Unit 4 IoT (1) Highlighted PDF

Download as pdf or txt
Download as pdf or txt
You are on page 1of 16

Implementation of Device integration in IoT

Implementing device integration in IoT (Internet of Things) involves connecting and enabling
communication between various IoT devices, sensors, and actuators to collect, transmit, and process
data. Below are the steps and considerations for implementing device integration in an IoT system:

1. Device Selection: Choose the IoT devices and sensors that are suitable for your application.
Consider factors such as sensor types (temperature, humidity, motion, etc.), communication
protocols (Wi-Fi, Bluetooth, Zigbee, LoRaWAN, cellular, etc.), and power requirements.

2. Communication Protocols: Select appropriate communication protocols for your IoT devices.
Common protocols include MQTT, CoAP, HTTP/HTTPS, WebSocket, and others. Ensure that
the chosen protocols are compatible with your devices.

3. Device Registration: Develop a mechanism to register and identify IoT devices in your
system. Each device should have a unique identifier, such as a MAC address or device ID,
which helps in tracking and managing them.

4. Network Configuration: Set up the necessary network infrastructure to connect your


devices. This may include configuring Wi-Fi routers, gateways, or cellular modems,
depending on the communication methods used by your devices.

5. Data Ingestion: Create a data ingestion layer to receive data from IoT devices. This layer can
be hosted on the cloud, on-premises, or at the edge, depending on your architecture. Ensure
that it can handle data streams from multiple devices simultaneously.

6. Security Measures: Implement robust security measures to protect data in transit and at
rest. Use encryption (e.g., TLS/SSL) for secure communication, authentication to ensure
device identity, and access control to manage device permissions.

7. Data Storage: Choose an appropriate data storage solution to store the data generated by
IoT devices. This can be a cloud-based database, a distributed storage system, or an on-
premises solution. Ensure scalability and redundancy for data reliability.

8. Data Processing: Implement data processing logic to analyze, transform, and filter the
incoming data. This may include data normalization, aggregation, and real-time analytics
based on your application's requirements.

9. Device Management: Develop a device management system to remotely configure and


monitor IoT devices. This includes firmware updates, over-the-air (OTA) updates, and
monitoring device health and status.

10. Integration with Applications: Integrate the processed IoT data with your applications or
services. Create APIs or connectors that allow your applications to access and utilize the IoT
data for real-time insights and decision-making.

11. Scalability and Load Balancing: Design the system to handle a growing number of IoT
devices and data volume. Implement load balancing and auto-scaling mechanisms to ensure
performance and availability.

12. Testing and Quality Assurance: Thoroughly test the IoT device integration system, including
functional testing, performance testing, and security testing. Ensure that it meets the
defined requirements and performs reliably.
13. Monitoring and Alerting: Set up monitoring tools to continuously monitor the health and
performance of IoT devices and the integration infrastructure. Implement alerts and
dashboards for proactive maintenance and issue resolution.

14. Documentation: Document all aspects of your IoT device integration implementation,
including device configurations, data schemas, API documentation, and troubleshooting
guides.

15. Compliance and Regulations: Ensure that your IoT system complies with relevant industry
standards and regulations, especially if you are handling sensitive or regulated data.

16. Lifecycle Management: Plan for the entire lifecycle of IoT devices, including provisioning,
activation, decommissioning, and recycling or disposal of devices.

Remember that IoT device integration can vary significantly depending on your specific use case and
technology stack. It's essential to carefully plan, implement, and maintain your IoT integration
solution to ensure its success. Additionally, consider leveraging IoT platforms and frameworks to
streamline the integration process and reduce development effort.
Data acquisition and integration
Data Acquisition
i. Data acquisition means collecting data from IoT or M2M devices. The data transfers after
the interactions with a data acquisition system of the application. The application interacts
and communicates with a number of devices for acquiring the needed data. The devices send
data on demand or at programmed intervals. Data of devices communicate using the network,
transport and security layers.
ii. When devices have the ability to configure themselves, an application can setup them for
data. The system, for example, can configure devices to send data at predefined intervals. The
frequency of data generation is determined by the device settings. For example, the system
can design an umbrella device to obtain weather data from an Internet weather service once
every week.
iii. Every hour, Vending machine can be programmed to communicate machine sales
statistics and other information. The Vending machine system can be programmed to
communicate instantly in the event of a problem or when a specific chocolate flavour requires
the fill service.
iv. The data-adaptation layer allows application to customize data sending after filtering or
enriching. The application-to-device’s gateway can do transcoding, data management, and
device management. Data management may include data integration, compaction, and fusion.
vi. Device-management software maintains the device ID or address, activation, configuring
registering, deregistering, attaching, and detaching
Data Integration
i. Data integration is the process of combining data from different sources into a single,
unified view. Data integration ultimately enables analytics tools to produce effective,
actionable business intelligence.
ii. There is no universal approach to data integration. However, data integration solutions
typically involve a few common elements, including a network of data sources, a master
server, and clients accessing data from the master server.
iii. In a typical data integration process, the client sends a request to the master server for
data. The master server then intakes the needed data from internal and external sources. The
data is extracted from the sources, then consolidated into a single, organize data set.
Integration helps businesses succeed
iv. Advantages of data integration are Improves collaboration and unification of systems,
saves time and boosts efficiency, reduces errors (and rework), delivers more valuable data.

To assess traffic on a busy highway, researchers used a ThingSpeak, Raspberry Pi 3


Model B, and a webcam to construct an Internet of Things (IoT) based smart real-time
traffic monitoring system. The Raspberry Pi device should then be connected to the
machine for deployment and configuration of the traffic-monitoring algorithm. ThingSpeak is a
cloud aggregator that allows you to store, analyse, and visualize data online.

Analytics are employed everywhere in the Internet of Things (IoT):


1. The edge node
2. Offline on the desktop
3. In the cloud
We've demonstrated how to monitor real-time traffic data on a busy roadway and how to collect
real-time data on the cloud. After storing the aggregated data on the cloud, the
edge device can perform analytics, and then the cloud data can be analysed. The
diagram shows a workflow analysis framework and how to execute online analysis on
cloud data. ThingSpeak is used for both storage and analysis in this system. As indicated in Fig,
the suggested architecture based on the Internet of Things might
include four layers. The Sensing layer, Data processing layer, IoT sensor data analysis
layer, and Application layer are the four layers.

1. Session layer - This layer is responsible for collecting data from various IoT devices.
In the suggested framework, this is the lowest layer that can be found. It gathers
different heterogeneous types of data gathered from heterogeneous IoT devices
such as sensors, actuators, protocols, and data formats, among others. The large
volume of big data is sent to the next level layer, the Data processing layer, for data
analysis and raw data processing.

2. Data processing layer - This layer is responsible for obtaining the required data from
the raw data obtained by the physical layer. Feature extraction and Foreground Detector data are
the two sub-modules that make up the data processing layer. To manage Realtime data, feature
extraction is employed to extract raw data using camera vision algorithms such as SPP-net and
GMM. This layer is responsible for obtaining the required data from the raw data obtained by
the physical layer. Feature extraction and Foreground Detector data are the two sub-modules that
make up the data processing layer. To manage the real-time data, feature extraction is employed
to extract raw data using camera vision algorithms such as SPP-net and GMM.

3. Data analysis layer - This is the most important layer in the proposed framework.
This layer has had the Simulink model applied to it. It will use the data once it has been
pre-processed on the data processing layer to annotate the IoT sensor data. The
Simulink model is used to reduce the ambiguity of the foreground detector data. The
domain knowledge systems based on ThingSpeak have employed context-aware data to forecast
real-time traffic statistics in order to reduce traffic congestion and unexpected delays.

4. Application layer - This layer is used to analyse and visualize the data collected from
IoT sensors for users on the ThingSpeak cloud platform. In real-time settings, this layer
generates traffic alerts and vehicle counts. These data can be used in IoT smart city
applications such as smart traffic, smart homes, smart grids, and smart health, among
others.
Fig 1: The proposed framework for IoT sensor data analysis
what is data store

A data store is a generic term used to describe a repository or storage system where data is stored,
organized, and managed. Data stores can take various forms and serve different purposes, depending
on the specific needs and requirements of an organization or application. Here are some common
types of data stores:

1. Databases: Databases are structured data stores that use a systematic approach to organize
and manage data. They can be relational databases (e.g., MySQL, PostgreSQL) or NoSQL
databases (e.g., MongoDB, Cassandra), depending on the data model and requirements.

2. File Systems: File systems are used to store unstructured or semi-structured data in files and
folders. They are commonly used for storing documents, images, videos, and other types of
files. Examples include NTFS for Windows and ext4 for Linux.

3. Data Warehouses: Data warehouses are specialized databases designed for storing and
analyzing large volumes of historical data. They are often used in business intelligence and
analytics to support reporting and decision-making.

4. Key-Value Stores: Key-value stores are a type of NoSQL database that store data as a
collection of key-value pairs. Examples include Redis and Amazon DynamoDB.

5. Document Stores: Document stores are NoSQL databases that store data in semi-structured
documents, typically in formats like JSON or XML. MongoDB is a popular example.

6. Columnar Stores: Columnar databases store data in columns rather than rows, which can be
more efficient for certain analytical workloads. Apache Cassandra and Amazon Redshift are
examples.

7. Object Stores: Object stores, like Amazon S3, are designed for storing large volumes of
unstructured data, such as images, videos, and backups, and are often used in cloud
computing environments.

8. Memory Stores: Memory stores keep data in memory for ultra-fast access times but may not
be suitable for long-term storage. Examples include Redis and Memcached.

9. Graph Databases: Graph databases are designed for storing and querying graph-like data
structures, making them suitable for applications like social networks and recommendation
systems. Neo4j is a well-known graph database.

10. Blockchain: Blockchain technology is a distributed and decentralized data store that
maintains a secure and immutable ledger of transactions. It is commonly associated with
cryptocurrencies like Bitcoin.

11. Data Lakes: Data lakes are storage repositories that can hold vast amounts of raw data in its
native format until it's needed. They are often used for big data and analytics purposes.

The choice of a data store depends on factors such as data volume, data structure, query
requirements, scalability needs, and the specific use case of an application or organization. Many
modern systems use a combination of these data stores to meet their various data storage and
processing needs.
What are the different schemas for a data store?

In the context of data storage and databases, a schema is a blueprint or structural design that defines
the organization, format, and relationships of data within a database. There are several types of
schemas, each with its own characteristics and use cases. Here are some of the most common
database schemas:

1. Single-Schema (Monolithic Schema):

• In a single-schema design, all data is stored in a single, unified structure.

• It is simple and easy to manage for small applications or when the data structure is
straightforward.

• Not suitable for complex or evolving data requirements as changes can be


challenging to implement without affecting the entire database.

2. Flat Schema:

• In a flat schema, data is stored in a single table without any complex relationships.

• It's suitable for storing simple, non-relational data where relationships between
entities are minimal.

3. Hierarchical Schema:

• Data in a hierarchical schema is organized in a tree-like structure, with parent-child


relationships.

• Often used for representing data where there are strict one-to-many relationships,
like organizational charts.

4. Star Schema:

• Commonly used in data warehousing, the star schema consists of a central fact table
surrounded by dimension tables.

• Fact tables contain quantitative data, while dimension tables contain descriptive
information.

• This schema simplifies complex queries for data analysis and reporting.

5. Snowflake Schema:

• A variation of the star schema, the snowflake schema further normalizes dimension
tables into sub-dimensions.

• It reduces data redundancy but can make queries more complex.

6. Normalized Schema:

• In a normalized schema, data is organized to minimize redundancy by breaking it into


smaller related tables.

• This design is used to ensure data integrity and reduce data anomalies.

• While it's good for data consistency, it can lead to more complex queries.
7. Denormalized Schema:

• In contrast to normalized schemas, denormalized schemas store redundant data to


improve query performance.

• It's often used in situations where read-heavy operations are common and the cost
of storage is not a concern.

8. NoSQL Schemaless:

• NoSQL databases like MongoDB or Cassandra often use a schemaless approach.

• Data is stored in a flexible, schema-free format, which can adapt to changing data
requirements.

• It's suitable for scenarios where data structures are dynamic or evolving rapidly.

9. Graph Schema:

• Used in graph databases like Neo4j, a graph schema defines nodes and relationships
between them.

• Ideal for representing complex, interconnected data, such as social networks or


recommendation systems.

10. XML Schema:

• Used in XML databases, an XML schema defines the structure of XML documents.

• It specifies the elements, attributes, and their relationships.

The choice of schema depends on the specific requirements of the application, the complexity of the
data, and the anticipated query patterns. Many databases also allow for a hybrid approach, where
different schema types can be used within the same database to optimize performance and flexibility
for different use cases.
Unstructured data
It is the data which does not follows to a data model and has no easily
identifiable structure such that it cannot be used by a computer program easily.
Unstructured data is not organised in a pre-defined manner or does not have a pre-defined
data model, thus it is not a good fit for a mainstream relational database.

Characteristics of Unstructured Data:


i. Data neither conforms to a data model nor has any structure.
ii. Data cannot be stored in the form of rows and columns as in Databases
iii. Data does not follow any rules
iv. Data lacks any particular format or sequence
v. Data has no easily identifiable structure
vi. Due to lack of identifiable structure, it cannot used by computer programs easily

Sources of Unstructured Data:


Web pages, Images (JPEG, GIF, PNG, etc.), Videos, Memos, Reports, Word documents
and PowerPoint presentations, Surveys

Key differences between structured and unstructured data


While structured (quantitative) data gives a “birds-eye view” of customers, unstructured
(qualitative) data provides a deeper understanding of customer behaviour and intent. Let’s
explore some of the key areas of difference and their implications:
Sources: Structured data is sourced from GPS sensors, online forms, network logs, web
server logs, etc., whereas unstructured data sources include email messages, word-processing
documents, PDF files, etc.
Forms: Structured data consists of numbers and values, whereas unstructured data consists of
sensors, text files, audio and video files, etc.
Models: Structured data has a predefined data model and is formatted to a set data structure
before being placed in data storage (e.g., schema-on-write), whereas unstructured data is
stored in its native format and not processed until it is used (e.g., schema-on-read).
Storage: Structured data is stored in tabular formats (e.g., excel sheets or SQL databases)
that require less storage space. It can be stored in data warehouses, which makes it highly
scalable. Unstructured data, on the other hand, is stored as media files or NoSQL databases,
which require more space. It can be stored in data lakes which makes it difficult to scale.
Uses: Structured data is used in machine learning (ML) and drives its algorithms, whereas
unstructured data is used in text mining.

Device data storage - Unstructured data storage on cloud/local server


i. Unstructured data storage use is very common as many systems have to upload
attachments, images, press releases and document management functions.
ii. Unstructured data is often relatively large, it will take up more bandwidth and a certain
server computing power, which has some influence on some of the high performance
requirements of servers.
iii. When the application needs to use more clustering support, the traditional way will be
more difficult to work with. As many servers are invaded by Trojan virus due to vulnerability
upload files.
iv. For unstructured data storage in a traditional file system (Cloud storage is not required), a
directory of the file system must be editable.
v. Technically advanced cloud storage has certain advantages. Cloud storage in the form of
object storage for storing and reading in charge of the actual content of the document.
vi. Cloud storage of high stretch, massive high reliability, duplicate files merge, will help
improve the quality of storage service. Unstructured data cloud storage architecture built on
this design, the hierarchy shown in Figure 2.
vii. Application layer provides unstructured data application interfaces. These interfaces
various types of storage applications such as online storage, network drives, video, data
hosting and software download services.
viii. Session layer is responsible for user management, rights assignment, space allocation
and storage security policy, the layer depending on the security level. Develop the different
security programs to ensure data security.
ix. The role of the data layer is a unified management of unstructured data and metadata.
Unstructured data volume level from MB to GB sizes. Its metadata information, such as data
identification, file length, type and other attribute information, the total length of no more
than 1 KB

Figure 1Cloud storage tier architecture of unstructured data

You might also like