IEEE Consumer Electronics Magazine - March/April 2020 - 98

Special Section on Implementable Humanitarian Technology

Figure 1. Cloud-IoT convergence through a middleware fog.

Fog and cloud both provide similar resources
such as storage, computation, memory, and
infrastructure, and yet compliment each other.
A key difference is their location. Fog is in close
proximity to the IoT nodes in contrast to the
cloud which is further away and supports
more general functions (e.g., delay-tolerant data
storage, compute-intensive tasks). Hence, fog
becomes a de-facto middleware between health
IoT and the cloud, that is capable of providing
quick response to time-sensitive applications
(such as telemedicine), avoid unnecessary data
communication to the core network (by providing data aggregation and trimming to wireless
body area networks (WBANs) and other healthcare IoTs), and providing necessary privacy
measures to healthcare services.
The cloud receives filtered and structured data
and performs necessary operations on it. It

applies various machine learning techniques,
ontology, semantics, and data analytics to learn
more about the data and convert it into knowledge.12 In this way, the cloud creates tailored and
context-aware services (e.g., early diagnosis). Various services offered by the cloud can be accessed
by different stakeholders, such as doctors, hospitals, blood banks, insurers, other cloud service
providers, IoT service providers, and so on.
Figure 2 shows the layered architecture of a
health fog and the main functions performed at
each layer. The tasks a typical health fog has to
perform generally include interfacing (via
routers or gateways) with the underlying IoT
nodes or the cloud, performing the middleware
operations (such as data filtering, security measures, device/sensor monitoring), and transporting the ready-to-send data to the cloud. Sensors,
WBANs, wearables, and so on, are all part of the
physical and virtualization layer. The fog
receives the raw and unstructured data from the
underlying nodes. To be scalable and cost-effective, fog can deploy VS and VSNs. In complex
environments and dynamic conditions, VS can
be managed more efficiently, enabling contextaware service provisioning. The received data is
then handled by the monitoring layer which is
responsible for logging the data. The monitoring
layer is also responsible for managing the power
consumption of the sensors by conserving it
though proper sleep intervals as well as finding
other sources of power, such as human body
heat, solar, and wind. Next, the data goes to preprocessing layer which structures the unstructured data by applying various data filtering and
trimming techniques.
Once the data is comprehensive and structured, various local services can be provided at
the middleware layer. Such services may include a
quick health fix, alerts, precautionary measures,
information about local healthcare institutions
and physicians, and assistance for telemedicine
and high-definition video streaming.
Data storage requires privacy and security
measures at the middleware. The local storage
layer is in charge of such tasks. Light-weight
security protocols and data integrity techniques
need to be incorporated at this layer. Afterwards, data is stored for a certain period of time
by the temporary storage layer. Since the fog

IEEE Consumer Electronics Magazine

IEEE Consumer Electronics Magazine - March/April 2020

Table of Contents for the Digital Edition of IEEE Consumer Electronics Magazine - March/April 2020