Robust protocols for the Industrial Internet of Things Elvis Vogli Politecnico di Bari,Telematics Lab - Dipartimento di Ingegneria Elettrica e dell Informazione Via Edoardo Orabona 4, 70125 Bari, Italy elvis.vogli@poliba.it http://telematics.poliba.it Abstract. The Internet of Things (IoT) paradigm envisions a world with an enormous amount of connected devices. Among the IoT technologies a key role is played from the short range wireless communications. One of the most used standards in this context is IEEE 802.15.4. On the top of it the IETF has already defined an IPv6 enabled protocol stack. Furthermore the Time Slotted Channel Hopping (TSCH) Medium Access Control (MAC) was introduced in the recent IEEE 802.15.4e amendment to further improve energy efficiency and reliability. Despite the big standardization efforts there are still many technical issues to be faced. The purpose of my research project falls in the short range wireless communications context applied to industrial IoT scenarios and the optimization of related protocol stacks. In the following there are described the main findings related to media access, network formation, security and data interoperability. Keywords: IEEE 802.15.4, Time Slotted Channel Hopping, Industrial Internet of Things, Machine to Machine 1 Introduction The IoT represents, in the context of networking and services, one of the most relevant innovations of the third millennium [1]. At the same time, the IoT entails an unprecedented gamut of new technical challenges, only partially addressed so far [2]. In industrial IoT, the power efficient IEEE 802.15.4 standard plays a key role. On the top of it, a full Internet protocol stack is standardized by Internet Engineering Task Force (IETF) [3]. In particular, the IETF IPv6 over Low power Wireless Personal Area Networks (6LoWPAN) adaptation layer implements Internet Protocol version 6 (IPv6) and enables universal Internet connectivity, the IETF Routing Over Low power and Lossy networks (ROLL) defines a routing protocol, and finally the IETF Constrained Application Protocol (CoAP) defines a lightweight application protocol and enables Representational State Transfer (REST) architecting in constrained devices. In addition in 2012 it was published the latest amendment IEEE 802.15.4e which adds more reliability and energy efficiency to the existing MAC. One of the most important features introduced in this amendment is the TSCH mode. The sloted medium access enables an efficient use of the power, whereas the channel hopping improves the communication reliability making it more suitable for industrial applications. In TSCH it is defined a sloted structure like the one shown in
2 Elvis Vogli SCORE@POLIBA Workshop: Authors Instructions Fig. 1. The time is divided in slotframes, each slotframe in slots and there are 16 available channels. Each mote synchronized to the network should receive in the Enhanced Beacon (EB) advertised from the network coordinator, the slotframe size and the Absolute Slot Number (ASN), where ASN is the total number of timeslots that has elapsed since the start of the network. The motes also in order to communicate should be scheduled in the same cell which is identified from the slot-offset (S o f f - the relative position of a slot within a slotframe) and the channel-offset a(c h o f f - frequency hopping parameter). The physical frequency is then determined according to the eq. 1. In the standard it is only defined how is executed a schedule but the effective scheduling is out of scope of the standard and a lot of work needs to be done in order to take advantage of this features. f = F {(ASN + c h o f f ) mod n c h } (1) To overcome this issues the IETF working group IPv6 over the TSCH mode of IEEE 802.15.4e (6TiSCH) has been created to integrate the latest MAC to the protocol stack. In my research activity i have exploited solutions and optimizations related to TSCH and other problems of this protocol stack which can as far as now be categorized in the following areas: 1) energy efficiency and reliability 2) security 3) data interoperability. In the following sections there are briefly described the main outcomes of my research activity. In sec. 2 four fast synchronization algorithms are implemented, in sec. 3 it is evaluated a traffic aware scheduling for TSCH, sec. 4 introduces a security framework for 802.15.4 based networks and in sec. 5 there are considered data interoperability issues in a M2M scenarios. Finally sec. 6 draws some conclusion and future researches. Moreover all the solutions have been evaluated experimentally in real testbeds using the OpenWSN framework [4]. Fig. 1: The schedule structure of the TSCH mode MAC. 2 Fast Synchronization scheduling The EBs used to advertise a TSCH network can be sent in the first slot of each slotframe. Once a mote receives an EB, it synchronizes to the sloted structure and starts sending on its own EBs in order to expand the network coverage. The frequency in which these frames are sent will change each time according to eq. 1.
Robust protocols for the Industrial Internet of Things 3 On the other side a new node that wants to join the network will be listening at only one of the n c h available frequencies. The frequency mismatch between the transmitter and the receiver, and collisions between EBs sent from neighboring motes can make the joining time of new motes to be very long. During the network formation, especially when the number of motes is consistent, a long joining time can be a serious obstacle for the deployment. A joining node before synchronization works with 100% duty cycle therefore the joining time can become also a relevant factor in energy consumption. In order to speed up the joining operations a solution is presented in [5]. Four different scheduling algorithms which regulate the EB transmission speed up the joining operations especially in high density networks. The algorithms purpose is to take advantage from each mote taking part in the network by appropriately scheduling beacon transmission of different nodes. It is defined a multi-slotframe structure in which every new mote will schedule its EB transmission. For each scheduled mote the corresponding cell will be occupied according to the used algorithm function. According to the way the scheduled slots are added (i.e, fill the corresponding position) to the general slotframe structure these algorithms have been called Random Vertical filling (RV), Random Horizontal filling (RH), Enhanced Coordinated Vertical filling (ECV) and Enhanced Coordinated Horizontal filling (ECH) (i.e, the red and yellow colored slots in the figure fig. 1). In the random algorithms the same S o f f (or c h o f f ) is used from every mote but they choose randomly the c h o f f (or S o f f ). In the coordinated versions it is used a collision avoidance mechanism and in addition if a mote has no energy limitation it can be scheduled to transmit in each slotframe. The algorithms have similar performance improvements but they offer different architectural choices. In addition the low computational complexity can make these algorithms very useful in the network setup phase. 3 Decentralized Traffic Aware Scheduling Algorithm (DeTAS) Once the network is set up new slots need to be scheduled to enable application data. From the quality of the schedule will depend many performance parameters of the network (i.e., data delay the energy efficiency, and the reliability of the transmissions), therefore it should be carefully designed. The Decentralized Traffic Aware Scheduling algorithm (DeTAS) proposed and experimentally evaluated in [6] aims to resolve the schedule in multihop sensor network where sensor data are sent toward a central server. In this algorithm transmission and reception slots are alternated and scheduled in consecutive slots. A node, in order to build the schedule will send to its preferred parent the information about its requirements (i.e. number of slots needed) and will receive from it the starting slotoffset. By alternating the transmission and reception slots there are avoided possible queue congestions, because after a message is received, there will always be a transmission slot that follows where it can be forwarded. The channel-offsets are calculated independently in each mote. Given K, the distance of a mote from the network coordinator in number of hops, the transmission channel-offset is equal to (K 2)mod(W ), where W is the number of channels used from DeTAS and it determines the distance in
4 Elvis Vogli SCORE@POLIBA Workshop: Authors Instructions number of hops after which the same channel will be reused. W should be at least 3 in order to avoid interferences. Whereas the reception channel-offset is equal to (K 1)mod(W ). This algorithm was implemented and evaluated using a testbed with TelosB sensor motes in different scenarios. The experimental results confirm a good performance in terms of duty cycle, end-to-end delay, end-to-end and link Packet Loss Ratio. 4 Security framework for 802.15.4 networks The constrains in term of computational and storage capacity that characterize the devices in IoT pose serious limitations to security. The usual security mechanisms used in the normal Internet can not be used without being first adapted. The IEEE 802.15.4 standard supports the possibility to protect MAC packets by using symmetric-key cryptography techniques and it offers several security options and most of the devices commercially available implement hardware encryption primitives. However the standard relies on upper layers to orchestrate the usage of the multitude of security profiles and configuration settings, as well as to handle the creation and the exchange of encryption keys. In support of this functionality, in [7] was defined a standard compliant security framework. A Key Management Protocol (KMP) is proposed in which each two motes negotiate a link key which is used for symmetric cryptography starting from a pre-shared Master Key. In addition to the KMP protocol it is defined a lightweight mechanism that initializes a secure IEEE 802.15.4 domain. Experimental evaluation has shown that the enabling security features in constrained nodes requires relevant additional computational resources. This involves a not negligible growth of communication latencies. Such findings have to be carefully considered when designing and developing enhanced applications in future and secured Internet of Things (IoT) system. 5 Data interoperability of 6LoWPAN based networks in ETSI M2M. The protocol stack described in [3] provides efficiency and reliability as long as Internet connectivity. The devices implementing this stack can connect in the Internet and should be easily integrated in a more complex Machine to Machine (M2M) scenario. This however is not straight forward. If it is considered the European Telecommunications Standards Institute (ETSI) M2M communication standard, a problem rises from the fact that the resources are organized in a tree-like hierarchy [8]. ETSI M2M defines a middleware that follows RESTful architecture. Therefore the resources can be addressed using Uniform Resource Identifiers (URIs). As a consequence of the tree-like structure of the resources, the resulting URIs can become very long. In the case of a 6LoWPAN network long messages can be sent in multiple blocks (i.e., using blockwise transfer mode in CoAP). However URIs should be present in each message block and a long URI sometimes will not be even possible to fit in one packet because of the maximum size of a packet limited to 127
Robust protocols for the Industrial Internet of Things 5 bytes. Considering the various protocol headers even less bytes remain available to the application layer. A first solution to this problem is by mapping the ETSI M2M standard URIs to single byte URIs. This is possible because for each level of the tree there is a limited number of URIs. A second solution can be obtained using an intermediate proxy that relays in the border router. When a mote joins the M2M system it has to create and describe its resources by sending several long messages. These messages can be sent using CoAP blockwise operation but the subscription process would be very long due to the elevated number of blocks. This makes the first solution useful only for scenarios where the motes are stable in the network and the subscriptions are not frequent. Otherwise in scenarios with higher mobility the second solution can be used. 6 Conclusions The boom of IoT has created new opportunities and on the same time new challenges. An accurate design and optimization of the related protocols can make them more robust and reliable. Several issues of a Internet enabled protocol stack adopted in industrial domain have been analyzed so far. The solutions proposed fill some relevant gaps of the stack making it more reliable, energy efficient, secure and resolve some interoperability issues int M2M environments. However still more issues have to be considered. More robust and fault tolerant scheduling algorithms and more application scenarios need to be still evaluated. References 1. Atzori, L., Iera, A., Morabito, G.: The internet of things: A survey, Computer Networks, vol. 54, no. 15, pp. 2787 2805, Elsevier, New York(2010) 2. Whitmore, A., Agarwal, A., Xu, L. D.: The Internet of things-a survey of topics and trends, Information Systems Frontiers, pp. 1-14, (2014) 3. Palattella, M., Accettura, N., Vilajosana, X., Watteyne, T., Grieco, L., Boggia, G., Dohler, M.: Standardized Protocol Stack for the Internet of (Important) Things, IEEE Communications Surveys & Tutorials, pp.1389 1406, (2013) 4. Watteyne, T., Vilajosana, X., Kerkez, B., Chraim, F., Weekly, K., Wang, Q. Glaser, S. D., Pister, K. S. J.: OpenWSN: a Standards-Based Low-Power Wireless Development Environment, Transactions on ETT, pp. 480-493, Aug. (2012) 5. Vogli, E., Ribezzo, G., Grieco, L.A., Boggia, G.: Fast Join and Synchronization Schema in the IEEE 802.15.4e MAC, Submited to: IEEE WCNC 2015 - Workshop - E2IoT 6. Accettura, N., Vogli, E., Pallatella M. R., Grieco, L. A., Boggia, G., Dohler, M.: Decentralized Traffic Aware Scheduling in 6TiSCH Networks: Design and Experimental Evaluation, submited to: IEEE Internet of Things Journal 7. Sciancalepore, S., Piro, G., Vogli, E., Boggia, G., Grieco, L. A.: On securing IEEE 802.15.4 networks through a standard compliant framework, EuroMed Telco Conf. 2014 8. Ben Alaya M., Banouar, Y., Monteil T., Chassot C., Drira K.: OM2M: Extensible ETSI-compliant {M2M} Service Platform with Self-configuration Capability, Procedia Computer Science, pp.1079-1086,(2014)