The 2010 International Conference on Advanced Technologies for Communications
WiSeCoMaSys: a Tool for Data Collection and Management of Wireless Sensor Networks Andreas Timm-Giel Communication Networks Hamburg University of Technology Schwarzenbergstr. 95E, 21073-Hamburg, Germany
[email protected]
Vo Que Son, Bernd-Ludwig Wenning, Carmelita Görg Communication Networks, University of Bremen Otto-Hahn-Allee 1, 28359-Bremen, Germany {son, wenn, cg}@comnets.uni-bremen.de
measure statistic parameters such as packet reception rate (PRR) or end-to-end delay, which are necessary for optimizing network performance. SWAT [4] is integrated with an SQL server to enable users to measure the performance of sensor networks and to visually display the results in reports. Borrowing ideas from SWAT and Octopus, this paper work develops a component-based tool which allows users to collect the data packets from a sensor network to analyze and visualize them. In addition, it also measures many statistic parameters such as PRR and end-to-end delay for evaluating the networks. Developers can add or modify this tool easily based on their applications because it is designed in separate and open components. This paper is structured in 5 sections: section I is the introduction to the scope of this paper, and the description of the WiSeCoMaSys architecture is discussed in section II. Section III describes the functionalities and formats of packets. Experiment measurements using this tool are shown in section IV. Finally, conclusions are given at the end of this paper.
Abstract— Currently, many specialized tools for monitoring and control of sensor networks are existing and being used separately. In this paper, an architecture of a component-based tool, named WiSeCoMaSys, is presented to integrate data monitoring from sensor networks, visualization, measurements and network management in a unique system. WiSeCoMaSys helps developers to debug their network deployments easily in many scenarios. Meanwhile, users can change a variety of parameters at any layer of the sensor node architecture to respond to network changes or optimize the performance. Moreover, statistic functionalities are also integrated in this tool to measure and evaluate the real time status of the entire network. With a completely integrated architecture, WiSeCoMaSys is a promising tool to interact with sensor networks visually in many sensing applications.
I.
INTRODUCTION
With the development of digital electronics, low-cost, lowpower, distributed processing sensor nodes have been proposed for use in a wide range of applications such as environmental monitoring and environment observation. Featuring sensing, computation and communication capabilities, such as ad-hoc networking and distributed processing, Wireless Sensor Networks (WSNs) allow telemetry, information collection and information management, which can be suitable for many applications. Different from other conventional networks, the monitoring of WSNs involves monitoring of both the network state and sensing parameters. Besides, controlling WSNs includes controlling the network configuration and the settings of sensing applications. Hence, the need for an integrated powerful tool to monitor the whole sensor network and manage their configuration is indeed necessary. There have been many research works to develop the tools to monitor the status of sensor networks and to also control the network configuration. MViz [1] is a TinyOSv.2 [18] tool which is used to display the data visually. However, this tool cannot configure the deployed networks. Surge [3] is another tool which allows users to display the connectivity of networks and to conduct some simple configuration such as changing the sampling frequency; however it is only used in TinyOS v.1. Octopus [1] is an advanced solution that supports three operation modes or can reconfigure the network. However, Octopus is a general tool; therefore, it does not support context-aware sensing applications and cannot
978-1-4244-8876-6/10/$26.00 ©2010 IEEE
II.
WIRELESS SENSOR NETWORK COLLECTION AND MANAGEMENT SYSTEM (WISECOMASYS)
WiSeCoMaSys, as its name says, is a tool which can be used to pull data out of sensor networks, visualize the received data and manage the networks over the air from the users’ side. In order to integrate the monitoring, measurement and management into one system, the goals of our design consist of the following key points: • Support modularized components for easy modification. • Collect data from the network, analyze and visualize them in tables and charts. • Log received data to files for later in-depth analysis. • Support context-aware sensing applications [6]. • Support centralized management and control per sensor node. Decisions about changing network parameters are made by WiSeCoMaSys, which has a comprehensive view of the whole network. • Measure statistics parameters such as PRR or message rate. • Operate at the gateway or via internet using serial communication features [8]. • Support alerts to users by email or SMS (Short Message Service) for critical events. WiSeCoMaSys consists of a high-level tool developed in
33
Java and an embedded nesC-based [16] application uploaded onto the sensor nodes and the gateway. The high-level tool architecture of WiSeCoMaSys is shown in Figure 1 with several main parts as follows: • Packet Monitor listens to the serial port for incoming packets from a sensor network. When it receives a packet, it will update the database based on the type of the packet (data packet or control packet). • Packet Injector puts the required command from Network Control in a request packet and sends this packet to the destination node. • Database stores all information of nodes and interfaces with other components. • Alert checks the pre-defined conditions and sends warning messages to users when a condition is matched. • Logger records all pre-defined values of nodes to a file after the database is updated. • The GUI has many panels to provide the interface between users and the network. It includes the following parts: o Statistics Measurements contains processes which calculate packet reception rate, message rate, etc. o Topology Viewer displays the network topology with related metrics. o Data Display shows all the information extracted from packets and measurements in a table. o Graph Display shows the user-understandable values after analyzing such as temperature, humidity, or battery level of each node in charts. o Network Control allows users to manage each node by sending specific commands to that node. o Network Status displays the status of the network and configuration of nodes. o Setting is the interface where users can change the parameters of the logging process, and customize displayed parameters of the network topology, etc.
a node located in a specific place to the sink via a gateway. The embedded application running in the sensor nodes uses the multi-hop opportunistic routing protocol published in [5] to forward packets from the originating nodes to the gateway via intermediate nodes, and these packets will be captured by WiSeCoMaSys. It also supports configuring the context-aware sensing application described in [6] and the localization techniques described in [7]. There are two kinds of messages used for data collection in networks: data packet for collecting the sensed values from sensor nodes, control packet for replying to a request from users. 1) Data packet The format of a data packet includes three following parts: • Routing information: includes several fields for routing: o Sink (2 bytes): the address of the destination sink o TTL (time-to-live) (1 byte): used to eliminate loop problems in the network. • Application information: contains the following fields: o Mote (2 bytes): the address of the sensor node which originates this packet. This field is also used by the routing layer for cycle suppression. o Sequence (2 bytes): used to recognize and avoid duplicated packets. o Time (4 bytes): a time stamp embedded in each message to indicate the global sending time. o Type (1 byte): an indicator of whether the receiving packet is a data packet or a control packet. o Reading (8 bytes): used to transmit temperature, humidity, light and internal voltage of sensor nodes. This field is based on the four sensors of a standard TelosB mote [9]. o Position (6 bytes): the location of the sender, which is given by the distributed localization process in [7]. • Management (10 bytes) contains the management information as following: o Hop count (1 byte): the number of intermediate nodes through which a packet has to travel to reach WiSeCoMaSys o Next-hop node (2 bytes): the address of the next-hop to which a packet is forwarded. This next-hop is selected by routing protocol [5]. o Backup next-hop node (2 bytes): the address of a back node which is used to forward packets in case the Nexthop node has loop problems [5]. o Receive Signal Strength Indicator (1 byte): used to measure the link quality [9] [5]. o Energy consumption (4 bytes): the current energy consumed in a sensor node. 2) Control packet Because the directions of data packets and control packets are similar (from sensor nodes to gateway), control data can be put in the field Reading of the data packet format with a specific reply mask. When a node receives a request, if the command in this request needs a reply, it will put the corresponding information in the data packet and a specific value of that request in the Reply field to complete a control packet. Then it sends this packet to the gateway using the
Figure 1: Architecture of WiSeCoMaSys.
WiSeCoMaSys can be used in many applications such as: • Habitat monitoring [14]. • Environmental condition monitoring in data warehouses, storages or harbors [5]. • Tracking and monitoring of logistic items inside containers [10] [16]. III.
COLLECTION, MEASUREMENT AND MANAGEMENT
A. Data collection Data collection is an important feature because most sensing applications need to transmit sensed physical parameters from
34
routing protocol. Based on the Reply field, WiSeCoMaSys can extract the information in the received control packet correctly. Data collection can be also used for debugging such as tracing a variable in a process of a node. With WiSeCoMaSys, developers can declare any field in the header of data packets or control packets to carry debugging information. At the gateway, WiSeCoMaSys will capture and display this debugging information easily. In addition, the format of the data and control packet can be easily changed by using Message Interface Generator [19] with the struct definitions of message types in a specific embedded nesC-based application.
In flat WSNs, WiSeCoMaSys can access WSNs via a gateway, which bridges the information between WiSeCoMaSys and the WSN. In Figure 2, looking from the outside, each sensor node in network has a management service that is responsible for managing the node configuration. This service communicates with WiSeCoMaSys using requests and replies and it works as a part of the application in a sensor node. The management information could be local node status, link quality, node configuration, etc. 1) Command dissemination Reconfiguration of the network requires sending information from the gateway to nodes inside the sensor network. The configuration is put in a specific command for dissemination. In order to send a command from WiSeCoMaSys to the network, the request packet is used for the purpose of carrying a command from user’s side to a specific node using [17]. The command can also be sent to a group of nodes or all nodes in the whole network based on the selection from the user. For reliable node configuration, a command can require a reply from a node. The format of the request packet has the following fields: • Address (2 bytes): the address of the destination sensor node or the broadcast address which allows all nodes to receive the packet. • Request (1 byte): is used to identify the command. • Parameter (8 bytes) contains the parameter information needed for the command. Table 2 shows the configuration which can be changed by commands at each layer. There are also commands to retrieve the configuration of each node from the sensor network.
B. Statistic measurements In order to have a deep understanding of the network operation, besides the real time status information collected from the network, the Statistics Measurement module also provides some useful statistic parameters that need to be measured over a longer period. Many parameters shown in Table 1 are supported by WiSeCoMaSys. These parameters are necessary for the user to evaluate the performance or optimize the configuration. Other parameters, which are not included in WiSeCoMaSys, can be easily added to the Statistics Measurement component. TABLE 1 PARAMETERS MEASURED BY WISECOMASYS Metric
Explanation
Packet Reception Rate (%)
The ratio of the number of successfully received packets to the number of sent packets
Packet rate (message/s)
The number of packets received per second
End-to-End delay (ms)
Delay between the time of sending this packet and the time of receiving this packet at WiSeCoMaSys
TABLE 2 CONFIGURATION SUPPORTED BY WISECOMASYS Layer PHY
The number of intermediate nodes through which a packet has to travel to reach WiSeCoMaSys
MAC
Power consumption (mW)
Power consumed in a sensor node
Routing
Remaining battery (%)
Remaining capacity of two AA batteries, which supply power for each node.
Hop count
Configuration - Power level settings [9] - Radio channel [9] - Duty cycle [13] - Soft/hard-Acknowledgement [12] - Beacon power [11] - Beacon period [5] - Data sampling period - Auto/Query/Context-aware mode [1] [6] - Context-aware rules [6]
C. Management
- Awake/sleep mode [1] - Localization parameters, localization techniques [7] - Number of transmission retries and retry delay Application
- Time synchronization [5] - Reset network/nodes remotely - Logistics information (package, container identifiers) when sensor networks are deployed inside container for tracking logistic items [5] [10] - Sensing services provide which sensors are available in nodes and can be used for measurements. - Memory management is used for storing configuration or buffering the packets when necessary
Figure 2: WiSeCoMaSys and management service in flat WSNs.
35
be also displayed in this panel. In this figure, the node labeled GW is the gateway, the gray nodes are the anchor nodes and the black ones are the localized nodes which can perform the localization process [7] based on the information provided from anchor nodes. The circle of node 9 indicates that this node is currently selected. A physical map can be loaded from an image file to be displayed in the background.
2) Alert mechanisms WiSeCoMaSys also has an interface for users to configure the alarm settings using logical conditions. These alarms are set for each sensing value and will trigger one of the following corresponding actions when the condition is matched: • Showing a message on the screen, • Sending an alert email to an address list, • Sending an alert SMS to a mobile number. 3) Remote access over Internet In some scenarios, users may require access to the sensor network from a remote location. For example, if the sensor network is deployed inside a container to monitor the environmental conditions of goods, the user can access the network via infrastructure networks (e.g. WLAN, GPRS). WiSeCoMaSys supports remote communication by using multiple SerialForwarders in TinyOS [18], which work as a client-server model. Figure 3 shows the remote access between WiSeCoMaSys and the sensor network. There are two instances of SerialForwarder: one runs at the gateway and the other runs at the remote location. These instances communicate with each other via TCP/IP networks. Hence, WiSeCoMaSys connects to the SerialForwarder in the remote location to receive incoming packets forwarded by the first SerialForwarder.
Figure 4: The network topology is built by WiSeCoMaSys.
2) Data visualization In Figure 5, the Data Display interface shows all the real data collected from networks in a summarized table. Users can set the thresholds of each parameter for alert such as temperature, humidity, etc. by generating a logic condition in the Alarm Setting part. The corresponding actions have also been selected from a list box. The Alert component checks these conditions in a pre-defined period. If a condition matches, it will execute the corresponding given action such as sending an alert SMS. This allows users to keep informed about critical events from the network in emergency cases, when users cannot access WiSeCoMaSys directly.
Figure 3: Remote access over Internet using multiple SerialForwarders.
IV.
EXPERIMENTS IN A LIVE WSN
A. Deployment of a live WSN A real deployment of WSNs is set up in NW1 building, (ComNets, University of Bremen) to monitor the environmental conditions in the room such as temperature and humidity. The deployment includes 14 TelosB motes [9] that are powered by two full AA batteries each to form a sensor network. Using WiSeCoMaSys, all nodes are configured to send their data packets every 10 seconds. In the next sections, the previously described features of WiSeCoMaSys are shown through cases. B. Monitoring, Measurement and Management 1) Topology and connectivity When receiving a packet from the sensor network (data or control packet), WiSeCoMaSys extracts the Management information in the header of the packet to build a map of the topology. Users are free to choose what they want to display in this panel. Figure 4 shows an example of the network topology captured by WiSeCoMaSys. Each link between two nodes is described by a solid line with the RSSI value. Other parameters such as sampling period, backup next-hop [5] can
Figure 5: Data Display panel in WiSeCoMaSys.
36
The Graph panel in WiSeCoMaSys can also plot the recent sensing data of multiple nodes in real time. Users can select which readings and which nodes they want to display in this Graph panel. The plots can be zoomed in/out or changed at many time scales. 3) Measurements When WiSeCoMaSys receives a packet, its Statistics Measurements component calculates or updates statistics parameters. The Statistics panel provides many types of parameters for display such as data rate, end-to-end delay, and power consumption. For example, Figure 6 illustrates the real time PRR of nodes in the experiment. Most of them have a good PRR which is over 90% and some of them (e.g. node 8 and node 1) have a normal PRR (approximately 85%).
which needs to be reconfigured in the Topology panel and after that, users have to choose and send a required command to the chosen node. The information about this action is displayed in a console window. The context-aware sensing application [6] can be also configured in this module to save energy or to improve the PRR because it can reduce the traffic in the network. 5) Context-aware sensing application The ability to be aware of the contexts can make the sensor nodes more intelligent to reduce the redundant generated traffic over time. A context-aware sensing model [6] is implemented in WiSeCoMaSys to allow users to set the context rules for each sensor node to send data packets when a rule is matched. The supported sources of context can be temperature, humidity, light, battery, and the trigger time to send data packets.
Figure 6: PRR is measured and plotted by WiSeCoMaSys. Figure 8: Number of receiving packets in context-aware mode [6].
4) Network management
For example, in Figure 8, node 13 is set to run in contextaware mode with the rule: If Temperature is greater than 38 (°C) then send packets. One can see that node 13 strictly follows the pre-defined context rule above because the number of received packets only increases when the condition is satisfied. Context-aware sensing applications also save energy of the whole network to prolong the operation time. C. Alert for users As mentioned previously, WiSeCoMaSys can warn users when there are errors or unexpected events happening with the sensor network if the alerts are set. This allows users to be informed about critical alerts from the network in emergency cases, when users cannot access WiSeCoMaSys directly. For example, in Figure 9, an alert SMS message is sent to a mobile number to notify about a low sensed light strength from the sensor node 2 in the SMS.
Figure 7: Network Status and Control interface.
With the Network Status and Control modules shown in Figure 7, users can get and see all the settings of multiple nodes or an individual node in the network. These parameters are summarized in a table. In order to reconfigure a parameter for a node, the user has to do two steps: selecting the node
37
Figure 9: An alert SMS.
V.
[16] A.Timm-Giel, K. Kuladinithi, M. Becker, C. Görg: “Wireless Sensor Networks in Wearable and Logistic Application” - CRUISE Workshop, page 2, Greece, June, 2006. [17] Dissemination - TEP118, http://www.tinyos.net/tinyos-2.x/doc [18] TinyOS – http://www.tinyos.net [19] MIG, http://www.tinyos.net/tinyos-1.x/doc/nesc/mig.html
CONCLUSIONS AND OUTLOOK
In summary, this paper presents an advanced tool, WiSeCoMaSys, to monitor, analyze, visualize, and manage sensor networks. With the help of this tool, it is believed that developers can save time in debugging and testing their deployments. Having a rich set of features, it also allows users to evaluate the performance of their applications easily. Adaptive control is an interesting issue for improvement. Based on the network information collected, an improved version of WiSeCoMaSys could be able to automatically perform necessary computation and adjust the reconfiguration for nodes to enhance a better performance. Another issue for future work is to apply standards such as 6LowPAN or ZigBee stack in this tool. REFERENCES [1]
[2] [3] [4]
[5]
[6]
[7]
[8] [9] [10]
[11]
[12] [13]
[14]
[15]
A.G.Ruzzelli, R.Jurdak, M.Dragone, A.Barbirato, G.M.P.O’Hare, S.Boivineau,V.Roy: “Octopus: A Dashboard for Sensor Networks Visual Control”. In 14th International Conferenceon Mobicom Demos, SanFrancisco, CA, USA, September 2008. MViz - http://www.tinyos.net/tinyos-2.x/apps/MViz/ Surge Demo: http://firebug.sourceforge.net/surgepics.html K. Srinivasan, M. A. Kazandjieva, M. Jain, E. Kim, P. Levis:” SWAT: enabling wireless network measurements”. In Proceedings of the 6th ACM conference on Embedded network sensor systems, pp. 395396, Raleigh, NC, USA, November 2008. V.Q. Son, B.L. Wenning, A. Timm-Giel, C. Görg, “A Model of Wireless Sensor Networks using Opportunistic Routing in Logistic Harbor Scenarios”. In Proceedings of LDIC2009 Conference, pp. 214-223, Bremen, Germany, August 2009. V.Q. Son, B.L. Wenning, A. Timm-Giel, C. Görg: “A model of Wireless Sensor Networks using Context-awareness in Logistics Applications”. In Proceedings of the ITST2009 Conference, pp. 2-7, Lille, France, October 2009. V.Q. Son, B.L. Wenning, A. Timm-Giel, C. Görg: “Localization using Opportunistic Routing in Wireless Sensor Network for Logistics”. In Proceedings of the ASOR2009 Conference, CD-ROM publication, Queensland, Australia, September 2009. Serial Communication - TEP113, http://www.tinyos.net/tinyos-2.x/doc TelosB datashseet - available at http://www.xbow.com. R. Jedermann, C. Behrens, R. Laur , W. Lang, “Intelligent Containers and Sensor Networks Approaches to apply Autonomous Cooperation on Systems with limited Resources”. In Understanding Autonomous Cooperation & Control in Logistics – The Impact on Management, Information and Communication and Material Flow. Springer, , pp. 365392, Berlin, 2007. B.-L. Wenning, A. Lukosius, A. Timm-Giel, C. Görg, S. Tomic: “Opportunistic Distance-Aware Routing in Multi-Sink Mobile Wireless Sensor Networks”. In Proceedings of ICT-MobileSummit 2008, CDROM publication, Stockholm, 2008. Packet Link Layer – TEP127, http://www.tinyos.net/tinyos-2.x/doc J. Polastre, J. Hill, D. Culler: “Versatile low power media access for wireless sensor networks”. In Proceedings of the 2nd international conference on Embedded networked sensor systems, pp. 95–107, Baltimore, MD, USA, 2004. B.L. Wenning, D. Pesch, A. Timm-Giel, C. Görg, “Environmental Monitoring Aware Routing in Wireless Sensor Networks”. In Proceedings of the IFIP International Federation for Information Processing, Volume 284/2008, pp. 5-16, Toulouse, France, September 2008 D. Gay, P. Levis, R. Behren, M. Welsh, E. Brewer, D. Culler: “The nesC Language: A Holistic Approach to Networked Embedded Systems”. In Proceedings of Programming Language Design and Implementation (PLDI) 2003, pp. 1-11 , San Diego, California, USA, 2003.
38
