IOT Topologies

General Topology Sensors and devices connect to IOT Communications Networks like M2M (3G/4G), LoRa, SigFox and private local Networks such as Z-Wave, Zigbee, Bluetooth, WIFI and Threat. Smart Grid, Smart Homes, Smart water networks, intelligent transportation, early warning systems are based on IOT communication networks to connect to the world. A Wireless Sensor Network (WSN) is a network of sensor nodes to measure light, heat, pressure and other physical phenomena. There are three kinds of Wireless sensor networks : star, tree and chain. Security, identity, trust management, dynamic characteristics and terminals are important qualities to be considered for wireless sensor networks. Wireless sensor networks protect the information and resources from security attacks like Sybil, DoS and abnormal nodes. There are evolving technologies in the areas of node security, cryptography, key management, authentication, authorization, network security, tamper-proofing, encryption, secure routing and secure data aggregation. Public networks like low-power wide area networks (LPWAN) operate in 800-900 MHz bandwidth and transmit 100-140 messages per day. Small sensors on low-power networks operate on regular batteries for 10-15 years. M2M networks can transmit large data over 3G/4G networks. Global international roaming is allowed on these networks.Local Networks consist of local gateways covering short range distances which allow sensors and actuators. They exchange information using local RF technologies like WIFI or Fixed internet lines, ZigBee,Z-Wave, WIFI, Bluetooth Low Energy (BLE) and Threat. These technologies have a very low duty cycle. Zigbee networks have hundreds of nodes in a network.Bluetooth is limited two nodes and Wi-Fi to 15 devices per access point. Topologies used in IOT deployment networking architectures are Point to Point, Star and Mesh Networks. Each IOT deployment network topology has different attributes and capabilities. Typical attributes which are used for selection are range, power consumption, reliability, scalability, bandwidth, data rate, flexibility, interoperability and cost. A broader solution consists of a mix of networking technologies to address device to device communication. Point-to-Point Network Two network nodes are connected directly in a point to point network. Communication happens between the two nodes or devices. Point-to-point networking is simple and cost is low. Scalability is an issue as it cannot go beyond two nodes. The network range is one hop and the transmission range for a single device. Typically one node is connected through internet gateway or another network which gives access to the users to use the device. Star Network The sensor nodes are connected to one central hub which will be the gateway node in the star network. The hub will be exposed to the external world through gateway node. Central hub is a common connection point for all nodes in star network. The performance of the star network is consistent, fast and predictable. It has low latency and high throughput. Data packet travels one hop to reach from hub and sensor or at most two hops between two sensors. Reliability is high in star network as faults and devices can be isolated. Single link is required for each device and isolation of individual devices is easy. The range of a single device is limited to the transmission range. Single point of failure is the gateway node in star network. Exchange and storage of data internally is possible when the gateway fails. Mesh Network Mesh Network has gateway, router and simple sensor nodes. The gateway node is exposed to the external world. Router nodes act as a relay for other nodes and capture and share their data. Collaboration happens in the network with neighbour nodes for sharing the data. In a mesh network, all nodes are within the transmission range of atlas one router node. Routers share the data packets received to gateway node. Mesh network topology is used for long range and broad area coverage. The network can scale up to thousands of nodes. A high density of coverage with a broad arrangement of sensors and devices can be provided in this network. Mesh network has flexibility to face High Radio Frequency challenges and obstacles. Self healing and packet retransmission features mitigate intermittent network interruptions. This type of network has high degree of network resilience. Recommendation Mesh Networks are more complex than point to point or star network topologies. Network latency is higher in mesh networks. This is because of multiple network hops from sensor to gateway. In a smart building application, the energy usage is analyzed and monitored in the building which consists of floors and rooms in each floor. Monitoring and control points are deployed on each floor. Occupancy, pressure, humidity, gas, light and temperature sensors are installed on each room and floor. Sensor data will be aggregated in the central management centre in the building. The applications in central management centre will monitor, analyse and optimise energy usage in the building. Utilisation patterns are identified from the data collected. Higher the resolution of information related to heat and occupancy provides better insight. Zigbee mesh networking architecture is preferred in the building management applications. The other applications include surveillance, industrial automation, logistics, inventory and asset management. There are many standards for different protocols in the networking of IOT sensors. The different protocols are WiFI, WiFi LP, Bluetooth, Bluetooth LE, Zig bee, Z-Wave , EnOcean, 6LoWPAN and others. Some of the challenges related to the sensors and devices are malfunctioning of sensor or device, improperly provisioned SIM card and wrongly set APN within the device. Network latency, bandwidth, jitter malfunctioning and stability are the quality of service issues related to network technologies. Cost management and growing interconnections are important challenges in the IOT network deployment. There are pool plans to manage cost like Fixed, Flexible and Add-on rate pool plans to manage a set of devices and scale the network with additional devices and sensors. Integrity of corporate network, Access controls, user management, audit and reporting of devices status and information are managed by the standard IOT platforms. There are open source IOT frameworks like Leshan & Wakama (LWM2M implementations), Kura & Mihini (M2M gateways), OneM2M, Eclipse Smart Home and Eclipse SCADA. Messaging Protocols Mosquitto is an open source MQTT message broker service. It uses the MQTT protocol for the device to communicate by sending and receiving messages. Among the message brokers that support MQTT, Mosquitto is a small and light weight implementation of MQTT v3.1/3.1.1. Mosquitto runs well on small compute models like the Raspberry PI and Intel Edison. Arduino IDE can be used for building communication services. Mosquitto is based on Eclipse which is a lightweight server implementation of MQTT protocol. Sensors and actuators are the sources and destinations of MQTT messages. Message Queuing Telemetry Transport (MQTT) is a standard protocol. Mosquitto is a bridge which connects to other MQTT based messaging servers. Bridge has features of passing MQTT messages from source to destination. Mosquitto based application consist of components connecting to a messaging server, subscribing to a topic and publishing to the topic. Andy Stanford-Clark of IBM and Arlen Nipper - past CTO of EuroTech are the MQTT Technology founders. MQTT is used for real time tracking of assets and fleets. The event information can be efficiently sent and received with MQTT and OWA 11 A devices. The other area where MQTT is used are home automation related to power monitoring, lighting control and gardening. Disaster & emergency alerting management, Energy Monitoring, Burglar detection system and Location aware messaging systems use MQTT. MQTT services are deployable on CloudMQTT, HiveMQ, AWS IoT and Azure IoT Hub. Publish-Subscribe pattern is used to publish and subscribe messages to topics. The mosquitto broker will distribute messages to clients subscribed on a topic. JSON and XML are typically used for distributing messages. JSON is a preferred choice for IOT applications. IOT devices use standard protocols. Small devices use MQTT and CoAP as the code footprint is small. CoAP is the constrained application protocol from Constrained Resource Environments IETF group. Message Queuing Telemetry Transport (MQTT) is message-oriented and typical architecture is based on a client/server model. Sensor is a client and it connects to the broker over TCP. Message is published to a topic (an address). Clients subscribe to multiple topics. The publish subscribe model is used by MQTT to communicate one-to-one, one-to-many and many to one clients. The typical scenario in different real life industries is remote sensors sending messages to the central system for processing and central system sending control commands to the devices. Sensors can send the condition of the devices to the server. MQTT enables connectivity to internet for data transportation. MQTT brokers have application level Quality of service types which are At most once Delivery, Delivered at Least Once and Delivered Exactly Once.In the case of At most once delivery service level, response is not expected and there are no retry semantics. Delivered at Least once service level ensures delivery of message at-least once and duplicate message can be there. The message is delivered once and only once in the case of Delivered exactly once. MQTT messages can be persisted on the broker. The most recent persistent messages are stored. Client receives the persisted message when one subscribes to the topic. MQTT brokers do not persist message to back up inside the server. The credentials are required for authenticating clients connecting to the server. TCP connection is encrypted with SSL/TLS to ensure privacy. CoAP is the constrained application protocol for transferring the documents. It is designed for constrained device needs. CoAP packets are easy to generate and smaller than HTTP TCP flows. CoAP does not use TCP but uses UDP. UDP broadcasting and multicasting is used for addressing through CoAP. CoAP is based on the client & server model. GET, PUT, POST and DELETE methods are allowed on CoAP services. It interoperates with HTTP and REST web methods. CoAP can be used on packet-based communication protocols like SMS. Confirmable or Non-Confirmable messages are two types of application-level service quality. Fire and Forget are Non confirmable messages. Content Negotiation is supported by CoAP. Accept options are used by clients to express a preferred representation of resources. Reply with a Content Type option for servers. DTLS capable CoAP devices messages can be transported after encryption using RSA, AES or ECC and AES. Servers can send stream changes to clients using CoAP requests and responses. CoAP client can discover resources along with their metadata. MQTT is many communication protocol used by multiple clients through a central broker. CoAP is one to one protocol between client and server for state transfer. MQTT broker creates a long lived outgoing TCP connection to the clients. UDP packets are used by CoAP brokers for communications. MQTT has no feature to label message or any other management. CoAP has support for content negotiation.