Worldwide, the IoV (Internet of Vehicles) market was valued at $66,075 million in 2017. By 2024, the forecast is $208,107 million (Allied Market Research). IoV uses a wireless network to help vehicles connect with each other or other traffic infrastructure and is one of the newest achievements in developing smart cities. As such, it uses smart technology and data to solve sustainability challenges.
The Internet of Vehicles (IoV) is a subset of the Internet of Things (IoT). There’s a growing need for intelligent and connected devices that help us in daily life. IoT devices – smart mobiles, refrigerators, watches, fire alarms, medical sensors, fitness trackers, sensors of autonomous cars, or any other smart objects – satisfy that desire, and the market is booming.
Nowadays IoT sensors, actuators, gadgets, appliances and machines are used in almost every industry, as well as day to day life. One area the Internet of Things is making a notable impact is the automotive industry. In recent years, there’s been huge interest in the Internet of Vehicles (IoV), one of the major branches of IoT.
What is meant by the Internet of Vehicles (IoV)?
To answer the question ‘What is the Internet of Vehicles?’, we need to define several terms. Let’s start with vehicular ad hoc networks (VANETs) – sometimes referred to as Intelligent Transportation Networks.
VANETs are a subgroup of MANETs (mobile ad hoc network), applied to the automotive industry, facilitating vehicle-to-vehicle communication (V2V), and communication with roadside equipment. Over time, vehicular ad hoc networks have evolved into a broader term: Internet of Vehicles.
The Internet of Vehicles allows safe and efficient exchange of vehicle-to-vehicle information and also between vehicles and infrastructures – using VANETs.
IoV solutions combine intra-vehicle networks, inter-vehicle networks, and vehicular mobile internet. Furthermore, IoV technology uses all the sensors that are connected to the ECU (Electronic Control Unit) by CAN Bus (Controller Area Network), allowing microcontrollers to communicate with each other without a host. With that, the vehicular cloud is created – vehicles connected to the cloud backend server using a mobile network.
How is a car connected to the internet?
Connected cars hook up to the internet via a wireless local area network (WLAN). Connected cars are part of the Internet of Things and the Internet of Vehicles, allowing remote control of some aspects and communication with other products. For example, connected cars can alert drivers to traffic, plan routes, and adapt cruise control for more efficient traffic management.
There are five main ways vehicles connect with surroundings and communicate:
- V2I: Vehicle to Infrastructure
- V2V: Vehicle to Vehicle
- V2C: Vehicle to Cloud
- V2P: Vehicle to Pedestrian (or Vehicle to Human)
- V2X: Vehicle to Everything
IoV has benefited from the improvement of big data technologies and mobile networking. IoV solutions differ from vehicle networking technologies like V2V communication and vehicular networks.
With a cloud-based approach, connected vehicles work within a cloud data center, vehicle statuses are uploaded through wireless communications, and vehicle data is usually treated as a data stream.
There are three major challenges to designing an IoV ecosystem that’s capable of supporting these services:
- High reliability
- High scalability
- Desirable trade-off between accuracy and real-time processing
Once the IoV service is up and running, it needs to run continuously – unexpected crashes or downtime may have critical consequences.
What are the layers that comprise IoV? There are multiple options for implementing and designing IoV systems, from online transaction processing (OLTP) and online analytical processing (OLAP) to quickly processing big data to extract real-time insights. Let’s look into each one by one.
A type of processing consisting of executing multiple transactions in parallel/concurrently. OLTP uses data uploaded by vehicles and stores it in a database. It also provides additional services that analyze the data and uses it in different ways. Despite this kind of system being mature and highly reliable, it doesn’t scale very well horizontally.
This architecture is often used to provide large-scale data analytics services. Using an OLAP solution for IoV scenarios requires integrating many different platforms and processing systems such as Kafka, Storm, or Spark Streaming.
Big Data streaming
Big data streaming is more common in new applications. It’s a process whereby large streams of real-time data are processed using a processing engine, ensuring accurate processing with low latency.
Compared to traditional OLAP/OLTP solutions, big data analytics platforms provide scalability – crucial for a domain, where hundreds of thousands of data events can be transferred in almost no time.
Common solutions like OLAP/OLTP are useful for small-scale fleets where relational databases aren’t bottlenecks for the entire system.
Examples of IoV applications
Now we’ve delved into IoV architecture, what about specific use cases? Read on for a selection of ridesharing companies - platforms matching passengers and drivers with vehicles for hire via mobile or website applications.
Based on real-world ridesharing company UCAR in China, CarStream was deployed and maintained for around three years in industrial usage, with over 30,000 vehicles. It provides similar services to Lyft and Uber but also has unique characteristics.
With CarStream, drivers come directly from the company, allowing the installation of an Onboard Diagnostic connector (OBD) to collect different types of data. OBD data includes speed, engine RPM, and vehicle error code. With near real-time processing of vehicle data, the system is capable of monitoring driver behavior and reacting to specific outliers, ensuring safety and trip quality.
The system also has both offline and online engines in the processing layer. For batch processing they use Hadoop, and for real-time processing, Jstorm. That combination of offline and online is widely used in CarStream to achieve top performance for real-time processing.
The data bus layer in the platform is implemented with Apache Kafka, because it’s easy to integrate with most stream processing engines. JStorm is written in Java, providing support for this language.
The most popular application among ridesharing companies, Uber offers paid passenger rides and boasts one of the biggest Apache Kafka deployments in the world.
According to their website, Uber processes trillions of messages (petabytes of data) each day using a pub/sub message system to send events data to both drivers and riders.
At Uber, business is driven by stability and efficiency of uninterrupted data flow across the platform. Apache Kafka plays a key role, providing a way to prevent the platform from crashing and offering the opportunity to create a disaster recovery plan.
Similar to Uber, Lyft provides car passenger ridesharing services. To process data, they use Apache Beam with the Apache Flink runner streaming processing framework. It is used to run feature engineering and orchestration jobs.
Lyft has built scalable pipelines that process real-time events with extremely low latency. The company generates around 100 features at about three million geohashes per minute, equating to approximately 400 billion features a day.
Why Apache Flink? Several reasons, with some of the most important being:
- Providing low latency and stateful computations
- Support of event time processing
- Exactly-once processing – no matter how many times data processing is re-tried, it will always run once
- Failure recovery
Internet of Vehicles as an emerging sector
How will the Internet of Things affect vehicles? The trailblazing relationship embeds technology in vehicles, connecting them to the Internet, IoT devices, and other vehicles. IoV technology and architecture includes sensors, OLTP, OLAP, the cloud, and big data streaming, with use cases including ridesharing apps Uber and Lyft.
The global IoV market is thriving. Being part of the Internet of Vehicles allows cars, buses, etc., to communicate in real-time with roadside infrastructure, drivers, and pedestrians – the applications are exciting and still evolving.
Keen to learn more about this rapidly emerging market? Get in touch with our expert team of developers.
More posts by this author