Elsevier

Computer Networks

Volume 75, Part A, 24 December 2014, Pages 1-17
Computer Networks

Operational information content sum capacity: From theory to practice

https://doi.org/10.1016/j.comnet.2014.08.017Get rights and content

Abstract

This paper considers Quality-of-Information (QoI) aware resource allocation policies for multiuser networks. QoI is a recently introduced composite metric which is impacted by a number of attributes of information communicated from the source(s) to the destination(s), and as such differs from traditional quality-of-service metrics considered to date. The focus of this work is defining the Operational Information Content Sum Capacity (OICC-S) of a network, achieved by the set of information attributes supported that maximize sum quality of the network. This quality is defined as a function of the information attributes provided by the source input, as well as the channel induced attributes that impact the QoI delivered to the destination(s). Optimum rate allocation to maximize the output sum quality of information and achieve OICC-S of the network for various settings is provided, and demonstrated to differ from the solution that provides maximum throughput, making QoI-awareness necessary in resource allocation. Insights arising from the analysis are provided, along with those from practical scenarios.

Introduction

Traditional approaches for resource allocation based on Quality of Service (QoS) perform network operations that are agnostic to the application or the information content. Such approaches may prove suboptimal for task-oriented networks where the main goal is sound decision making. Several examples for such tasks involve crowd-sourcing, participatory sensing-type applications, as well as tactical networks. To this end, a new paradigm which emphasizes the quality of information by viewing the network as an information source, and developing methods to satisfy information quality requirements at the end user is necessary.

To characterize information quality, there is growing interest in moving from traditional QoS metrics as throughput, packet delivery ratio, fairness, and delay towards new notions of quality associated with information. This effort includes introducing new attributes which characterize the value of information relevant to the specific application [1], [2]. Attributes such as provenance, accuracy, precision, reliability, corroboration, credibility, age/freshness, and timeliness have been used to define the quality of information [1], [2], [3], [4]. Event detection applications for QoI are studied in Refs. [1], [5]. Recently, there have also been studies which focus on QoI-based scheduling [6], [7], [8]. In [9], [7], we have optimized delivered QoI for scenarios with randomness in either channel conditions or traffic, focusing on a source–destination pair. In [6], we have introduced the concept of operational information content sum capacity and demonstrated initial associated theoretical results for a multisource scenario. In this work, we build on [6] to provide a comprehensive study to address QoI-aware network system optimization from both theoretical and practical aspects.

We consider the following scenario. A network is sent tasks sequentially from an end user, and users with sensing capabilities respond to these tasks. We are interested in the set of information attribute vectors that the network can support. Moreover, we identify which of these vectors of information attributes are most useful in terms of decision making associated with the task through a Quality-of-Information function. We denote the maximum sum QoI achieved by these information attribute vectors supported by the network as the Operational Information Content Sum Capacity (OICC-S) of the network. Proposed recently, the notion of Operational Information Content Capacity (OICC) is an indicator of the decision making capability that the collection of sources and links, i.e., the network can provide [2]. As such, it differs from, for instance, the Network Utility Maximization (NUM) framework where the traditional utility is a function of the flow rates [10]. While there have been recent efforts to include delay-dependent terms in the NUM framework [11], [12], we take the viewpoint of optimizing of QoI metrics such as accuracy by file size adaptation at the sources. Another main difference is that while the NUM framework deals with optimal rate adaptation, we include optimization of the attributes at the source in addition to optimal rate allocation. Although the concept of QoI by itself is associated with information generated by a single source, OICC-S captures the interaction of multiple sources or flows and the physical layer they share. More specifically, we address the problem of sum quality maximization via optimal rate allocation given the application specifications and network constraints.

Among the attributes which can effect QoI and OICC-S, we focus on the effects of source-specific attributes as accuracy and timeliness.1 Information attributes as accuracy, precision and completeness are indicators of the initial information content and the success of generating information at the sources. Timeliness, which measures the availability of information relative to the time it is needed, is related with success of network delivery. We choose accuracy and timeliness since these two attributes together capture both source and network dependent factors on quality. Accordingly, the overall OICC-S maximizing optimization framework involves both source- and link-level decisions. These sets of attributes possess a trade-off such that improving source attributes can degrade timeliness for a given network. We consider several models for QoI that depends on these two metrics.

We consider various network scenarios with the objective of maximizing the sum quality of the system, i.e., achieving the OICC-S. The main issue we address is obtaining the balance between source attributes, specifically accuracy, and timeliness for the given network, by rate allocation. QoI is a composite function of these source- and network- based attributes, hence maximization of sum quality calls for new treatment compared with the network-centric NUM framework. We first provide theoretical results for a two-user multiple access channel (MAC). For this scenario, it is well known that max weight scheduling maximizes throughput for this model by operating at one of two corner points for the MAC capacity region [13]. In contrast, here, we demonstrate that arbitrary points on the dominant face of the rate region can be optimal rate points to attain OICC-S. Next, we demonstrate that OICC-S optimizing rate allocation strategies significantly differ from throughput-maximizing rate allocation for a several canonical topologies operating with practical protocols as TDMA and CSMA/CA operating with several widely used commercial applications. We conclude, based on the analysis and the simulations that rather than focusing on maximizing the number of bits in resource allocation, QoI-aware policies are necessary to maximize the decision-making capability of a network. The organization of the paper is as follows. In Section 2, we present the basic model and QoI definitions. Next, in Section 3 we formally define the OICC-S. We provide theoretical results associated with rate allocation and information attribute optimization problems to achieve the OICC-S for different settings in Section 4. Sections 5 Case study: TDMA based network with multiple applications, 6 Case studies: Canonical and arbitrary network topologies with CSMA/CA present scenarios with widespread applications and practical network settings. We conclude the paper in Section 7.

Section snippets

QoI: Definitions, user and application perspective

QoI is a composite, multi-dimensional metric that captures the trade-offs of several components to characterize the information ultimately delivered to the application. QoI as determined by an application is a function of both intrinsic and contextual metrics. Intrinsic metrics are those that are valued independently of the use of the information. For example, the freshness of information, i.e., its age, is a function of when the information was generated, and once delivered will have the same

Operational Information Content Sum Capacity (OICC-S)

We consider a scenario where tasks are issued from an end user in a tactical network. Tasks arrive with a random interarrival time greater than Tmin. We assume that at most one task is processed by the network at any time. Information sources Si,i=1,,K are capable of responding to this task and focus on independent events and possibly possess or generate different types of information related with the task. Moreover, each information source can respond to the task with potentially up to J

Case study: The multiple access channel

In this section, we shall concentrate on uplink scenarios. More specifically, we consider the two-user Multiple Access Channel (MAC) shown in Fig. 8, and the two-user Time-Division-Multiple-Access (TDMA). The results can be readily generalized to more than two users. These constitute basic and inspiring models for analytical OICC-S characterization, which involve multiuser issues as proper rate allocation between users. This rate allocation is dependent on QoI functions and information

Case study: TDMA based network with multiple applications

In the previous section, we have analytically demonstrated the necessity for QoI-aware scheduling and optimization, as the optimal solutions can deviate from traditional QoS-based network solutions. While this is a fundamental result, recall from Section 4 that it is not tractable to theoretically analyze any given network scenario and application in detail.

To that end, in this section to make the analysis tractable, we relax the constraints on the application characteristics. Specifically, in

Case studies: Canonical and arbitrary network topologies with CSMA/CA

Having presented QoI characteristics of several real-world applications and how multiple such applications can be scheduled among two sources, it is interesting to study the behavior of the OICC-S achieving resource allocation in more complicated scenarios, which have been previously used in order to understand the complications posed by wireless multi-hop networks in realizing scheduling and congestion control schemes. We have taken a number of such canonical scenarios as well as a

Conclusion

In this paper, we propose methods for QoI based evaluation in multiuser networks. We characterize the maximum sum output QoI provided by information-vectors supportable by the network as the OICC-S. For OICC-S formulation, we focus on the effect of network delivery and timeliness on information with specific accuracy attributes. We first theoretically characterize rate allocation schemes in order to attain OICC-S for the most basic multiuser network model, specifically a two-user MAC. Next, we

Acknowledgements

Research was sponsored by the U.S. Army Research Laboratory under the Network Science Collaborative Technology Alliance, Agreement Number W911NF-09-2-0053. The views and conclusions contained in this document are those of the author(s) and should not be interpreted as representing the official policies, either expressed or implied, of the U.S. Army Research Laboratory or the U.S. Government. The U.S. Government is authorized to reproduce and distribute reprints for Government purposes

Ertugrul Necdet Ciftcioglu received his B.S. degree in Electrical and Electronics Engineering from the Middle East Technical University (METU), Ankara, Turkey in 2004, M.S. degree in Electronics Engineering and Computer Science from Sabanci University, Istanbul, Turkey in 2006, and Ph.D. degree in Electrical Engineering from The Pennsylvania State University, PA in 2012. He has been a Research Associate in the Department of Computer Science and Engineering, The Pennsylvania State University

References (26)

  • J. Edwards et al.

    Quality of information-aware mobile applications

    Pervasive Mobile Comput.

    (2014)
  • G. Anastasi et al.

    Energy conservation in wireless sensor networks: a survey

    Ad Hoc Netw.

    (2009)
  • S. Zahedi, C. Bisdikian, A framework for QoI-inspired analysis for sensor network deployment planning, in:...
  • A. Bar-Noy, G. Cirincione, R. Govindan, S. Krishnamurthy, T.F. LaPorta, P. Mohapatra, M. Neely, A. Yener,...
  • B. Liu, P. Terlecky, A. Bar-Noy, R. Govindan, M.J. Neely, Optimizing information credibility in social swarming...
  • Z.M. Charbiwala, S.Z.Y. Kim, Y. Cho, M.B. Srivastava, Toward quality of information aware rate control for sensor...
  • E.N. Ciftcioglu, A. Yener, R. Govindan, K. Psounis, Operational information content sum capacity: formulation and...
  • R. Urgaonkar, E.N. Ciftcioglu, A. Yener, M.J. Neely, Quality of information aware scheduling in task processing...
  • E.N. Ciftcioglu et al.

    Maximizing quality of information from multiple sensor devices: the exploration-exploitation tradeoff

    IEEE J. Sel. Top. Signal Process.

    (2013)
  • E.N. Ciftcioglu, A. Yener, Quality-of-information aware transmission policies with time-varying links, in: Proc. MILCOM...
  • F. Kelly

    Charging and rate control for elastic traffic

    Eur. Trans. Telecommun.

    (1997)
  • H. Xiong et al.

    Delay-aware cross-layer design for network utility maximization in multi-hop networks

    IEEE J. Sel. Areas Commun.

    (2011)
  • A. Szwabe, P.M.P. Walkowiak, Delay-aware NUM system for wireless multi-hop networks, in: Wireless Conference 2011 –...
  • Cited by (3)

    Ertugrul Necdet Ciftcioglu received his B.S. degree in Electrical and Electronics Engineering from the Middle East Technical University (METU), Ankara, Turkey in 2004, M.S. degree in Electronics Engineering and Computer Science from Sabanci University, Istanbul, Turkey in 2006, and Ph.D. degree in Electrical Engineering from The Pennsylvania State University, PA in 2012. He has been a Research Associate in the Department of Computer Science and Engineering, The Pennsylvania State University from 2012 to 2014. He has also been a visiting scholar at Northwestern University, IL, University of Southern California, CA and BBN Technologies, MA. Since September 2014, he is a Postdoctoral Researcher at IBM Research, Yorktown Heights, NY. His research interests are cross-layer design and resource allocation for wireless communication networks, particularly stochastic network optimization for relaying, cooperative communications and multiuser networks, and recent emphasis on network science.

    Antonios Michaloliakos graduated in 2009 from the department of Electrical and Computer Engineering of the National Technical University of Athens. He is currently pursuing a PhD in Electrical Engineering at the University of Southern California. His interests span the areas of probabilistic modeling and design of networks and cross-layer optimization in new generation wireless networks. He is currently working on analytic modeling of MU-MIMO systems.

    Aylin Yener received the B.Sc. degree in electrical and electronics engineering, and the B.Sc. degree in physics, from Bogazici University, Istanbul, Turkey; and the M.S. and Ph.D. degrees in electrical and computer engineering from Wireless Information Network Laboratory (WINLAB), Rutgers University, New Brunswick, NJ. Commencing fall 2000, for three semesters, she was a P.C. Rossin Assistant Professor at the Electrical Engineering and Computer Science Department, Lehigh University, PA. In 2002, she joined the faculty of The Pennsylvania State University, University Park, PA, where she was an Assistant Professor, then Associate Professor, and is currently Professor of Electrical Engineering since 2010. During the academic year 2008–2009, she was a Visiting Associate Professor with the Department of Electrical Engineering, Stanford University, CA. Her research interests are in information theory, communication theory and network science, with recent emphasis on green communications and information security. She received the NSF CAREER award in 2003. Dr. Yener previously served as a technical program chair or co-chair for various conferences for the IEEE Communications Society, as an associate editor for the IEEE TRANSACTIONS ON COMMUNICATIONS, as an associate editor and an editorial advisory board member for the IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS. She served as the student committee chair for the IEEE Information Theory Society 2007–2011, and was the co-founder of the Annual School of Information Theory in North America co-organizing the school in 2008, 2009 and 2010. Dr. Yener currently serves on the board of governors of the IEEE Information Theory Society as its treasurer.

    Konstantinos Psounis is an associate professor of Electrical Engineering and Computer Science at the University of Southern California. He received his first degree from the department of Electrical and Computer Engineering of National Technical University of Athens in 1997, the M.S. degree in Electrical Engineering from Stanford University, California, in 1999, and the Ph.D. degree in Electrical Engineering from Stanford University in 2002. Konstantinos models and analyzes the performance of a variety of wired and wireless networks and designs schemes and protocols to solve problems related to such systems. He is the author of numerous research papers on these topics which have received thousands of citations, and has faculty awards from multiple sources including the National Science Foundation, the Army Research Laboratory and CISCO Systems. He is a senior member of both IEEE and ACM.

    Thomas F. La Porta is the William E. Leonhard Chair Professor in the Computer Science and Engineering Department at Penn State. He received his B.S.E.E. and M.S.E.E. degrees from The Cooper Union, New York, NY, and his Ph.D. degree in Electrical Engineering from Columbia University, New York, NY. He joined Penn State in 2002. He is the Director of the Institute of Networking and Security Research at Penn State. Prior to joining Penn State, Dr. La Porta was with Bell Laboratories since 1986 where he was the Director of the Mobile Networking Research Department in Bell Laboratories, Lucent Technologies where he led various projects in wireless and mobile networking. He is an IEEE Fellow, Bell Labs Fellow, received the Bell Labs Distinguished Technical Staff Award in 1996, and an Eta Kappa Nu Outstanding Young Electrical Engineer Award in 1996. He also won a Thomas Alva Edison Patent Awards in 2005 and 2009. His research interests include mobility management, signaling and control for wireless networks, security for wireless systems, mobile data systems, and protocol design. Dr. La Porta was the founding Editor-in-Chief of the IEEE Transactions on Mobile Computing and served as Editor-in-Chief of IEEE Personal Communications Magazine.

    Ramesh Govindan received his B. Tech. degree from the Indian Institute of Technology at Madras, and his M.S. and Ph.D. degrees from the University of California at Berkeley. He is a Professor in the Computer Science Department at the University of Southern California. His research interests include routing and measurements in large internets, wireless sensor networks, and mobile computing systems.

    This work was presented in part at the 14th International Conference on Information Fusion, FUSION 2011, July 2011.

    View full text