Table 2.

Software used for the modules implementation.

Run Time	Message Broker	RabbitMQ: It is an open source streaming platform that supports different message brokers to provide fault-tolerant, scalable, high-throughput and low-latency data pipelines of queuing real time IoT data streams using a publish-subscribe mechanism.
	IoT Device Connector	Cisco Kinetic: This is a scalable, secure commercial system that can be used to extract, compute and move the data tuples to the right applications at the right time. There are three integral parts of the Cisco Kinetic platform: Edge & Fog Processing Module (EFM), Gateway Management Module (GMM) and Data Control Module (DCM).
	Data Link	Cisco Kinetic Connector: As a feature of EFM, Cisco Kinetic Connector provides a wide array of data links developed by Cisco, Third Party, and Open Source Community. It supports connectivity between compute nodes, message brokers, and IoT devices.
Stream Processing & Analytics	Data Flow Editor	Cisco Kinetic Dataflow Editor: This is a feature in EFM that can be used to customize, modify, and manage data flows with a graphical layout. It also offers a convenient interface to create and debug data flows.
	Parser	JSON parser: JSON objects are mainly exchanged between the computational nodes in our system. Therefore, the parser is used to encode the data structures to JSON strings and decode them back to dictionary, list, tuple, boolean or other numerical data types.
	Stream ML Library	Scikit-Multiflow: It offers main packages to assist the users with handling and learning from their data streams such as stream generators, learning methods, change detectors, and evaluation methods.
	Processing library	Python: For dealing with structured incoming data streams and detecting different data patterns, we have developed the algorithms to take action when the events happen. A variety of built-in Python libraries, such as numpy and scipy, were used to develop our algorithms.
Admin/Control	In-memory Database	RethinkDB: It is an open-source, distributed document-oriented database for real time changing feeds. It allows the developers to push the continuous queries to retrieve the results in real time using ReQL query language.
	Visualization (Historical Data)	Superset: Aiming to extract the insights from the historical/processed data, we have employed Superset, which is a new ongoing incubation at the Apache Software Foundation.
	Visualization (real time Data)	Grafana: It is an open-source platform capable of monitoring and analyzing the dynamic data incoming from IoT devices, which we used for our streaming real-time data.
Provision & Orchestration		Aiming to mitigate difficulties in managing, distributing and updating the system, we have installed Apache Ambari and Apache Zookeeper in our network of compute nodes. The Apache Ambari package is then used to configure and install the other main modules of our IoT architecture.
Security & Governance		For the security, we have also configured Wazuh which is an open source system for integrity monitoring, and threat and intrusion detection to protect our compute nodes. It consists of many functions such as security analytics, vulnerability detection, file integrity monitoring, and configuration assessment.