1 IoT Fundamentals: Definitions, Architectures, Challenges, and Promises
13
In summary, the main challenges of IoT can be listed as below:
• Scale: Connecting to billions of active connected IoT devices is a big challenge,
and the current communication models and technologies should be adjusted
to address scalability challenges. In this context, emerging IoT technologies
such as decentralized IoT network (e.g., edge/fog computing), peer-to-peer
communications, and blockchain can be helpful.
• Heterogeneity: IoT in its nature consists of a plethora of devices with different
interfaces and communication protocols, and thus there is a necessity to form a
common way to abstract the underlying heterogeneity.
• Privacy: All the collected data must be kept secure and anonymous when
necessary.
• Data ownership: Who is the owner of machine-generated data (MGD)? The
entity that owns the IoT device or the manufacturer of the device (e.g., in
connected cars)?
• Cybersecurity: Defeating attackers who seek to control, steal, or mislead is vital.
• Legal liability: Who is responsible when something goes wrong with an algorithm or an automated decision?
• Sensors: Technically, sensors must be inexpensive, accurate, and energy efficient.
• Networks: Transferring data and commands must be secure, reliable (correct
and timely), and robust, despite operating in a noisy, busy, dangerous, or harsh
environment.
• Big data: Connected devices continuously and simultaneously generate large
volume and different varieties/forms of data, and thus IoT should be able to
address time, resources, and processing capabilities.
• Analysis: The data must be properly interpreted and analyzed with fidelity to its
meaning, especially if automated actions are taken based on data outcomes.
• Interoperability: There is a fierce competition to lead this burgeoning field, and
all players must work together to be functional and to protect investments and
must do so with fairness and integrity.
1.1.5 IoT and Big Data
Data coming from the Internet of Things is unlike data from the past in at least two
important dimensions. First, the large amounts of data being generated demand a
new data management approach. Traditional methods need to be adapted or entirely
new approaches need to be discovered to handle diverse data constantly streaming
from many sources. The second dimension is the nonuniformity of the data. Often
the raw data is unstructured, or may come in several different formats, or may
even change depending on the context. The new data management techniques must
cope with these challenges. Up until now the discussion has been about big data
without formally defining it. Big data is a large set of structured, unstructured, and
13
In summary, the main challenges of IoT can be listed as below:
• Scale: Connecting to billions of active connected IoT devices is a big challenge,
and the current communication models and technologies should be adjusted
to address scalability challenges. In this context, emerging IoT technologies
such as decentralized IoT network (e.g., edge/fog computing), peer-to-peer
communications, and blockchain can be helpful.
• Heterogeneity: IoT in its nature consists of a plethora of devices with different
interfaces and communication protocols, and thus there is a necessity to form a
common way to abstract the underlying heterogeneity.
• Privacy: All the collected data must be kept secure and anonymous when
necessary.
• Data ownership: Who is the owner of machine-generated data (MGD)? The
entity that owns the IoT device or the manufacturer of the device (e.g., in
connected cars)?
• Cybersecurity: Defeating attackers who seek to control, steal, or mislead is vital.
• Legal liability: Who is responsible when something goes wrong with an algorithm or an automated decision?
• Sensors: Technically, sensors must be inexpensive, accurate, and energy efficient.
• Networks: Transferring data and commands must be secure, reliable (correct
and timely), and robust, despite operating in a noisy, busy, dangerous, or harsh
environment.
• Big data: Connected devices continuously and simultaneously generate large
volume and different varieties/forms of data, and thus IoT should be able to
address time, resources, and processing capabilities.
• Analysis: The data must be properly interpreted and analyzed with fidelity to its
meaning, especially if automated actions are taken based on data outcomes.
• Interoperability: There is a fierce competition to lead this burgeoning field, and
all players must work together to be functional and to protect investments and
must do so with fairness and integrity.
1.1.5 IoT and Big Data
Data coming from the Internet of Things is unlike data from the past in at least two
important dimensions. First, the large amounts of data being generated demand a
new data management approach. Traditional methods need to be adapted or entirely
new approaches need to be discovered to handle diverse data constantly streaming
from many sources. The second dimension is the nonuniformity of the data. Often
the raw data is unstructured, or may come in several different formats, or may
even change depending on the context. The new data management techniques must
cope with these challenges. Up until now the discussion has been about big data
without formally defining it. Big data is a large set of structured, unstructured, and