstract

In recent years, we have witnessed substantial exploitation of real-time streaming applications, such as video surveillance system on road crosses of a city. So far, real world applications mainly rely on the traditional well-known client–server and peer-to-peer schemes as the fundamental mechanism for communication. However, due to the limited resources on each terminal device in the applications, these two schemes cannot well leverage the processing capability between the source and destination of the video traffic, which leads to limited streaming services. For this reason, many QoS sensitive application cannot be supported in the real world. In this paper, we are motivated to address this problem by proposing a novel multi-server based framework. In this framework, multiple servers collaborate with each other to form a virtual server (also called cloud-server), and provide high-quality services such as real-time streams delivery and storage. Based on this framework, we further introduce a (1ϵ) approximation algorithm to solve the NP-complete “maximum services”(MS) problem with the intention of handling large number of streaming flows originated by networks and maximizing the total number of services. Moreover, in order to backup the streaming data for later retrieval, based on the framework, an algorithm is proposed to implement backups and maximize streaming flows simultaneously. We conduct a series of experiments based on simulations to evaluate the performance of the newly proposed framework. We also compare our scheme to several traditional solutions. The results suggest that our proposed scheme significantly outperforms the traditional solutions.

Introduction

In recent years, we have witnessed the popularity of real-time streaming applications in the various environments, such as the video surveillance systems  [1], [2], [3]. Not only can the police monitor the risky streets and road crosses, but also general public can employ these applications to enhance security and safety in their business and life. In these applications, there will be many cameras installed in order to cover as large area as one can. For example, Nasution and Foroughi et al. proposed methods to detect and record various posture-based events of interest in elder monitoring application at various home surveillance scenarios  [4], [5], [6]. In particular, due to the popularity of smart phones and wide accessibility of wireless connections, people are able to use portable devices to attain the events of interest from remote cameras  [7].

Due to the limited resources on each terminal device in the applications, current schemes of the real-time streaming applications cannot well leverage the processing capability and storage space between the source and destination of the video traffic. For this reason, many QoS sensitive application cannot be supported in the real world. For example, there are traditionally mainly two general architectures for real-time services. The first one is the client–server (C/S) scheme where the terminals stream the data captured in the places of interest to the rendezvous server. Remote users can then retrieve the data from the server. This scheme is simple but incurs heavy burdens on the server. It also does not scale well when more terminals of monitoring get involved [8]. Another widely-adopted scheme is the peer-to-peer scheme (P2P). The P2P scheme greatly alleviates the burdens on rendezvous servers, and improves the applications scalability [8], [9]. For example, Pudlewski [10] proposed a point-to-point real-time video transmission scheme over the Internet in conjunction with a compression method that is error resilient and bandwidth-scalable. However, any terminal in the P2P scheme acts as both client and server. This leads to the heavy burden on the transmission source. Moreover, because the lightweight server (every terminal) is in short of storage space and processing capability, the P2P scheme cannot support streaming data retrieval. The situation becomes worse when the connections are not stable. The quality of service (QoS) will become unacceptable. To date, It has long been a big challenge to realize the real-time streaming over best-effort packet networks. On the other side, some real-time streaming data is required to be backed up for later retrieval. For example, the data produced by video surveillance system is usually required to be backed up for a long time, which can be retrieved for taking evidence.

In this paper, we proposed a novel framework based on multi-server to address the above problem. In this framework, we take the real-time data streaming and storage into account at the same time, and multiple servers will incorporate with a virtual server (also called cloud-server) to provide both forwarding and storage services. Since our newly proposed framework can fully utilize the resources from dynamic networks [11], the multiple servers provide better robustness in case one transmission channel becomes congested, and achieve high throughput on end users. We carried out a series of experiments in order to test and evaluate the performance of our proposed framework. Compared to the traditional single server scheme, the best server and route/path could be picked out in our solution. For the perspective of users, the service will be provided by a super server, and the multiple servers will coordinate with each other to maximize the simultaneous services.

We summarize the contributions of this paper as follows:

  • We proposed a novel framework based on multi-servers to collaboratively provide services for each end user. This framework exploits the diversity of network and efficiently realize the real-time streaming and backup at the same time.

  • Based on the proposed framework, we innovatively formulated a brand-new problem in providing real-time streaming services: Maximum Services (MS). This problem aims to maximize the total number of streaming flows and/or services. We proved that this problem is actually an NP-complete problem.

  • For the real working applications of real-time streaming services, we innovatively inserted a two-relay restriction into the framework, and therefore simplified the MS problem. We designed an approximate algorithm with foreseeable performance bounds to the optimal solution. Extensive simulations were carried out to validate the effectiveness of the proposed methods for real working applications.

  • A new metric Possible Loss (PL) was introduced to evaluate the storage and backup performance. An efficient storage algorithm is designed to achieve the minimized PL.

    The rest of this paper is organized as follows: Section 2 presents related work. The details of the framework are presented in Section 3. The approximate algorithm of MS problem is presented in Section 4, followed by Section 5, which is about the performance evaluation. Section 6 concludes this paper.

Section snippets

Related work

Clearly, the amount of data in our industry and the world is exploding. Data is being collected and stored at unprecedented rates. For example, video streaming has become a popular application. In 2010, the number of video streams increased 38.8% to 24.92 billion even without counting the user generated videos [12]. Video news clips, movies and TV shows, and videos made and shared by the general public are watched by millions of people every day [13], [14]. A considerable amount of data

Multi-server based framework

In this section, we first introduce the multi-server based framework which includes multiple cooperative servers. Based on this framework, the problem aiming at maximizing the number of simultaneous streams is formulated.

Algorithms for realistic applications

Since the MS problem is NP-complete, a polynomial time algorithm cannot be found. Therefore, we seek to simplify the problem for a practical solution. We note that in reality, if there are too many forwarding hops among the servers, it will lead to both long delay and a high loss rate that may not satisfy the real-time requirement. Moreover, multi-hop forwarding among servers will lead to extra communication cost between servers. In this section, we introduce several special cases of the MS

Performance evaluations

This section presents the evaluation of the algorithms using MATLAB. We compare the proposed algorithms (denoted as MS-1 and MS-2) with the traditional “one-server” method and P2P method (MS-0). As a baseline, we also run the brute-force method to get the optimal solution when multi-hop forwarding are allowed among servers, which is denoted as MS-n. In the simulations, by saying “P2P capacity”, we mean the capacities from sources to servers or the capacities from servers to users. To simulate

Conclusions

In this paper, we studied the networking aspect of real-time streaming applications. We designed a novel real-time streaming framework where multiple servers form into a cloud-server to collaboratively provide services. Our framework combined the advantages of both the traditional client/server scheme and P2P scheme. First, a server group acts as the streaming servers, which are more powerful. Secondly, the multiple servers can backup streaming data, which can improve the data security if a

Acknowledgment

Above work was supported in part by grants from the National China 973 Project no: 2015CB352401 and National Science Foundation of China (NSFC) Key Project Fund No. 61532013 and the National Natural Science Foundation (NSF) of China under Grant nos. 61572206, 61202468, 61305085, 61472451, 61272151, 61370007, 61302094 and the Natural Science Foundation of Fujian Province of China (No. 2014J01240) and the Promotion Program for Young and Middle-aged Teacher in Science and Technology Research of

Tian Wang received his BSc and MSc degrees in Computer Science from the Central South University in 2004 and 2007, respectively. He received his PhD degree in City University of Hong Kong in 2011. Currently, he is an assistant professor in the National Huaqiao University of China. His research interests include wireless networks, Internet of things and mobile computing.

References (29)

  • P. Quax, W. Vanmontfort, R. Marx, et al. Overlay network based optimization of data flows in large scale...
  • L. Cai et al.

    P2P traffic identification based on transfer learning

    Proceedings of the IEEE International Conference on Granular Computing (GrC)

    (2013)
  • S. Pudlewski et al.

    Compressed-sensing-enabled video streaming for wireless multimedia sensor networks

    IEEE Trans. Mob. Comput.

    (2012)
  • M. Koerner et al.

    Multiple service load-balancing with openflow

    Proceedings of the 13th IEEE International Conference onHigh Performance Switching and Routing (HPSR), 2012

    (2012)
  • Cited by (14)

    • Data collection from WSNs to the cloud based on mobile Fog elements

      2020, Future Generation Computer Systems
      Citation Excerpt :

      Simulation results are demonstrated in Section 7, and Section 8 concludes the paper. Nowadays, with the explosive growth of data collected by sensors, the demand for data collection in WSNs becomes more important and stringent [14,15]. Most traditional data collection systems are based on multi-hop transmission between sensors [16].

    • Fog-based storage technology to fight with cyber threat

      2018, Future Generation Computer Systems
      Citation Excerpt :

      As fog computing defines, it is closer to the ground than the cloud. The transfer rate between fog computing layer and other layers is faster than the rate directly between cloud layer and the bottom layer [39]. 2.

    • Reliable wireless connections for fast-moving rail users based on a chained fog structure

      2017, Information Sciences
      Citation Excerpt :

      Train passengers often use their smartphones to watch online videos on services such as YouTube [30] and play games such as Hearthstone [27]. These online video services and games typically have stringent performance requirements, including sufficient connectivity time, short delay, and high bandwidth [37]. For example, online games require consistent network coverage to smoothly deliver streaming data.

    • A privacy-preserving model for cloud data storage through fog computing

      2022, International Journal of Computer Aided Engineering and Technology
    • Substituting missing values in end-To-end internet performance measurements using k-nearest neighbors

      2018, Proceedings - IEEE 16th International Conference on Dependable, Autonomic and Secure Computing, IEEE 16th International Conference on Pervasive Intelligence and Computing, IEEE 4th International Conference on Big Data Intelligence and Computing and IEEE 3rd Cyber Science and Technology Congress, DASC-PICom-DataCom-CyberSciTec 2018
    View all citing articles on Scopus

    Tian Wang received his BSc and MSc degrees in Computer Science from the Central South University in 2004 and 2007, respectively. He received his PhD degree in City University of Hong Kong in 2011. Currently, he is an assistant professor in the National Huaqiao University of China. His research interests include wireless networks, Internet of things and mobile computing.

    Yiqiao Cai received the B.S. degree from Hunan University, Changsha, China, in 2007, and the Ph.D. degree from Sun Yat-sen University, Guangzhou, China, in 2012. In 2012, he joined Huaqiao University, Xiamen, China, where he is currently a lecturer with the College of Computer Science and Technology. He is interested in differential evolution, multiobjective optimization, and other evolutionary computation techniques.

    Sheng Wen graduated with the degree in CS from Central South University of China in 2012. He is currently working toward the PhD degree at the School of Information Technology, Deakin University, Melbourne, Australia, under the supervision of Prof. Wanlei Zhou and Yang Xiang. His focus is on modeling of virus spread, information dissemination, and defence strategies of the Internet threats. He is a student member of the IEEE.

    Weijia Jia is currently a full Professor in the Department of Computer Science in Shanghai Jiaotong University. He received BSc and MSc from Center South University , China in 1982 and 1984 and Master of Applied Sci. and PhD from Polytechnic Faculty of Mons, Belgium in 1992 and 1993 respectively, all in Computer Science. His research interests include next generation wireless communication, distributed systems, multicast and anycast routing protocols.

    Guojun Wang received the BSc degree in geophysics, the MSc degree in computer science, and the PhD degree in computer science, from Central South University, China. He is currently a professor at Guangzhou University and Central South University. He has been an adjunct professor at Temple University, Philadelphia, PA; a visiting scholar at Florida Atlantic University, Boca Raton, FL; a visiting researcher at the University fault tolerant of Aizu, Japan; and a research fellow at the Hong Kong Polytechnic University. His research interests include computer networking, Internet of things, cloud computing, and information security. He is a distinguished member of the CCF, and a member of the IEEE, ACM, and IEICE.

    Hui Tian received his BSc and MSc degrees in Wuhan Institute of Technology, Wuhan, China in 2004 and 2007, respectively. He received his PhD degree in Huazhong University of Science and Technology, Wuhan, China. He is now an associate professor in the National Huaqiao University of China. His research interests include network and multimedia information security, digital forensics and information hiding.

    Yonghong Chen received B.Sc. degrees from Hubei National University, and M.Eng. and Ph.D. degree degrees from Chognqing University, Chongqing, China, in 2000 and 2005 respectively. He is currently the professor of College of Computer Science and Technology, Huaqiao University, Xiamen, China. His research interests include network security, watermarking and nonlinear processing.

    Bineng Zhong received the B.S., M.S., and Ph.D. degrees in computer science from the Harbin Institute of Technology, Harbin, China, in 2004, 2006, and 2010, respectively. From 2007 to 2008, he was a Research Fellow with the Institute of Automation and Institute of Computing Technology, Chinese Academy of Science. Currently, he is an Associate Professor with the School of Computer Science and Technology, Huaqiao University, Xiamen, China, and he is also a Post-Doc with the School of Information Science and Technology, Xiamen University, Xiamen, China. His current research interests include pattern recognition, machine learning, and computer vision.

    View full text