530
B. Farahani et al.
Late-binding ELT approach, leading to more flexibility and faster access to
all data at any time responding to any and all future needs.
3. Data warehouse: It is a highly structured repository used mainly for reporting
and representing an abstracted picture of the e-Health system. The data stored
in this repository can be uploaded from the data lake or from the operational
systems (such as sales). However, note that before storing any data, we need to
process, model, and give the data a specific structure.
4. Data flow manage and orchestration: This layer is responsible for managing,
automating, and orchestrating the cloud sources (e.g., data ingestion, data
storage, data processing, and visualization).
5. User, device, and data management: The cloud integrates data from multiple
sources. It captures data from many fog nodes and stores the data in a safe and
secure manner. In this way, the data is always there to be accessed by those
engaged in patient care. This platform seamlessly integrates with non-sensor
sources such as EHRs, e-prescriptions, web sources, and more. As a result,
patients, physicians, or any other member of a patient’s care team can access
vital health data when needed. This significantly increases collaboration across
all disciplines, increasing the efficiency of the healthcare plan. Moreover, cloudbased platforms offer a unified schema to capture and query transactions. In
doing so, the versatility to create new applications is increased. This module is
also used to manage users, groups, devices, and fog nodes and access permissions
and roles.
6. Big Data analytics: This is a key component for analyzing medical data. The
use of analytics allows the platform to use event- and rule-based processing,
data mining, machine learning, and automated reasoning-based algorithms on
stored historical records. This way, the platform can make meaningful insights
about patient health. Having these early health insights could be a game-changer
for a patient who can begin to take preventative action against an otherwise
fatal ailment. The configurations of the connected e-Health devices can also be
adjusted using the extracted insights. For instance, users may alter the frequency
and type of information collected, as well as the multimedia (images and videos)
resolution. Note that, there are two different engines in the analytical module.
The first one handles all requests that are subject to (near) real-time constraints.
The second one deals with batch data and extracting historical intelligence.
7. Output integration: This is typically based on an Enterprise Service and Integration Bus such as Apache Camel with a rich set of protocols (e.g., REST
API, Message Broker, Websocket, etc.) that enables connection with any system,
application, or portal. It should also be noted that this module can exploit inmemory databases (such as Redis and HBase) to answer fast incoming queries
by merging and caching the results from warehouse, data lake, analytics, etc.
However, this database only stores very recent data and results.
B. Farahani et al.
Late-binding ELT approach, leading to more flexibility and faster access to
all data at any time responding to any and all future needs.
3. Data warehouse: It is a highly structured repository used mainly for reporting
and representing an abstracted picture of the e-Health system. The data stored
in this repository can be uploaded from the data lake or from the operational
systems (such as sales). However, note that before storing any data, we need to
process, model, and give the data a specific structure.
4. Data flow manage and orchestration: This layer is responsible for managing,
automating, and orchestrating the cloud sources (e.g., data ingestion, data
storage, data processing, and visualization).
5. User, device, and data management: The cloud integrates data from multiple
sources. It captures data from many fog nodes and stores the data in a safe and
secure manner. In this way, the data is always there to be accessed by those
engaged in patient care. This platform seamlessly integrates with non-sensor
sources such as EHRs, e-prescriptions, web sources, and more. As a result,
patients, physicians, or any other member of a patient’s care team can access
vital health data when needed. This significantly increases collaboration across
all disciplines, increasing the efficiency of the healthcare plan. Moreover, cloudbased platforms offer a unified schema to capture and query transactions. In
doing so, the versatility to create new applications is increased. This module is
also used to manage users, groups, devices, and fog nodes and access permissions
and roles.
6. Big Data analytics: This is a key component for analyzing medical data. The
use of analytics allows the platform to use event- and rule-based processing,
data mining, machine learning, and automated reasoning-based algorithms on
stored historical records. This way, the platform can make meaningful insights
about patient health. Having these early health insights could be a game-changer
for a patient who can begin to take preventative action against an otherwise
fatal ailment. The configurations of the connected e-Health devices can also be
adjusted using the extracted insights. For instance, users may alter the frequency
and type of information collected, as well as the multimedia (images and videos)
resolution. Note that, there are two different engines in the analytical module.
The first one handles all requests that are subject to (near) real-time constraints.
The second one deals with batch data and extracting historical intelligence.
7. Output integration: This is typically based on an Enterprise Service and Integration Bus such as Apache Camel with a rich set of protocols (e.g., REST
API, Message Broker, Websocket, etc.) that enables connection with any system,
application, or portal. It should also be noted that this module can exploit inmemory databases (such as Redis and HBase) to answer fast incoming queries
by merging and caching the results from warehouse, data lake, analytics, etc.
However, this database only stores very recent data and results.