OLIMPO, an ad-hoc wireless sensor network simulator for optimal scada-applications

OLIMPO, AN AD-HOC WIRELESS SENSOR NETWORK SIMULATOR FOR OPTIMAL SCADA-APPLICATIONS J. Barbancho, F.J. Molina, C. Le´on, J. Ropero and A. Barbancho Departamento de Tecnolog´ıa Electr´onica, University of Seville, C/ Virgen de ´ Africa, 7, Seville 41011, Spain Phone.: (+034) 954 55 71 92, Fax: (+034) 954 55 28 33. e-mail: [email protected] ABSTRACT This paper introduces OLIMPO, an useful simulation tool for researchers who are developing wireless sensor communication protocols. OLIMPO is a discrete-event simulator design to be easily reconfigured by the user, providing a way to design, develop and test communication protocols. In particular, we have designed a self-organizing wireless sensor network for low data rate. Our premise is that, due to their inherent spread location over large areas, wireless sensor networks are well-suited for SCADA applications, which require relatively simple control and monitoring. To show the facilities of our simulator, we have studied our network protocol with OLIMPO, developing several simulations. The purpose of these simulations is to demonstrate, quantitatively, the capability of our network to support this kind of applications. KEY WORDS Ad-hoc wireless sensor network, multi-hop Packet Radio Network (PRN), Event Discrete Scheduler (EDS), network simulator, SCADA applications. 1 Introduction Wireless data networks have received research interest in recent years due to the desire for connectivity. New multimedia services, such as World Wide Web, e-mail, and data file transfers have been provided by different working groups. Two very well known examples are the 802.11 and the 802.15 Institute of Electrical and Electronics Engineers (IEEE) working groups. 802.11 has designed several protocols defining the Wireless Local Area Network (WLAN) standard [1]. 802.15 works in the design of several standards based on networks that employ no fixed infrastructure and have communication links less than 10 meters (known as Personal Operating Space, POS) in length (Wireless Personal Area Networks, WPAN) [2]. Both organizations are developing the definition of protocols with data throughputs greater than 100 Mb/s. However there are other potential wireless network applications, with relaxed throughput requirements. These applications are industrial control and monitoring, home automation and consumer electronics, security and military sensing; asset and inventory tracking; supply chain management; intelligent agriculture; and health monitoring [3]. These low-data-rate applications involve sensing of one form or another. This is the reason why these networks are usually called wireless sensor networks, or LowRate WPANs (LR-WPANs), because they require shortrange links without a preexisting infrastructure. Some of the characteristics of an LR-WPAN are described by the standard 802.15.4 [4] and are related below:  Over-the-air data rates of 250 kb/s, 40kb/s, and 20 kb/s  Star or peer-to-peer operation  Allocated 16 bit short or 64 bit addresses  Allocation of guaranteed time slots (GTs)  Carrier sense multiple access with collision avoidance (CSMA-CA) channel access  Fully acknowledge protocol for transfer reliability  Low power consumption  Energy detection (ED)  Link quality indication (LQI)  16 channels in the 2450 MHz band, 10 channels in the 915 MHz band, and 1 channel in the 868 MHz band. We have focused our research in networks that support this kind of applications. In particular, we are very interested in technologies for Automatic Meter Reading functions (AMR) and Distribution Automation (DA) for utilities applications (water, gas and electricity) [5][6]. These applications are often distributed throughout many computers. However, customer data polled from sensor networks, queries, remote control orders and network management messages must be processed by a unique computer named Utility Controller (UC). This entity works as a master in a Supervisory Control and Data Acquisition (SCADA) system. UC controls and collects data from many remote units (sensor nodes). There are many studies which show that solutions based on low power radio networks are viable and that they offer the best cost/permformance ratio [7][8][9]. Since the early 90’s, radio and microcontroller technologies have allowed the development of smart sensor networks. Many research groups have focused their investigations in the network optimization. However neither of them have develop a network for optimizing this kind of applications. The environment in which we are developing our study has the following features:  Communication is usually between nodes and the UC.  Nodes are densely deployed within the range area.  Topological changes are possible (a node can be added or eliminated if radio range changes).  Nodes must implement multi-hop mechanisms to let some kind of packets reach the destiny node.  Nodes can be located without pre-planning.  Nodes are prone to failure.  Nodes have no mobility. Ad-hoc networks are best suited to support these features. We propose (in this paper) a simulation tool to study a multi-hop self-organizing wireless sensor network. We have called this simulator OLIMPO. This paper is organized as follows. In section 2, a brief summary of sensor network performances objectives is given. In section 3, our network protocol is presented. In section 4, we describe the OLIMPO design. In section 5, the results of our network simulation over OLIMPO are discussed. Concluding remarks are given in section 6 and future works are listed in section 7. The goal of this paper is to outline the design of our self-organizing multi-hop network protocol, showing results of several experiments made with our network simulator, OLIMPO. 2 Sensor Network Performance Objectives Wireless sensors design must be developed assuming the following devices’ characteristics:  Low power consumption and low supply voltage.  Low cost.  Flexibility.  Short transmission range.  Embedded processor implementing a distributed intelligence process [10].  Microelectromechanical systems (MEMS), actuators and radio technology integrated in a reduced space [11]. App App App App App Net Machine Router Network Manager Llc BROADCAST MacControl BROADCAST Llc MacControl MAC Routing Table Proximity Table ST ACK Inf To Base ACK ExR ACK/NoFra NoACK NoFra Link ExR Fragment Collect Alarm Telecontrol Telemetry Poll a meter LData.Ind LData.Conf LData.Req LinkQueue LData.Conf LError.Ind LData.Ind SAP 0 SAP 1 SAP 2 SAP 3 Figure 1. OLIMPO protocol stack In that sense, there is a trend to develop and commercialize this kind of embedded systems. This is the case of  an international association of semiconductor companies, software providers, system integrators, and original equipment manufacturers (OEMs) such as ZMD  , ATMEL  CompXs  , Microchip  , betronic  and chipcon  [12]. Another example of this trend is neuRFon  sensor, from Motorola  Labs, a relative simple device that has been inspired in sensory receptors of the nervous system, which are capable of detecting changes in the environment due to stimuli [13]. Given the nature of these nodes, our network must assume some limitations in the network performance. Due to the low radio power (less than 100 mW), the low battery life and the noisy RF band (Industrial Scientific Band, ISM) the data throughtput will be quite low. Simulations made by Morotola  with his neuRFon  network show an average data throughput of 40 kbps with a raw bit rate of 250 kps [14][15]. Furthermore, the packets’ multi-hop necessity for reaching the end, increase the messages latency. Experiments made by Morotola  with a neuRFon  network, formed by 64 nodes with ten meters of radio range and a maximal number of four hops in order to reach the end, show a messages latency of 2.6 seconds [14][15]. Another very important issue we have to consider is the duty channel time. This time is closely related with the channel contention. That’s why we must consider the influence of the design of the Medium Access Control (MAC) layer on the collision effect. Slotted and unslotted chan- nel access are provided in several MAC designs. We have measure this item in our simulation. 3 Self-Organizing Sensor Network Protocol and SCADA Application Many network protocols have been proposed recently [16][17][18][19][20]. However, we have designed a network protocol to optimize SCADA applications [21] [22]. We have focused our research on a modified unslotted CSMA-CA channel access mechanism. Each time a node wishes to transmit data frames, it will wait for a random period. If the channel is idle, the device will transmit the data. If the channel is found to be busy, the node waits for another random period before trying to access the channel again. The design has been inspired following the ISO model of architecture for Open Systems Interconnection (OSI) [23]. The protocol stack defined uses two link layers protocol as depicted in figure 1. With this structure every node is capable to process -transmit and receivebroadcast packets while maintaining a data link dedicated to other node. It is possible to select the type of data link -with or without fragmentationand the simultaneous number of them. Acknowledgement service is available. Our multi-hop network protocol creates a spanning tree logical structure over the physical topology. All nodes develop an algorithm to maintain this logical backbone where the root of the tree -we refer it as base stationwill have access to all network nodes. In addition, every node will be able to route its packets to the base station. Nodes always know the path to root through the parent node, but any selected algorithm must allow the root node (UC) to have an approximate image of current tree topology. This way, packets from nodes to the root do not contain information about routing. Conversely, messages from root to nodes generally use an explicit routing scheme, so packets must contain information about the path. Packet header size must be optimized because radio frames should be as short as possible to reduce the transmission error ratio, effective bandwidth, etc. The collecting algorithm is shown in figure 2. An application layer in all the nodes receives the order. This node passes the token to one of its children (figure 2(2)), and waits for a response. When it is received, (figure 2(3)) the node sends data to the root, and it passes the collect order to the next child (figure 2(4)). Only when there are no more children to be polled, will the node answer with its own data (figure 2(6)). Only one token is being passed between parent and children, so no path header information is necessary, and the answer token is simultaneously flowing to the root. To prevent a lost token, a timeout period guarantees the token passes to the next child. 1 2 34 5 6 0 100 120 125 Token sent Token to parent Data to base station Figure 2. Collecting scheme Node structure creation Global initiation Channel access MAC execution Link Manager execution Enabled timers refresh Event? Upper layer execution NODE EXECUTION Yes No Figure 3. Execution model for a node in an OLIMPO simulation Figure 4. 127 nodes network simulation, with OLIMPO 4OLIMPO, simulator design The motivation to develop a new network simulation tool is the need for capability to track specific traffic -SCADA packetsfrom one node to base station through multi-hops. There are other network simulators such as OPNET [24], ns-2 [25], WinneuRFon [17], SensorSimII [26]. However we have designed a new simulator based on the Event Discrete Scheduler (EDS), to support our networks applications over a high number of sensor nodes. OLIMPO has been create with C++ Borland  Builder 6, creating a friendly Graphical User Interface (GUI). OLIMPO is open source, providing an useful tool to simulate new communication or application protocols for other researchers. The current version of OLIMPO provide a graphical experiment configuration. User can create several experiments changing easily the simulation parameters, e.g.: timers, random process, radio range, nodes deployment. . . OLIMPO also provide several simulation modes - MAC step, Logical Link Control (LLC) step, network step and application stepproviding a way to debug protocols. While executing the simulation many reports are generated in real time filling a text file, that is closed and saved when the simulation ends. Figure 3 shows the execution model for a node, and figure 4 shows the GUI created. 5OLIMPO Simulation and Results The simulated network model includes a maximal number of 127 sensor nodes populated over 600 meters x 600 meters area. The deployment of nodes has been chosen in an hexagonal distribution, forming a dense network, where every node has a radio range of 60 meters and consequently there is a maximal number of neighbors of 6. Figure 4 shows the spanning tree generated. Each node establish a 0 5 10 15 20 25 30 35 40 0 1 2 3 4 5 6 7 Number of nodes Average end-to-end delay (sec) Figure 5. Average End to End Delay logical parent-child relationship with its neighbors, looking for the best path to the root. The number above every node represents the MAC address. Node 0 -shown in yellow color in figure 4is assumed to be the root and its radio range is depicted with a circle. It’s placed in the center of the network topology. The simulation was generated with a unslotted CSMA-CA channel access configuration, and packets with and without acknowledgement. Our objectives were to measure the performance issues related in section 2 and to obtain efficiently SCADA application data from all the nodes. We have created several experiments, incrementing the number of networked nodes until 127, although OLIMPO hasn’t restrictions on the maximal number of networked nodes. After the simulation we have obtain some relevant results. The time for spanning tree creation is related in table 1. These times depend on the configuration times (e.g.: Table 1. Time for Spanning Tree Creation Number of nodes Time for Creation (sec) 7 7 14 11 21 23 28 25 37 31 64 68 127 130 time to beacon). Anyway, we consider it to be reasonable good result because this formation time is a very short time compared with the typical SCADA application time. After the spanning tree was created, we executed the SCADA application proposed. Data application were collecting by the UC successfully, obtaining the following network performance results: 0 5 10 15 20 25 30 35 40 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Number of nodes Average collision time (%) Figure 6. Average Collision Time 0 5 10 15 20 25 30 35 40 0 0.5 1 1.5 2 2.5 3 3.5 4 Number of nodes Average duty time (%) Figure 7. Average Duty Time  The average data packet delivery end-to-end delay performance for 20 hours of simulation run time is shown in figure 5. We can see that this time depend directly on the number of hops. Once we increase the number of nodes, the number of hops also increases, and, consequently, the average data packet delivery end-to-end delay increases too.  Figure 6 shows the average collision time for the all nodes. Collision time never is above 0.6 % for all nodes and 1.5 % for base station.  Data throughput is obtained applying the equation 1     "!!#%$'&( )*!#+"!!#%$'&,!-(/.10 2345! 0 !# (1) With a raw bit rate of 4800 bps, we have obtain a throughput of approximately 4655 bps.  The simulation shows an average duty time of 2.43.6% as depicted in figure 7. 6 Conclusion We have presented OLIMPO, a network simulator, as an useful tool to develop SCADA applications over wireless sensor networks. We have developed a network protocol that constructs and maintains the spanning tree. To illustrate the capabilities of OLIMPO, we have generated several experiments, presenting simulation results showing network formation time, data packet delivery response time, collision time, throughput and average duty time. We have obtained relevant results for SCADA applications. 7 Future works For the proposed simulator, we are developing new protocols, such as Low Duty cycle MAC (LD-MAC), slotted CSMA-CA channel access, Mediation Device (MD) protocol and Distributed MD protocol. Due to the good results, obtained with OLIMPO, we have begun to work on the implementation of some SCADA applications over real sensor nodes. References [1] Institute of Electrical and Electronics Engineers. IEEE 802.11-1999(ISO/IEC 8802-11:1999) IEEE Standard for Information Technology - Telecomunications and Information Exchange between Systems - Local and Metropolitan Area Networks - Specific Requirements - Part 11: Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) Specifications. IEEE Press, New York, USA, 1999. [2] Jos´e A. Guti´errez. IEEE 802.15.4: A developing standard for low-power, low-cost wireless personal area networks. In IEEE Network, number 5 in v. 15, pages 12–19, New York, USA, October 2001. IEEE, IEEE Press. [3] D. Estrin et al. Instrumenting the world with wireless sensor networks. In Acoustics, Speech, and Signal Proccesing, v. 4, pages 2033–2036, New York, USA, 2001. IEEE, IEEE Press. [4] Institute of Electrical and Electronics Engineers. IEEE 802.15.4-2003 IEEE Standard for Information Technology - Telecomunications and Information Exchange between Systems - Local and Metropolitan Area Networks - Specific Requirements - Part 15.4: Wireless Medium Access Control (MAC) and Physical Layer (PHY) Specifications for Low-Rate Wireless Personal Area Networks (LR-WPANs). IEEE Press, New York, USA, 2003. [5] A. Bond. The water industry (automatic meter reading). IEE Colloquium on ’Low Power Radio and Metering’. IEE, London, UK, (Digest No.1994/060), 1994. [6] B. Adams and M. Philips. Trends towards standard communications for metering. Ninth International Conference on Metering and Tariffs for Energy Supply. IEE, London, UK, (Conf. Publ No.462), 1999. [7] A.M. Fox. The business case for radio based AMR. IEE Colloquium on ’Low Power Radio and Metering’. IEE, London, UK, (Digest No.1994/060), 1999. [8] S. Mak and D. Radford. Design considerations for implementation of large scale automatic meter reading systems. IEEE Transactions on Power Delivery, 10(Digest No.1994/060), Jan 1995. [9] F. J. Molina, M. G. Gordillo, J. Luque, and J. Barros. Radio network architecture for automatic meter reading. Conf´ erence Internationale des Grandes R´ eseaux ´ Electiques, Krakow POLAND, 1999. [10] R. Estrin, J. Heideman, and S. Kumar. Next century challenges: Scalable coodination in sensor networks. In ACM/IEEE International Conference on Mobile Computing and Networking, v. 5, pages 263– 270, 1999. [11] S. Tilak, N. Abu-Ghazaleh, and W. Heinzelman. A taxonomy of wireless microsensor network models. In ACM Mobile Computing and Communications Review, MC2R, 2002. [12] Patrick Kinney. Zigbee technology: Wireless control that supply works. In Communication Design Conference. IEEE, 2003. [13] L. Hester, Y. Hung, O. Andric, A. Allen, and P. Chen. neuRFon  NetForm. A self-organizing wireless sensor network. 2003. [14] Jian Huang, Yan Huang, Edgar Callaway, Qicai Shi, and Bob 0’Dea. Simulation of a low duty cycle protocol. In OPNETWORK 2001, Washington, D.C., August 2001. [15] Yan Huang, Jian Huang, Lance Hester, Anthony Allen, Oleg Andric, Priscilla Chen, and Bob 0’Dea. Opnet simulation of a multi-hop self-organizing wireless sensor network. In OPNETWORK 2002, Washington, D.C., August 2002. [16] J. Broch, D.A. Maltz, D. B.Johnson, Y.C. Hu, and J. Jetcheva. A performance comparison of multi-hop wireless ad hoc network routing protocols. MOBICOM, Dallas, TX, Aug 1998. [17] Edgard H. Callaway Jr. Wireless Sensor Network. AUERBACH PUBLICATIONS, Boca Raton, USA, 2003. [18] J. J. Garcia-Luna-Aceves and M. Spohn. Source-tree routing in wireless networks. Proc. IEEE ICNP 99, 7th ntl. Conference on Network Protocols, Toronto, Canada, Oct 1999. [19] Couto and B Aguayo. Performance of multihop wireless networks: Shortest path is not enough. Proceedings of the HotNets, Princeton, New Jersey, Oct 2000. [20] R.W. Heinzelman, A. Chandrakasan, and H. Balakrishnan. Energy-eficient routingprotocols for wireless microsensor networks. In Hawaii International Conference on System Sciences (HICSS ’00), Jan 2000. [21] Francisco J. Molina, Julio Barbancho, and Joaqu´ın Luque. Application protocol for SCADA on a wireless metropolitan area network. In C.E. Palau Salvador, editor, Communication Systems and Networks, pages 227–232, Benalm´adena, Spain, September 2003. IASTED, ACTA Press. [22] Francisco J. Molina, Julio Barbancho, and Joaqu´ın Luque. Automated meter reading and SCADA application for wireless sensor networks. In Samuel Pierre, Michel Barbeau, and Evangelos Kranakis, editors, Ad-Hoc, Mobile and Wireless Networks, number 2865 in Lectures Notes in Computer Science, pages 223–234, Montr´eal, Canada, October 2003. ADHOCNOW 2003, Springer-Verlag. [23] H. Zimmermann. OSI reference model - the ISO model of architecture for open systems interconnection. In IEEE Trans.Commun., number 4 in v. COM28. IEEE, 1980. [24] http://www.opnet.com/nsnam/ns/. [25] http://www.isi.edu. [26] C. Ulmer. http://users.ece.gatech.edu/ grimace/research/sensorsimii/. 2000.