Module 3 Introduction To IoT
Module 3 Introduction To IoT
• Unstructured data on the internet, like text, videos, and sensor data,
lacks(no) predefined formats, posing challenges for querying
algorithms. NoSQL querying languages are often used due to their
flexibility in handling the diverse and dynamic nature of this data.
• Time-critical data allows for remote processing in locations like clouds. Tolerates latency of a
few seconds.
• Normal data, associated with less time-sensitive domains, can be processed leisurely without
urgent time constraints. Tolerates latency of a few minutes to a few hours milliseconds.
• This classification aids in determining when and where to process data based on its criticality.
Processing Topologies
• Selecting the right processing topology is crucial for efficient IoT architecture,
impacting network bandwidth, energy consumption, and processing latencies.
• Two primary topologies are On-site, involving local processing, and Off-site,
which further divides into Remote processing (external servers) and
Collaborative processing (shared processing among devices). The intelligent
choice between these topologies optimizes resource usage
On-site processing
• On-site processing, performed at the data source, is crucial for applications
with low latency tolerance, such as healthcare and flight control systems.
• This topology ensures rapid and robust handling of data, preventing potential
catastrophic (if the processing infrastructure is slow then generated data will
be missed before processing) consequences in real-time systems like those
with rapid data generation rates.
Off-site processing
• Off-site processing, unlike on-site paradigms, allows for cost-effective handling of
data with permissible latencies, suitable for large-scale IoT deployments.
• The off-site approach facilitates data transmission to either remote locations, such
as servers or the cloud, or to collaborative processing nodes, enabling flexible and
cost-efficient solutions for diverse IoT scenarios i.e., (shared processing among
devices).
Remote Processing
• In off-site processing topology, sensor node sense the data, forwarding it to a
remote server or cloud for centralized processing.
• This setup offers cost and energy savings, allowing reuse of powerful computing
resources, facilitating scalable solutions. However, it heavily relies on network
connectivity and consumes significant network bandwidth.
Collaborative processing (shared processing
among devices)
• By aggregating the processing power of nearby nodes, it enables local collaborative data
processing.
• The collaborative processing topology, ideal for areas with limited network connectivity,
economically addresses large-scale deployments in remote regions.
• By leveraging (increasing min
intelligence) nearby processing nodes, it
reduces latencies, conserves bandwidth,
and is particularly beneficial for
applications like agriculture with
infrequent data processing needs.
IoT Device Design and Selection
Considerations
• The selection of a processor for an IoT sensor node is a critical aspect of
defining an IoT solution, with several key considerations:
• Size:
• Energy:
• Cost:
• Memory:
• Processing Power:
• I/O Rating:
• Add-ons:
IoT Device Design and Selection
Considerations (cont..)
Size:
• Crucial for form factor and energy consumption, smaller form factors are preferred for many IoT
applications, especially wearables, as larger form factors tend to consume more energy.
Energy:
• The energy requirements of a processor significantly impact the sustainability of IoT-based sensing
solutions. Higher energy needs lead to more frequent battery replacements, affecting long-term viability.
Cost:
• Processor cost, along with sensor costs, influences the deployment density of sensor nodes. Cheaper
hardware enables higher deployment density, making solutions more accessible and cost-effective.
IoT Device Design and Selection
Considerations (cont..)
Memory:
• Memory, both volatile and non-volatile, dictates the capabilities of IoT devices, influencing features
like local data processing, storage, and filtering. Higher memory capacities enhance functionality but
may increase costs.
Processing Power:
• Vital for accommodating sensors and supporting on-site processing, processing power determines
the types of applications a device can handle. Applications dealing with video and image data
require higher processing power.
IoT Device Design and Selection
Considerations (cont..)
I/O Rating:
• The input-output (I/O) rating of the processor influences circuit complexity, energy usage, and
compatibility with various sensing solutions. Newer processors with lower I/O voltage may require
additional circuitry for legacy technologies, impacting costs.
Add-ons:
• Support for add-ons such as ADC units, built-in clock circuits, USB and Ethernet connections, and wireless
access capabilities enhances the robustness and usability of a processor or IoT device. Availability of these
options expedites solution development, making the device more attractive to users and developers.
Processing Offloading
• Remote Server: A remote server with good processing power is used with
IoT-based applications to offload processing from resource-constrained IoT
devices.
• Rapid scalability may be an issue with remote servers, and they might be
costlier and harder to maintain compared to cloud solutions.
• The naive approach to offloading data from IoT devices involves a simple, rule-
based method without much sophisticated (advanced) decision-making.
• While easy to implement, this approach is not recommended for most of the
scenarios.
• Not all parameters can be equally prioritized in maximizing the overall QoS.
Instead, it employs a bargaining strategy, attempting to find a balance where
the qualities of certain parameters are reduced while others are enhanced.
• Relies on past behavior and trends in data flow within the IoT architecture.
• Eg., Machine learning, use historical data to make decisions and continually
improve optimization strategies.