© 2018 IJRAR January 2019, Volume 6, Issue 1
www.ijrar.org (
E-ISSN 2348-1269, P- ISSN 2349-5138
)
IJRAR19J1651
International Journal of Research and Analytical Reviews (IJRAR)
www.ijrar.org
1389
The hierarchical architecture is composed of the following three layers:
Terminal layer: This is the layer closest to the end user and physical environment. It consists
of various IoT devices, for
example, sensors, mobile phones, smart vehicles, smart cards, readers, and so on. Specially, though the mobile phones and smart
vehicles have the computing power, we only utilize them as the smart sensing devices here.
These devices are widely
geographically distributed in general. They are responsible for sensing the feature data of physical objects or events and
transmitting these sensed data to upper layer for processing and storage [9].
Fog layer: This layer is located on the edge of the network. Fog computing layer is composed of a large number of fog nodes,
which generally including routers, gateways, switchers, access points, base stations, specific fog servers, etc. These fog nodes are
widely distributed between
the end devices and cloud, for example, cafes,
shopping centers, bus terminals, streets, parks, etc.
They can be static at a fixed location, or mobile on a moving carrier. The end devices can conveniently connect with fog nodes to
obtain services. They have
the capabilities to compute, transmit and temporarily store the received sensed data. The real-time
analysis and latency-sensitive applications can be accomplished in fog layer. Moreover, the fog nodes
are also connected with
cloud data center by IP core net- work, and responsible for interaction and cooperation with cloud
to obtain more powerful
computing and storage capabilities [9].
Cloud layer: The cloud-computing layer consists of multiple high-performance servers and storage devices, and provides
various application services, such as smart home, smart transportation, smart factory, etc. It has powerful computing and storage
capabilities to support for extensive computation analysis and permanently storage of an enormous amount of data. How- ever,
different from traditional cloud computing architecture, not all computing and storage tasks go through the cloud. According to
the demand-load, the cloud core modules are efficiently managed and scheduled by some control strategies to improve utilization
of the cloud resources [18].
