Enterprise-grade IoT Platforms: Benefits and Components
As more firms begin to turn IoT strategies into realities, IoT platforms are becoming central piece for enterprises to address the immense complexity of deploying and managing an IoT solution. Imagine the in-house effort required in deploying and managing thousands of devices in the field, receiving and analyzing the huge swaths of data generated by these devices, developing applications to process and deliver insights on this data, and then of course do this all securely. An enterprise-grade end-to-end IoT platform can address these complexities by simplifying the process of deploying, managing, analyzing insights and securing an end-to-end IoT solution. By purchasing and implementing a turn-key platform with these seamlessly integrated components — instead of developing and/or integrating a solution in-house — an organization can realize a superior ROI through lower TCO and faster time-to-market.
While enterprises use IoT platforms to optimize internal business processes and gain new insights, OEM solution manufactures that produce critical components of IoT solutions are also starting to see the benefits of partnering with IoT platform providers and their respective ecosystems, in order to deliver a proper end-to-end experience for its customers. For example OnFarm, a smartfarm provider that provides farming solutions to measure crop moisture levels, pestitside levels, and weather impacts, has partnered with PTC’s Thingworx platform to provide end-users an easy way to deploy and manage their OnFarm solution for users.
IoT Platform Components
Given the constant evolution of IoT Platforms and the amount of new entrants — there are over 400 platform vendors as per a November 2016 report from IoT Analytics — narrowing down the list of true enterprise-grade IoT platforms can be difficult. Customers need to consider the following core components as essential parts of a true end-to-end enterprise-grade IoT platform:
- Connectivity: create and manage the link between devices, the Internet, and the customer data centre / cloud
- Device management: provisioning, operations, and management of a diverse array and large number of devices
- Streaming Analytics Engine: receive and transform data from devices into timely, relevant, and actionable insights
- Application Development and Integration: create and configure applications, and integrate with back-end systems
While I have omitted security from the list (I believe it to be a requirement rather than a component), it is nonetheless another very important part of any IoT solution because of the sensitive nature of information being processed as well as the large number of entry-points for hackers to potentially gain unauthorized access into the system. Not coincidentally, IoT security has lately been a hot topic due in part to a number of recent security breaches in live IoT environments.
Let’s dive into each of these components in a little more detail:
1) Connectivity: There are now a variety of ways that devices can connect to an enterprise IoT network. In the old days of “Machine to Machine”, expensive cellular data-only connections or wired connections were the primary method of communication between devices and the central server. However, with the rise of alternate low-power / low-consumption connectivity methods such ZigBee, Sigfox, and LoRa (all much more reliable and efficient than Wi-Fi communication), and the recent introduction of “LTE-IoT” networks by AT&T and others, an IoT Platform needs to handle a diverse array of current and future connectivity methods to be an effective solution for enterprises.
As well, there are a multitude of IoT application and messaging protocols that can be used to communicate data streams from end devices to the cloud gateway. A good platform should be able to support multiple protocols, such as MQTT, CoAP, and AMQP. While MQTT is seen as the most IoT-oriented messaging protocol due to its “light-weight”, a platform nonetheless has to be able to handle a multitude of protocols.
Cisco’s Jasper IoT platform was initially built to manage devices with SIM-enabled cellular connectivity, (which has evolved from 2G to now LTE-IoT). The Jasper interface offers a powerful user interface and analytics dashboard to manage and locate these devices and their data usage, and has developed a very strong telecom partner ecosystem along the way.
2) Device Management: IoT devices in the field can number in the thousands in large-scale implementations. For example, Awesense uses a platform to manage power grid electricity flow sensors along thousands of kilometers of power lines to pinpoint deficiencies in the power grid to minimize expensive technician calls and downtime. Of course, deploying and managing such a wide array of devices can be extremely challenging. A proper IoT platform can help mitigate this complex task be remotely monitoring and maintaining these sensors through an intelligent central user interface. The ability for a platform to perform remote diagnostics and patch management further adds to the efficiency of a deployment
End-point devices can also send data back to the user cloud in a layered approach, using gateways and edge computing devices which process and action on data outside of the central cloud gateway to minimize data flows and storage into the back-end database. As such, having such a wide array of devices requires an IoT platform to have powerful management capabilities to simplify the process of provisioning, configuring, and managing devices, including functions such as remote monitoring, testing, patch management and trouble shooting.
3) Streaming Analytics Engine: Collecting data is in important part of the IoT solution process, but analyzing and gaining actionable insights from it is where the real value lies. A good IoT platform needs to offer a real-time analytics streaming engine with the ability to add business rules, alerts, as well as a UX interface for the right user to receive these insights and alerts as required. For example, a properly developed IoT solution on a refrigerated food truck can monitor and stream temperature data in real time within the food compartment, and then notify the right contact if a given cooling unit malfunctions and the compartment temperature exceeds a government mandated safety threshold.
Global elevator manufacturer thyssenkrupp recently implemented elevator sensors and Microsoft’s Azure IoT Suite and Stream Analytics platform for predictive maintenance purposes, capturing and analyzing elevator data including motor temperature, shaft alignment, cab speed and door functioning. Using Azure advanced analytics, thyssenkrupp can predict with improved accuracy when an elevator will break down before it happens, dramatically reducing the costs of maintaining elevators and greatly increasing customer satisfaction rates.
4) Application Development and Integration: Many enterprise-grade IoT platforms come with their own pre-built IoT modules. For example, PTC’s Thingworx platform includes turn-key Asset Management and Workflow Management modules to simplify implmentation and management. However, organizations may also need to develop and/or customize their own applications to address specific internal business requirements, processes, and legacy systems. As such, a good IoT Platform will also allow developers to easily create code, business rules, alerts, and data management policies that can process the data that is being consumed and streamed in from various devices in the field. There are various ways to make this possible, through strong SDKs and development tools, scripting tools, and open-source APIs that integrate well into enterprise backend processes and systems such as SAP, Oracle, etc. Using open-source or industry-standard APIs and associated SDKs are also key to be an effective part of a IoT partner ecosystem, so that others can easily link and integrate their IoT specific modules and capabilities into existing platforms.
IBM’s Watson IoT platform boasts its built-in “cognitive API” library. This wide array of APIs contains built-in capabilities to receive unstructured data arriving from IoT sensors, analyze this large data set within the Watson cognitive engine, and then produce insights and analytics back out to the same API for users to review and process.
Of course, security concerns need to be considered as a critical requirement of any IoT platform, as the data generated by IoT solutions is potentially sensitive and can include information about operations and its customers. As such, is it critical for platform to address security at all levels of an IoT solution. An end-to-end IoT solution has a variety of potential breach points, including the IoT device, communication between devices, gateways/edge devices, and the cloud server applications and data storage itself (whether stored in a native datacentre or a public/private cloud). As such, IoT platform services should include security components covering network connectivity, authentication, identity management, and data loss management, as well as data intrusion prevention systems. This means having enterprise-grade security at the device-level (including any edge or gateway devices), device-initiated data communication, and cloud applications with encrypted identity-based authentication.
Worth the Cost?
With so more core components in a platform addressing a wide array of complexities in an IoT solution, an enterprise would be wise to invest in a turn-key IoT Platform to reduce time and costs in development, implementation, on-going maintenance, and additional hardware associated with an IoT solution. While a platform may introduce higher up-front costs in professional services for implementation, the long-term ROI benefits in lower TCO and faster-time-to-market will certainly outweigh these costs in the long run.
