Elsevier

Journal of Systems Architecture

Volume 72, January 2017, Pages 19-28
Journal of Systems Architecture

Highly-Efficient Bulk Data Transfer for Structured Dissemination in Wireless Embedded Network Systems

https://doi.org/10.1016/j.sysarc.2016.09.001Get rights and content

Highlights

  • We identify the limitations of the existing structured dissemination protocols in wireless embedded network systems (WENS) and propose an out-of-order, XOR enabled transmission mechanism for structured dissemination.

  • We propose a slot length optimization model, tailored for the proposed full-slot transmission mechanism to further enhance the propagation.

  • We incorporate the above components into ULTRA, a full-slot reliable bulk data transfer mechanism for structured bulk data dissemination.

  • We implement ULTRA in both testbed and simulation. The results show that ULTRA outperforms the state-of-the-art works.

Abstract

Recent years have witnessed the remarkable development of wireless embedded network systems (WENS) such as cyber-physical systems and sensor networks. Reliable bulk data dissemination is an important building module in WENS, supporting various applications, e.g., remote software update, video distribution. The existing studies often construct network structures to enable time-slotted multi-hop pipelining for data dissemination. However, the adopted transmission mechanism was originally designed for structureless protocols, and thus posing significant challenges on efficient structured data dissemination. In this paper, we investigate the problem of structured bulk data dissemination. Specifically, we propose reliable out-of-order transmission and bursty encoding mechanisms to transmit packets as many as possible in each transmission slot. As a consequence, the resulting transmission protocol (ULTRA) can fully utilize each transmission slot and propagate data in the network as fast as possible. The performance results obtained from both testbed and simulation experiments demonstrate that, compared to the state-of-the-art protocols, ULTRA can greatly enhance the dissemination performance by reducing the dissemination delay by 34.8%.

Introduction

Wireless embedded network systems (WENS) are composed of a number of low-power embedded devices which are capable of computing and wireless communications [1], [2]. WENS have been widely deployed recently to support Internet-of-things [3], [4], [5], [6]. Bulk data dissemination is used to distribute a large data object reliably from a sink node to all network nodes in WENS, becoming an essential building module for a variety of WENS systems, e.g., remote software management [7], [8], security patches [9], reprogramming [10], [11] and video distribution [12], [13].

The existing studies [14], [15], [16] often employ a Connected Dominating Set (CDS) structure for bulk data dissemination. In such a structure, a set of nodes are selected as core nodes, which are responsible for disseminating the object to the rest nodes that are one hop away from the core nodes. This facilitates the transmission and sleep scheduling to achieve more efficient data propagation. Structured dissemination leads to less broadcast overhead as compared to structureless dissemination which is prone to the broadcast storm problem, and hence, offers a good solution for dense and low-power WENS.

Specifically, a data object is divided into small pages, each of which consists of a number of packets [14], [15], [16]. Time is sliced into fixed slots, and the slot length approximates the transmission time of a page. Each node operates in three types of slots for transmission (T slot), reception (R slot) and sleep (S slot), respectively. Then the data propagation is done in a page-by-page manner: a node starts transmitting the next page only when the current page is entirely received by all its child nodes in the structure. The page-by-page design is motivated by two reasons. First, it enables multi-hop pipelining. Different pages can be simultaneously transmitted at different hops, thus reducing the overall propagation delay. Second, to establish retransmission and ensure eventual reliability, a bitmap indicating the missing packets is carried by a request message (REQ). The bitmap for the entire object may be too large for the limited payload length (e.g., 127 octets in 802.15.4 packets). Therefore, the bitmap for a page is much smaller and can be included in an REQ message. The eventual reliability is achieved when all pages are received.

While the above mechanism can roughly establish reliable page-level pipeline for structured protocols, it suffers from the following key limitations. First, the transmission slots are not fully utilized. In current dissemination protocols, a node transmits the packets requested by its child nodes. Although the slot length approximates a page transmission time, the slots are rarely full of transmissions since the requested packets in REQs are often fewer than the page size. Second, the three-way handshake, which works in the Advertisement-Request-Data (ADV-REQ-DATA) paradigm, produces redundant controlling overhead. The ADV messages are designed for two purposes: neighbor discovery and new data page declaration. However, when applied in the structured dissemination, neighbor discovery is no longer necessary since each node has fixed parent and children. If we disable the page-by-page transmission and transmit out-of-order packets to fully utilize the slots, the data page declaration is neither necessary.

In this paper, we investigate the transmission for structured data dissemination and propose a full-slot reliable transmission mechanism (called ULTRA) for bulk dissemination, which fully exploit the benefits of the structure. ULTRA has three salient features. 1) We break the page-by-page transmission manner and enable out-of-order packets transmission. Instead of only transmitting the packets within one page, a node transmits as many packets as possible in each slot. 2) ULTRA adopts a bursty coding scheme to fully utilize the transmission slots when there are not enough native packets during the propagation, which effectively recovers packet errors without the need of extra transmission rounds. While the above two designs greatly improve the data propagation speed, they pose new challenges in terms of reliability. 3) We propose a slot optimization model, specifically for the proposed transmission mechanism to minimize the propagation delay.

We implement ULTRA in low power wireless embedded TelosB nodes and conduct both testbed and simulation experiments. The evaluation results show that ULTRA outperforms the state-of-the-art bulk data dissemination mechanisms and greatly enhances the dissemination performance by reducing completion delay by 34.8%.

The major contributions of this paper are summarized as follows:

  • We investigate the limitations of the existing structured dissemination protocols in WENS and propose an out-of-order, XOR enabled transmission mechanism to fully utilize the transmission slots.

  • We propose a slot length optimization model, tailored for the proposed full-slot transmission mechanism to further enhance the propagation.

  • Based on the above components, we propose a holistic solution, i.e., full-slot reliable transmission for structured bulk data dissemination (ULTRA). Compared to the existing protocols, ULTRA can fully utilize each transmission slot and greatly reduces the negotiation overhead.

  • We implement ULTRA in both TelosB [17] testbed and TOSSIM [18], [19] simulation. The performance results show that ULTRA greatly outperform the state-of-the-art protocols in terms of the end-to-end delay.

The remainder of the paper is organized as follows. Section 2 summarizes the related works and compares ULTRA with them. Section 3 presents the motivation of our work. Section 3.2 identifies the design opportunities for efficient dissemination. Section 4 presents the main design of the ULTRA protocol. Section 5 evaluates ULTRA via testbed and simulation experiments. Section 6 concludes the work.

Section snippets

Related Works

The existing bulk data dissemination protocols can be mainly classified into two categories: structureless protocols and structured protocols.

Motivation and Design Opportunities

In most of the structured dissemination protocols for WENS, all network nodes first establish an underlying structure and start dissemination along the structure (for stationary networks). These protocols exploit various link characteristics to construct the core structure [14], [15], [16], [28], [29], [30] and ensure reliability using the three-way handshake mechanism, where a sender node first broadcasts ADV messages, declaring its source ID and the data it can provide. A receiver node which

Main Design of ULTRA

In this section, we present the design of ULTRA. We first give the high level overview and then present each building block of ULTRA in detail.

Performance Evaluation and Analysis

In this section, we conduct testbed and simulation experiments to study the performance of ULTRA, in comparison with the state-of-the-art work CoCo [16]. After the testbed study, we tune the link quality in the TOSSIM simulation to study ULTRA’s performance under different network conditions.

Conclusion

In this paper, we investigate the problem of structured bulk data dissemination in WENS and propose a full-slot reliable transmission mechanism (ULTRA) that can fully exploit the benefits of the established structures. Compared to the three-way handshake mechanism, ULTRA requires only one round negotiation and can fully exploit each transmission slot. Besides, the slot length is optimized for fast data dissemination. We conduct both testbed and simulation experiments to study the performance of

Acknowledgement

This work is supported by the National Science Foundation of China under Grant No. 61602095 and the Fundamental Research Funds for the Central Universities under Grant No. ZYGX2016KYQD098.

Zhiwei Zhao is currently an assistant professor at the College of Computer Science and Engineering in University of Electronic Science and Technology of China. He received his PhD degree at the College of Computer Science, Zhejiang University in 2015. His research interests include wireless computing, heterogeneous wireless networks, protocol design and network coding.

References (37)

  • M. Stolikj et al.

    Efficient reprogramming of wireless sensor networks using incremental updates

    in Proc. of PerCom

    (2013)
  • C. Miller et al.

    Reliable and efficient reprogramming in sensor networks

    ACM Transactions on Sensor Networks (TOSN)

    (2010)
  • J. Subramanian et al.

    UFlood: High-throughput flooding over wireless mesh networks

    Proceedings of IEEE INFOCOM

    (2012)
  • V. Naik et al.

    Sprinkler: A reliable and energy efficient data dissemination service for extreme scale wireless networks of embedded devices

    IEEE Transactions on Mobile Computing

    (2007)
  • L. Huang et al.

    CORD: Energy-efficient Reliable Bulk Data Dissemination in Sensor Networks

    Proceedings of IEEE INFOCOM

    (2008)
  • Z. Zhao et al.

    Exploiting link correlation for core-based dissemination in wireless sensor networks

    Proceedings of IEEE SECON

    (2014)
  • T. Datasheet, Crossbow...
  • P. Levis, N. Lee, M. Welsh, D. Culler, TOSSIM: Accurate and scalable simulation of entire TinyOS applications, in:...
  • Cited by (0)

    Zhiwei Zhao is currently an assistant professor at the College of Computer Science and Engineering in University of Electronic Science and Technology of China. He received his PhD degree at the College of Computer Science, Zhejiang University in 2015. His research interests include wireless computing, heterogeneous wireless networks, protocol design and network coding.

    Zi Wang received his BS degree at the College of Computer Science and Engineering at the College of Computer Science and Engineering in University of Electronic Science and Technology of China (UESTC). He is currently a MS student in UESTC. His research interests focus on wireless computing.

    Geyong Min is the Chair Professor and Director of High Performance Computing and Networking (HPCN) Research Group at the University of Exeter, UK. He received the PhD degree in Computing Science from the University of Glasgow, UK, in 2003, and the B.Sc. degree in Computer Science from Huazhong University of Science and Technology, China, in 1995. He joined the University of Bradford as a Lecturer in 2002, became a Senior Lecturer in 2005 and a Reader in 2007, and was promoted to a Professor in Computer Science in 2012. His main research interests include Next-Generation Internet, Wireless Networks, Mobile Computing, Cloud Computing, Big Data, Multimedia Systems, Information Security, System Modelling and Performance Optimization.

    Yue Cao received his PhD degree from the Institute for Communication Systems (ICS) formerly known as Centre for Communication Systems Research, at University of Surrey, Guildford, UK in 2013. Further to his PhD study, he was a Research Fellow at the ICS. Since October 2016, he has been the Lecturer in Department of Computer Science and Digital Technologies, at Northumbria University, Newcastle upon Tyne, UK. His research interests focus on Delay/Disruption Tolerant Networks, Electric Vehicle (EV) charging management, Information Centric Networking (ICN), Device-to-Device (D2D) communication and Mobile Edge Computing (MEC).

    View full text