4 Architecting IoT Cloud
227
deployment, all you need to do is deploying your code. The vendor is responsible
for handling all the operational requirements (e.g., scaling). On the other hand,
in virtual machine or container-based deployment, you are responsible for
monitoring and managing your system yourself. The details of serverless and
FaaS will be covered in Sect. 4.10. As an example, AWS Lambda, one of
the most well-known serverless deployment technologies, is compatible with
Python, Java, and Node.js services. To deploy a service, one needs just to upload
the corresponding microservices packaged as ZIP files. AWS Lambda handles
microservice requests by automatically running a sufficient number of instances.
However, this technology is not suitable for deploying long-running services.
The reason is that all requests in AWS must finish in 300 seconds and all services
must be stateless.
4.8 Data Visualization and Reporting Layer
The data visualization layer monitors project success and is the means by which
users perceive data value.
4.8.1 Data Visualization Frameworks
• Kibana: As stated earlier, Kibana is a data visualization framework based on
Elasticsearch. It enables users to understand data through a dashboard depicting
a group of visualizations. Users can resize or rearrange data visualizations at will
and save it to the dashboard so that it can be reloaded or shared. Kibana’s main
focus is enabling users to explore and analyze Elasticsearch’s log data. It does
not support other data sources, so if you are not utilizing Elasticsearch, Kibana
is not a viable data visualization option.
• Grafana: Grafana is an all-purpose, open-source graph composer and dashboard
that functions as a web application. Grafana provides built-in means for obtaining
data from 30+ sources (i.e., Elasticsearch). It is best suited to visualize continuous time-series and streaming data such as metric reporting or sensor data.
4.8.2 Business Intelligence Frameworks
• Tableau: Tableau is a wonderful tool for data visualization because it is widely
available and enables the user to manipulate big data. It has two derivatives
including Tableau Server and the Cloud-based Tableau Online, specifically
designed to handle big data from organizations. Utilizing Tableau does not
require a coding skill. It comes with a user-friendly dashboard with drag and
227
deployment, all you need to do is deploying your code. The vendor is responsible
for handling all the operational requirements (e.g., scaling). On the other hand,
in virtual machine or container-based deployment, you are responsible for
monitoring and managing your system yourself. The details of serverless and
FaaS will be covered in Sect. 4.10. As an example, AWS Lambda, one of
the most well-known serverless deployment technologies, is compatible with
Python, Java, and Node.js services. To deploy a service, one needs just to upload
the corresponding microservices packaged as ZIP files. AWS Lambda handles
microservice requests by automatically running a sufficient number of instances.
However, this technology is not suitable for deploying long-running services.
The reason is that all requests in AWS must finish in 300 seconds and all services
must be stateless.
4.8 Data Visualization and Reporting Layer
The data visualization layer monitors project success and is the means by which
users perceive data value.
4.8.1 Data Visualization Frameworks
• Kibana: As stated earlier, Kibana is a data visualization framework based on
Elasticsearch. It enables users to understand data through a dashboard depicting
a group of visualizations. Users can resize or rearrange data visualizations at will
and save it to the dashboard so that it can be reloaded or shared. Kibana’s main
focus is enabling users to explore and analyze Elasticsearch’s log data. It does
not support other data sources, so if you are not utilizing Elasticsearch, Kibana
is not a viable data visualization option.
• Grafana: Grafana is an all-purpose, open-source graph composer and dashboard
that functions as a web application. Grafana provides built-in means for obtaining
data from 30+ sources (i.e., Elasticsearch). It is best suited to visualize continuous time-series and streaming data such as metric reporting or sensor data.
4.8.2 Business Intelligence Frameworks
• Tableau: Tableau is a wonderful tool for data visualization because it is widely
available and enables the user to manipulate big data. It has two derivatives
including Tableau Server and the Cloud-based Tableau Online, specifically
designed to handle big data from organizations. Utilizing Tableau does not
require a coding skill. It comes with a user-friendly dashboard with drag and