Effective debugging usually involves watching program state to diagnose bugs. When debugging sensor network applications, this approach is often time-consuming and errorprone, not only because of the lack of visibility into system state, but also because of the difficulty to watch the right variables at the right time. In this paper, we present declarative tracepoints, a debugging system that allows the user to insert a group of action-associated checkpoints, or tracepoints, to applications being debugged at runtime. Tracepoints do not require modifying application source code. Instead, they are written in a declarative, SQL-like language called TraceSQL independently. By triggering the associated actions when these checkpoints are reached, this system automates the debugging process by removing the human from the loop. We show that declarative tracepoints are able to express the core functionality of a range of previously isolated debugging techniques, such as EnviroLog, NodeMD, Sympathy, and StackGuard. We describe the design and implementation of the declarative tracepoints system, evaluate its overhead in terms of CPU slowdown, illustrate its expressiveness through the aforementioned debugging techniques, and finally demonstrate that it can be used to detect real bugs using case studies of three bugs based on the development of the LiteOS operating system.

Network protocols are typically designed and tested individually. In practice, however, applications use multiple protocols concurrently. This discrepancy can lead to failures from unanticipated interactions between protocols. In this paper, we argue that sensor network communication stacks should have an isolation layer, whose purpose is to make each protocol’s perception of the wireless channel independent of what other protocols are running. We identify two key mechanisms the isolation layer must provide: shared collision avoidance and fair channel allocation. We present an example design of an isolation layer that builds on the existing algorithms of grant-to-send and fair queueing. However, the complexities of wireless make these mechanisms insufficient by themselves. We therefore propose two new mechanisms that address these limitations: channel decay and fair cancellation. Incorporating these new mechanisms reduces the increase in end-to-end delivery cost associated with concurrently operating two protocols by more than 60%. The isolation layer improves median protocol fairness from 0.52 to 0.96 in Jain’s fairness index. Together, these results show that using an isolation layer makes protocols more efficient and robust.

Current data collection protocols for wireless sensor networks are mostly based on quasi-static minimum-cost routing trees. We consider an alternative, highly-agile approach called backpressure routing, in which routing and forwarding decisions are made on a per-packet basis. Although there is a considerable theoretical literature on backpressure routing, it has not been implemented on practical systems to date due to concerns about packet looping, the effect of link losses, large packet delays, and scalability. Addressing these concerns, we present the Backpressure Collection Protocol (BCP) for sensor networks, the first ever implementation of dynamic backpressure routing in wireless networks. In particular, we demonstrate for the first time that replacing the traditional FIFO queue service in backpressure routing with LIFO queues reduces the average end-to-end packet delays for delivered packets drastically (75 % under high load, 98 % under low load). Further, we improve backpressure scalability by introducing a new concept of floating queues into the backpressure framework. Under static network settings, BCP shows a more than 60 % improvement in max-min rate over the state of the art Collection Tree Protocol (CTP). We also empirically demonstrate the superior delivery performance of BCP in highly dynamic network settings, including conditions of extreme external interference and highly mobile sinks. 1.

Abstract — When deployed in a real-world setting, many sensor networks fail to meet application requirements even though they have been tested in the lab prior to deployment. Hence, concepts and tools for inspection are needed to identify failure causes in situ on the deployment site. Tools for inspection should minimize the interference with the sensor network to, firstly, ensure that failures of the sensor network do not break the inspection mechanism, and, secondly, to ensure that the inspection mechanism does not change the behavior of the sensor network. In this paper, we propose passive distributed assertions (PDA) as a novel tool for identifying failure causes. PDA allow a programmer to assert certain predicates over distributed node states. Packet sniffing is used to detect failed assertions, thus minimizing the interference with the sensor network. 1 I.

When deployed in a real-world setting, many sensor networks fail to meet application requirements even though they have been tested in the lab prior to deployment. Hence, concepts and tools for inspection are needed to identify failure causes in situ on the deployment site. Tools for inspection should minimize the interference with the sensor network to, firstly, ensure that failures of the sensor network do not break the inspection mechanism, and, secondly, to ensure that the inspection mechanism does not change the behavior of the sensor network. In this position paper, we propose passive distributed assertions (PDA) as a novel tool for identifying failure causes. PDA allow a programmer to assert certain predicates over distributed node states. Packet sniffing is used to detect failed assertions, thus minimizing the interference with the sensor networks.

Abstract. Remote management is essential for wireless sensor networks (WSNs) designed to run perpetually using harvested energy. A natural division of function for managing WSNs is to employ both an in-band data plane to sense, store, process, and forward data, and an out-of-band management plane to remotely control each node and its sensors. This paper presents SRCP, a Simple Remote Control Protocol that forms the core of an out-of-band management plane for WSNs. SRCP is motivated by our target environment: a perpetual deployment of high-power, aggressively duty-cycled nodes capable of handling high-bandwidth sensor data from multiple sensors. The protocol runs on low-power always-on control processors using harvested energy, distills an essential set of primitives, and uses them to control a suite of existing management functions on more powerful main nodes. We demonstrate SRCP’s utility by presenting a case study that (i) uses it to control a broad spectrum of management functions and (ii) quantifies its efficacy and performance. 1

Abstract—Wireless sensor networks are envisioned to be an integral part of cyber-physical systems, yet wireless networks are inherently dynamic and come with varieties of uncertainties. One such uncertainty is wireless communication itself which assumes complex spatial and temporal dynamics. For dependable and predictable performance, therefore, link estimation has become a basic element of wireless network routing. Several approaches using broadcast beacons and/or unicast MAC feedback have been proposed in the past years, but there still lacks as a systematic characterization of the drawbacks and sources of errors in beacon-based link estimation in low-power wireless networks, which leads to ad hoc usage of beacons in routing. Using a testbed of 98 XSM motes (an enhanced version of MICA2 motes), we characterize the negative impact that link layer retransmission and traffic-induced interference have on the accuracy of beacon-based link estimation, and we show that data-driven link estimation and routing achieves higher event reliability (e.g., by up to 18.75%) and transmission efficiency (e.g., by up to a factor of 1.96) than beacon-based approaches. These findings provide solid evidence for the necessity of data-driven link estimation and demonstrate the importance of addressing the drawbacks of beacon-based link estimation when designing protocols for low-power wireless networks of cyber-physical systems. Index Terms—Low-power wireless networks, sensor networks, link estimation and routing, data-driven, beacon-based I.

Complex interactions and the distributed nature of wireless sensor networks make automated testing and debugging before deployment a necessity. A main challenge is to detect bugs that occur due to non-deterministic events, such as node reboots or packet duplicates. Often, these events have the potential to drive a sensor network and its applications into corner-case situations, exhibiting bugs that are hard to detect using existing testing and debugging techniques. In this paper, we present KleeNet, a debugging environment that effectively discovers such bugs before deployment. KleeNet executes unmodified sensor network applications on symbolic input and automatically injects non-deterministic failures. As a result, KleeNet generates distributed execution paths at high-coverage, including low-probability cornercase situations. As a case study, we integrated KleeNet into the Contiki OS and show its effectiveness by detecting four insidious bugs in the µIP TCP/IP protocol stack. One of these bugs is critical and lead to refusal of further connections.

Abstract—We describe grant-to-send, a novel collision avoidance algorithm for wireless mesh networks. Rather than announce packets it intends to send, a node using grant-to-send announces packets it expects to hear others send. We present evidence that inverting collision avoidance in this way greatly improves wireless mesh performance. Evaluating four protocols from 802.11 meshes and 802.15.4 sensor networks, we find that grant-to-send matches or outperforms CSMA and RTS/CTS in all cases. For example, in a 4-hop UDP flow, grantto-send can achieve 96 % of the theoretical maximum throughput while maintaining a 99.9 % packet delivery ratio. Grant-tosend is also general enough to replace protocol-specific collision avoidance mechanisms common to sensor network protocols. Grant-to-send is simple. For example, incorporating it into 802.11 requires only 11 lines of driver code and no hardware changes. Furthermore, as it reuses existing 802.11 mechanisms, grant-to-send inter-operates with current networks and can be incrementally deployed. I.

Abstract—Link estimation is a basic element of routing in low-power wireless networks, and data-driven link estimation using unicast MAC feedback has been shown to outperform broadcast-beacon based link estimation. Nonetheless, little is known about how different data-driven link estimation methods affect routing behaviors. To address this issue, we classify existing data-driven link estimation methods into two broad categories: L-NT that uses aggregate information about unicast and L-ETX that uses information about the individual unicast-physical transmissions. Through mathematical analysis and experimental measurement in a testbed of 98 XSM motes (an enhanced version of MICA2 motes), we examine the accuracy and stability of L-NT and L-ETX in estimating the ETX routing metric. We also experimentally study the routing performance of L-NT and L-ETX. We discover that these two representative, seemingly similar methods of data-driven link estimation differ significantly in routing behaviors: L-ETX is much more accurate and stable than L-NT in estimating the ETX metric, and, accordingly, L-ETX achieves a higher data delivery reliability and energy efficiency than L-NT (for instance, by 25.18 % and a factor of 3.75 respectively in our testbed). These findings provide new insight into the subtle design issues in data-driven link estimation that significantly impact the reliability, stability, and efficiency of wireless routing, thus shedding light on how to design link estimation methods for mission-critical wireless networks which pose stringent requirements on reliability and predictability. Index Terms—Low-power wireless networks, sensor networks, link estimation and routing, data-driven, beacon-based, distancevector routing, geographic routing I.