Cloud-Fog-Dew Architecture for Refined Driving
Assistance
The Complete Service Computing Ecosystem
Tushar S Mane.
TASM2M,
Total Automation Solutions,
Pune, India.
tushar.mane@tasind.com
Himanshu Agrawal.
Department of Computer Science,
Symbiosis Institute of Technology,
Pune, India.
himanshu.agrawal@sitpune.edu.in
Abstract— Mapping services like Google Maps, Bing Maps,
and Mapbox etc. have been assisting the vehicle drivers for real-
time traffic, shortest routes, locations and road topologies since
several years, and these services are evolving unceasingly. Road
accidents is a prominent cause of deaths in India and so does the
road conditions. State of art driving assistance systems guide the
drivers for sharp turns and road intersections but they lack in
providing the fine grained assistance like potholes, speed
breakers, sudden inclinations, and declinations on the road
surface. The foremost hurdles for such highly sensitive systems
are latency requirements, affordability, and high availability.
With cloud computing, we can have high availability and
affordability but it may not offer quick response due to its
considerable round trip time. Fog computing gives huge
advantage on latency side, geo-distribution, mobility, and
affordability but high availability (fault-tolerance) cannot be
always guaranteed due to its substantial geo-distribution.
Moreover, end nodes, in both the cases acts purely as clients.
With dew computing, we propose to allocate computing tasks on
end devices and make them service providers instead of service
consumers. The projected approach is heavily based on peer-to-
peer communication which would be complementary in nature
with cloud and fog. In this article, we explain how a driver can
experience a fine-grained driving assistance if we intertwine all
the three computing paradigms together. As per the knowledge
of authors, this is the first ever proposal till date which entangles
all three computing paradigms (Cloud, Fog and Dew) together.
Keywords— Dew Computing; Fog Computing; Cloud
Computing; Service Computing; Distributed Systems, Driving
Assistance.
I. PROLOGUE
Road conditions, especially in India, are often varying and
irregular. State of the road in the morning time may change in
the afternoon and one in the afternoon cannot be assured to be
the same in the evening. Reasons being, laying water pipelines,
installing electricity poles, traffic signals, telephone lines and
other maintenance activities. Nature can be the cause too.
Heavy rains, lightning strike, overflowed river water, overheat
can affect the road surface severely. Sometimes roads/ bridges
are too worn-out to deform or collapse overnight [1]. Also,
more than 40% of overall roads in India don’t have street
lamps for night vision. Even in cities, due to poor maintenance
of streetlights, roads are not always guaranteed to be lightened.
Unnoticed or unknown speed breakers or potholes, many
times, lead to accidents. Uneven road surfaces like sudden
inclinations or declinations are causes for many accidents too.
Solutions like Google Maps offer an excellent aid for
shortest route discovery, real-time traffic, location services and
route re-calculation [2]. They also offer, to some extent,
topological guidance like sharp turns and road intersections.
But they lack in fine-grained assistance like alerting for
impending speed breaker or potholes, a sudden change in slope
of the road etc. during the route. India is a developing country,
most of the geography of India is still unmapped. Except for
cities, wireless network infrastructure/ technology is still a
fascination for many villages in India. Also in cities, not
everyone can afford high-speed connectivity or adapt to
emergent technology as it gets available. Affordability is
directly proportional to penetration of technology, surplus
discounted solutions get accepted on a colossal scale,
regardless of region (urban/ rural).
II. I
NTRODUCTION
Cloud computing is trending on top in IT market since
2006. Virtualization, resource pooling, elasticity, self and on-
demand measured services, multitenant architecture etc. are
some of the sophisticated features of cloud which make it
market dominating solution. Cloud can be an economical
solution but faces two major problems. Lack of network
infrastructure is a major roadblock for use of cloud in rural
areas [3]. Whereas high latency makes it unsuitable for use
even in cities or suburban areas (network infrastructure
enriched areas) [4]. The system should support for significant
geo-distribution (load balanced distributed systems), high
mobility (dynamic environments) and low latency (at the edge
services). Fog computing stands as the best solution to
overcome these challenges. Instead of, single server, placed
remotely, serving large geographical areas, many lightweight
nodes are distributed over the objective regions; computation is
pushed down at the edge devices like routers and gateways to
cater low latency needs in mobile environments. To maintain
high availability/ fault tolerance, these nodes should be
ª*&&&