A cache-aware social-based QoS routing scheme in Information Centric Networks

https://doi.org/10.1016/j.jnca.2018.07.002Get rights and content

Abstract

Due to the rapid expansion of Internet and the huge proliferation of users, Internet has evolved from a host-centric model to a content-oriented model. This implies the in-adaptation of current TCP/IP architecture providing the best performance to end-users and the urgency of researching future Internet architecture. In future Internet, the named data rather than traditional IP address may become the thin waist of the hourglass model of networking. Therefore, in this paper, we propose a cache-aware social-based Quality of Service (QoS) routing scheme for Named Data Networking (NDN) in Information Centric Network (ICN). Three kinds of social relationships, namely neighbors (NB), interest friends (IF) and response friends (RF) are devised to describe the relationships among nodes. Thus, when there is a failure in doing Pending Interest Table (PIT) scheme, a forwarding scheme based on social relationships is done before doing Forwarding Information Base (FIB) scheme. Moreover, a caching policy and its corresponding replacement policy based on content popularity, cache space and neighbor caching information are proposed. Results from simulation experiments demonstrate that our proposed scheme has better performance, including a higher routing success ratio, than NDN routing mechanism.

Introduction

Internet, running on top of Transmission Control Protocol/Internet Protocol (TCP/IP), has played a more and more important role in modern society since it was devised in the 1960s. The Internet paradigm is a host-centric model which was developed in accordance with its early usage, such as providing connectivity and sharing resource. However, with the rapid development of telecommunication technology and various networked applications, Internet usage is evolving from the host-centric model to a content-oriented model, and things that people pay attention to are shifting from “where” they can get information to “what” that information actually is. To follow these evolutions, Internet has become more and more complex. For example, peer-to-peer (P2P) technologies and content delivery networks (CDN) have been developed to enable all interested users to access content as efficiently as possible (Passarella, 2012). But they place a lot of pressure on the current TCP/IP model which makes it difficult for Internet to offer the best performance to end-users. Moreover, according to the Cisco report, the number of Internet-connected devices is expected to increase twofold from 22.9 billion in 2016 to 50 billion by 2020, and the global IP traffic is expected to increase nearly threefold from 1.2 ZB in 2016 to 3.3 ZB by 2021 per year (Cisco, 2017). The prominence of Big Data aggravates the already heavy burden on Internet. Plenty of research communities have been motivated to develop the Information Centric Networking (ICN) paradigm, which has emerged as a promising candidate for the future Internet (Ahlgren et al., 2012). By naming information at the network layer, ICN favors the deployment of in-network cache, thus facilitating the efficient information delivery to end-users. Moreover, some challenges in the current Internet, for example multicast, seem to have been addressed by ICN because one content can be provided by multiple different providers, that is, one entry may contain multiple outgoing faces. The ICN model thus supports multicast by default (Vasilakos et al., 2015). Thus, ICN has become a crucial research focus in future Internet studies.

Nowadays, the research on future Internet attracts large numbers of researchers. As shown in Fig. 1, in the United States, NSF funded five projects, namely Named Data Networking (NDN), MobilityFirst, eXpressive Internet Architecture (XIA), NEBULA and ChoiceNet under its FIA in 2010, and the former three projects entered the next phrase in 2014. In European Union, FP7 and H-2020 funded some future Internet research, such as Publish-Subscribe Internet Technology (PURSUIT) between 2007-2013 and 2014–2020 respectively. In the above research, NDN has roots in CCN and changes the basic network service semantics from “delivering packet to a given destination” to “retrieving data with a given name”. It reserves the hourglass model of TCP/IP and reshapes Internet architecture by taking “named data” not IP in the thin waist (Jacobson et al., 2009). Therefore, we take NDN as the basic architecture used in this paper.

Although ICN has made significant progress, there are still some problems, such as routing, caching, and Quality of Service (QoS), which need to be addressed (Xylomenos et al., 2014). A conversion from being based on IP address to being based on content name makes traditional routing mechanism unable to be applied directly in ICN. Therefore, routing becomes an important and active research in ICN (Bari et al., 2012).

There are only two kinds of packets in NDN, namely Interest and Data. An Interest packet carrying the name of the desired data is sent by a source node, and a Data packet with the desired data can be returned by any node which holds the corresponding content in network. If the source node cannot receive any Data packet from the same interface where it forwards the Interest packet within a time period, it would get a timeout. Thus, a node should be aware of content distribution, otherwise it can only resort to blind forwarding with network load increasing and performance harmed. Therefore, a key issue of routing mechanism in NDN is how to know which nodes should be preferred when Interest packets are forwarded.

The rapid development of Internet, especially the emergence of web 2.0, changes the interactive way in modern society and strengthens social relationships among people. On one hand, human relationships in real world bring out plenty of online social networking services, such as Facebook and Twitter. On the other hand, virtual relationship in highly interconnected Internet accelerates the development of Internet and reversely influences human relationships. It is shown that more than 60% of current Internet traffic comes from social network and P2P transactions (Chen et al., 2013). There usually exist some kinds of social relationships among users if frequent communication occurs among their used terminals. Moreover, the communication process often occurs multiple times. Meanwhile, with the conversion of network usage from transmitting packets among hosts to communicating around content, social trend will be further strengthened in the future Internet. Therefore, routing performance would be improved as knowledge regarding social relationships among users is efficiently exploited, that is, when a routing request is received, a node should know which nodes can return the required results. Exploiting social relationships to guide routing has become a new research direction (Sofia et al., 2012).

However, the “relationship” among nodes is a complex and fuzzy concept. It can be divided into two categories, namely position and content. Two adjacent nodes in position just like neighbors in the real world. They have natural advantages to exchange data. Specially, some recent research in device-to-device (D2D) communications enlighten the work on data exchange and forwarding between adjacent nodes by taking advantage of their physical proximity (Fodor et al., 2012). On the other hand, two nodes having some similarity in content just like friends in the real world. People are increasingly relying on friends for sharing contents and interacting with each other (Yu et al., 2015). In NDN, the friend relationships mainly manifest in interest and content. For example, the probability of two nodes with similar interest and content satisfying the requirements of each other is higher than that of two nodes without any similarity. It should be noted that the ability of a node is considered an important factor in building social relationship. For example, a node with more content, stronger processing capability and higher connection degrees, just like the node with high betweenness in scale-free networks (Wu et al., 2012), would have a higher probability to satisfy the requirements from other nodes. Therefore, we exploit and apply neighbor relationship and friend relationship in network communication.

Meanwhile, in-network caching is an important characteristic of NDN. Caching received content makes a node respond to the subsequent Interest packets directly, thus the routing path can be shortened and transmission efficiency can be improved. Because the cache space of a node is limited, cache-everything-everywhere is unpractical, and thus being cache-aware makes Interest packets be forwarded to more appropriate nodes and routing performance improved (Zhang et al., 2015).

With the development of a variety of new networked applications, especially multimedia based, and the diversified demands of users, QoS support is more and more important to networks (Meddeb, 2010). For example, a bulk data transfer has the strict requirement on bandwidth and transfer reliability, while a video conference has the strict requirement on bandwidth, delay and delay jitter. Thus, in this paper, our proposed routing scheme considers networked application QoS requirements, specifically on bandwidth, delay, delay jitter, and error rate.

Based on the above discussion, we propose a Cache-aware Social-based QoS Routing scheme (CSQR) for ICN. The contributions of our work are summarized as follows:

  • (1).

    The CSQR not only exploits social relationships to forward Interest packets, but also is cache-aware and considers QoS constraints of network applications. To achieve the above functions, our routing scheme is built on modules, and each module which can be added or removed flexibly is responsible for the above different function. Thus, CSQR can be integrated into NDN and compatible with current routing mechanism easily.

  • (2).

    Three kinds of relationships, namely neighbor (NB), interest friends (IF) and response friends (RF) are built. They are based on node proximity, interest similarity, and node capability respectively, which jointly provide the most appropriate forwarding for Interest packets.

  • (3).

    A content popularity is proposed to denote the probability of a content to be requested later. Moreover, a caching policy and its corresponding replacement policy based on content popularity and neighbor cache information are presented. It makes node cache content reasonably and achieves a tradeoff between content distribution and node space.

The rest of this paper is organized as follows. The related work is reviewed in Section 2. The framework, model and problem formulation are introduced in Section 3. The algorithm description is given in Section 4. Simulation study is presented in Section 5. Conclusion is drawn in Section 6.

Section snippets

Related work

Recently, with the rapid development and huge success of social networks, researchers have begun to utilize social relationships among nodes to design applicable routing algorithms (Wei et al., 2014). For example, in Rothfus et al. (2013), a novel routing algorithm which uses similarity metrics from data mining in node contact history is proposed. In Li et al. (2013), social energy, a new metric which is generated via node encounters and shared by the communities of encountering nodes, is

System framework and model

In this section, we describe a framework of system, a network model proposed for NDN, and the request model dealt with in this paper.

QoS evaluation

The actual QoS parameters of a path rs,d from node vs to node vd are calculated respectively as follows:bwrs,d=minbwei,j,ei,jrs,ddl(rs,d)=ei,jrs,ddl(ei,j)jt(rs,d)=ei,jrs,djt(ei,j)er(rs,d)=1ei,jrs,d(1er(ei,j))where bw(rs,d), dl(rs,d), jt(rs,d), er(rs,d) and bw(ei,j), dl(ei,j), jt(ei,j), er(ei,j) are the available bandwidth, delay, delay jitter and error rate of rs,d and ei,j respectively.

To better evaluate the QoS along the routing paths for different networked applications, we

Experimental setting

We have implemented CSQR by C++ programming language on a personal computer with Intel Core i5-4570 @ 3.20 GHz Dual Core having 4 GB DDR3 RAM (Saxena and Raychoudhury, 2016). The operating environment is 64-bit Windows 10 professional. We compare CSQR with NDNR which is the basic routing mechanism in NDN. When a node receives an Interest packet, NDNR firstly checks whether there is matched content in its CS. If yes, it generates a corresponding Data packet and sends back to the incoming

Conclusion

In this paper, a cache-aware social-based QoS routing scheme is proposed for ICN. It builds three kinds of social relationships, namely NB, IF and RF, and deals with the diversified QoS requirements. Moreover, a caching policy and corresponding cache replacement policy based on content popularity and neighbor cache information are presented to achieve a reasonable tradeoff between content copy and node capacity. With the development of information communication technology, real society and

Acknowledgments

This work was supported by the Major International (Regional) Joint Research Project of NSFC under Grant No. 71620107003; the National Science Foundation for Distinguished Young Scholars of China under Grant No. 71325002; the MoE and ChinaMobile Joint Research Fund under Grant No. MCM20160201; the National Natural Science Foundation of China under Grant No. 61572123; the Program for Liaoning Innovative Research Term in University under Grant No. LT2016007; the PhD Science Initiation Foundation

Dapeng Qu received the B.S. degree and M.S. degree in applied mathematics and computer science from Central South University, Changsha, China, in 2003, and 2006 respectively, and Ph.D. degree in computer science from the Northeastern University, Shenyang, China, in 2012. He is currently an associate professor at the College of Information, Liaoning University. His research interests include computer networking and future Internet, etc.

References (41)

  • Z. Yu et al.

    Friend recommendation with content spread enhancement in social networks

    Inf. Sci.

    (2015)
  • G. Zhang et al.

    Caching in information centric networking: a survey

    Comput. Network.

    (2013)
  • A. Afanasyev et al.

    Packet Fragmentation in NDN: Why NDN Uses Hop-by-hop Fragmentation NDN Memo

    (2015)
  • B. Ahlgren et al.

    A survey of information-centric networking

    IEEE Commun. Mag.

    (2012)
  • M.F. Al-Naday et al.

    Quality of service in an information-centric network

  • M.F. Bari et al.

    A survey of naming and routing in information-centric networks

    IEEE Commun. Mag.

    (2012)
  • J.M. Blanquer et al.

    Quorum: flexible quality of service for Internet services

  • CERNET2

    CERNET2 Topology

    (2006)
  • K. Chen et al.

    From technological networks to social networks

    IEEE J. Sel. Area. Commun.

    (2013)
  • W. Chen et al.

    An optimal weights combination method considering both subjective and objective weight information in power quality evaluation

  • Cited by (0)

    Dapeng Qu received the B.S. degree and M.S. degree in applied mathematics and computer science from Central South University, Changsha, China, in 2003, and 2006 respectively, and Ph.D. degree in computer science from the Northeastern University, Shenyang, China, in 2012. He is currently an associate professor at the College of Information, Liaoning University. His research interests include computer networking and future Internet, etc.

    Xingwei Wang received the B.S., M.S., and Ph.D.degrees in computer science from the Northeastern University, Shenyang, China, in 1989, 1992, and 1998 respectively. He is currently a Professor at the College of Computer Science and Engineering, Northeastern University. He has published over 100 research papers. His research interests include cloud computing and future Internet, etc.

    Min Huang received the B.S. degree in automatic instrument, the M.S. degree in systems engineering, and Ph.D. degree in control theory from the Northeastern University, Shenyang, China, in 1990, 1993, and 1999 respectively. She is currently a Professor at the College of Information Science and Engineering, Northeastern University. She has published over 100 research papers. Her research interests include the modeling and optimization for the logistics and supply chain system, etc.

    Keqin Li received the B.S. degree in computer science from Tsinghua University, Beijing, China, in 1985, and Ph.D. degree in computer science from the University of Houston, Texas, USA, in 1990. He is currently a SUNY distinguished professor of computer science in State University of New York at New Paltz. He has published over 200 journal articles, book chapters, and research papers in refereed international conference proceedings. His research interests include parallel and distributed computing and computer networking, etc.

    Sajal K Das received the B.S. degree in computer science from Calcutta University, Kolkata, India, in 1983, the MS degree in computer science from the Indian Institute of Science, Bangaluru, India, in 1984, and the PhD degree in computer science from the University of Central Florida, Orlando, in 1988. He is currently with the Department of Computer Science, Missouri University of Science and Technology. He is the author of more than 400 published papers and more than 35 invited book chapters. He is a Fellow of the IEEE.

    Sijin Wu received the B.S. degree in computer science and technology from the Northeastern University, Shenyang, China, in 2015, and M.S. degree in computer application technology from the Northeastern University, Shenyang, China, in 2018. Her research interest is Information-Centric Networking.

    View full text