By Gal Marder, EVP of Strategy, JFrog
The Internet of Things (IoT) has evolved significantly since its inception: connecting physical devices to the internet for data transfer, communications, and remote controls. Early use cases, such as smart water metering systems, have enabled utilities to remotely keep track of water use, flow, and pressure. This helps in detecting anomalies in the water supply chain, eliminating manual readings, and enhancing real-time data visibility, contributing to water conservation, cost reduction and improved customer satisfaction. The core functions of smart meters revolve around collecting and regularly transmitting data to a cloud-based business application tailored for analytics. The software running on the devices is simple, updates are infrequent, and remote controls are mainly required during initial setup and troubleshooting.
In today’s world, virtually everything operates through software, with the number of IoT devices already surpassing the global population. In recent years, several tech (r)evolutions have paved the way for more sophisticated use cases, enabling “things” to operate more autonomously. Some of the most important advancements include:
- More capable sensors enable “things” to see, listen, and sense their surroundings. Advanced use cases, such as autonomous vehicles, require a combination of sensing technologies to “come alive”. Diversifying the collected data set enables the deployment of more sophisticated AI systems to ensure reliability under any condition.
- More capable hardware and distributed computing. The advancements in multi-core processing have allowed it to run workloads already on the devices themselves enabling real-time operations. This brings several benefits in deployment reliability, flexibility, scalability, and data privacy as processing can be done locally with less data being transmitted and stored in a centralized location.
- Microservices-based SW applications with embedded AI. Modern microservices-based SW applications facilitate flexible deployment options, allowing mission-critical services to be deployed closer to data sources, while non-critical workloads can be centralized. Containerization provides an essential layer of isolation and security while ensuring high availability of services. The SW running on IoT devices has evolved beyond simplicity, now incorporating embedded ML models. To contextualize this, a contemporary car incorporates roughly 100 million lines of code, but reaching full autonomy requires an astonishing 10X increase. This scale & complexity represents a formidable engineering challenge.
The evolution of IoT use cases – from reactive to adaptive.
IoT use cases can be broadly categorized into three groups based on “device intelligence” and the degree of autonomy in operations.
- Simple: Involves the automation of manual data collection. Devices transmit small amounts of data at predefined intervals to the core for processing and analytics. In-device SW is simple with infrequent updates. Remote controls are primarily needed during initial setup and troubleshooting. Use cases encompass environmental monitoring, asset tracking, and smart parking applications.
- Intermediate: Combines multiple sensing technologies and employs distributed computing. Higher volumes of data are being generated and processing can take place on-devices or at the edge. This allows devices to autonomously execute a variety of tasks. For instance, industrial robots equipped with machine vision capabilities can operate with heightened accuracy, orientation, and comprehension. Remote controls allow operators to convey instructions (inputs) regarding the robot’s operations and enable businesses to streamline processes remotely.
- Advanced: Encompass autonomous use cases with extensive use of remote sensing technologies. Devices employ powerful in-device processors to run AI applications for real-time analytics and operation. Devices can autonomously perform a diverse set of tasks and dynamically adapt to their surroundings. Remote controls are evolving towards dynamic conversations, involving both text and voice interactions with the devices.
As an advanced use case example, some of the world’s largest retailers are experimenting with drones in last-mile deliveries, where drones are used for delivering small to medium-sized parcels for short distances. The business advantages are evident, as drones are not impeded by traffic congestion, can optimize delivery routes, reduce human labor and environmental impact, as well as enhance CX through faster and more reliable service. Operators at the control center can monitor the live camera feeds, track the location of each drone, and receive telemetry data. In situations where the drone encounters unexpected obstacles or complex environments, operators can take full control of the devices remotely.
The rapid expansion of the IoT market has left gaps in cybersecurity posture and SW Supply Chain has become a critical attack vector.
IoT devices, like any other computing devices, use multiple SW components to carry out their tasks. These components require regular updates to fix bugs, introduce new features, and importantly, to address security vulnerabilities. Neglecting updates can result in unaddressed vulnerabilities, making devices exploitable by attackers. Consider a situation in which an unmanaged SW vulnerability could potentially allow a malicious actor to take control over the entire connected system, such as a vehicle. Alarmingly, a significant number of organizations lack confidence in their ability to defend against cyberattacks, primarily due to the absence of standardized and cohesive regulations worldwide. Organizations themselves bear the responsibility of implementing security best practices and guidelines to take over the control of their SW supply chain security. This starts with SW curation & vulnerability detection and extends over the full software lifecycle from development to device.
In conclusion, the integration of AI, Edge Computing and IoT has been the game changer for more advanced use cases, which differ significantly compared to the ones we have witnessed in the past. This Intelligent Edge paradigm shifts data processing and analytics closer to the source of data rather than relying solely on the cloud infrastructure. This means more software is running on devices and edge which needs to be proactively maintained and updated to ensure that systems remain secure, reliable, and up to date with the latest features, patches, and improvements. The ability to apply automation in registering, deploying, updating and supporting these intelligent edge devices at a scale is emphasized. Organizations should implement a continuous flow of small, frequent updates from development to devices, all while ensuring the security of the software supply chain. The security notion in IoT needs to be incorporated by design and by default (shift-left), not by response, with the target of making the systems free of exploitable vulnerabilities.
All the aforementioned principles closely align with the core tenets of DevSecOps for IoT. Interested in learning more about accelerated, secure SW development and delivery for IoT? Stop by the JFrog booth at KubeCon to speak with a solutions expert and see a demo.
##
ABOUT THE AUTHOR
Gal Marder brings over 20 years of experience across various tech leadership positions. Prior to joining JFrog, he was the co-founder and CEO of Trainologic – a cutting-edge DevOps technology consulting company that provided comprehensive development, consulting, and training services for leading software technologies. In 2018, JFrog acquired Trainologic with Gal taking on the role of VP of DevOps Acceleration, scaling and leading the team providing solutions engineering and professional services to JFrog’s global top enterprise customers.
