Communication advances Chapter | 2 19- in-vehicle intelligence should be enhanced, through enablers for supporting
novel concepts, such as, for example, 5G-D2D, so as to cater for ultra-low
latency next-generation V2X communications; - well-defined communication processes must be established, along with
interfaces for plugging-in disparate data sources enabling local and cross-
border CCAM. For example, the processing of 5G-oriented information can
be performed by mobile applications that constantly provide the RSSI/RSRP
signal level that the phone gets from the nearest BTSs (around ten stations
are used). This information can be combined with the location of BTS/Node-
Bs, the RAT (GSM, UMTS, LTE, etc.), and other information provided by
the 5G operator and using multi-angular calculations can be used to specify
the coordinates of a 5G-enhanced vehicle, its direction, its velocity, etc.
Virtualization and resource sharing are mainstream flexibilization strategies
to achieve stringent performance goals in cellular networks (Philip, Gourhant,
& Zeghlache, 2011). Diverse technologies (RAN virtualization, NFV, SDN)
have stemmed from this need for flexibility allowing, among other benefits, the
implementation of network slices—virtual networks with specific functionality
for a particular service or customer. This functionality can be used to differenti-
ate traffic classes with diverse requirements or even to implement virtual mobile
operators (VMOs).
For example, the recent SONATA project (Dräxler et al., 2017) has addressed
aspects such as orchestration of VMO resources. The MPC project has proposed
an SDN-based NFV-oriented mobile packet core to facilitate the dynamic provi-
sioning of network functions (Sama et al., 2015 ). Most existing approaches are
generalist, in the sense that they are neither optimized for particular applications
nor for particular traffic classes. Instead, they exploit virtualization to differenti-
ate the diverse cooccurring traffic classes in traditional telcos seeking to balance
the quality of service levels. For example, the exhaustive survey (Bizanis &
Kuipers, 2016) recognizes the interest of SDN technology to “group” IoT traf-
fic with similar requirements, but not the possibility of specialized IoT virtual
operators within a multioperator architecture.
Orchestration (Foukas, Nikaein, Kassem, Marina, & Kontovasilis, 2017)
and SDN technologies play key roles in the “specialized” scenario we foresee.
Regarding SDN control, it must react to migration decisions rerouting flows
between packet core processes. Besides, PHY-level RAN slicing may be inte-
grated within the same SDN control architecture. Multi-operator or multi-slice
SDN control should be driven by global optimization of the performance goals
of the actors in the scenario, according to its particular trade-offs (central pro-
cessing vs. edge computing, PHY costs vs. peak rates, etc).
Regarding RAN slicing, different works have studied it considering a particu-
lar PHY layer. For example, the COHERENT project (www.ict-coherent.eu) sepa-
rated the RAN data and control planes and coupled the latter with the virtualized
control functions. This way, it was possible to control the OpenAirInterface-based
RAN infrastructure following an SDN approach. The 5G-EmPOWER tool kit was
used to demonstrate a Wireless LAN hypervisor that could follow the dynamic