CFEC: An ultra-low latency Microservices-based In-Network Computing
Framework for Information-Centric IoVs
Abstract
The advancement of vehicular onboard units (OBUs) has led to
compute-intensive and delay-sensitive vehicular applications. Undeniably
edge-assisted static roadside computing terminal (sRCT) offers immediate
computations, a surge of smart
vehicles and intensive computation requests during crowded hours may
overload the sRCT, leading to performance degradation and intolerable
delays. Therefore, to facilitate proximate computations and achieve
ultra-low latency, this article envisions a Consortium of mobile
vehicular Fog, Edge, and Cloud (CFEC) an ultra-low latency
microservices-centric in-network computing framework for vehicular Named
Data networks (VNDN). CFEC develops a fog-profiler-assisted mobile
vehicular fog based on vehicles’ mobility patterns and available
resource characteristics to ensure reliable computation offloading and
reverse-path stability in a dynamic vehicular environment. Furthermore,
CFEC introduces an intermediary ZTMC controller that effectively filters
out underutilized sRCTs and routes computation requests to nearby,
filtered sRCTs, thus minimizing transmission time and accelerating
computations even during crowded hours. Simulations results revealed
that CFEC significantly reduces computational satisfaction delays by up
to 32.5%, 48.5%, and 31.9%, 51.025% against varying interest and
node rates, respectively while in extreme traffic conditions, CFEC
achieved an impressive computation satisfaction ratio of around 85%
compared with benchmark schemes.