Page 36 - ITU Journal Future and evolving technologies – Volume 2 (2021), Issue 2

ITU Journal on Future and Evolving Technologies, Volume 2 (2021), Issue 2

cess. The above works on optimal D2D cache placement the increasing number of users. In the distributed discov‑
are assumed to be implemented in a centralized manner. ery , users have to cope with the discovery process on
Moreover, these works did not address user and content their own, without any external assistance. It does not re‑
discovery, which is an important aspect of D2D commu‑ quire network coverage but is ineffective in the resource
nications. utilization.

Clustering is an important concept in ad hoc networks. It In the literature, several studies cover device and service
is used in order to group nearby users into a cluster and discovery. Authors have proposed a clustering algorithm
designate a cluster head, in order to facilitate distributed and a two‑step discovery procedure that aims to reduce
network functions. Clustering can be utilized in D2D com‑ signaling overhead and enables simultaneous discovery
munications in order to implement practical cache place‑ in [32] but cache placement and content delivery issues
ment and content delivery schemes. In the literature clus‑ are foreclosed. In [35] a hybrid scheme has been pro‑
tering in D2D communications is utilized in order to use posed in D2D, which includes a directdiscovery phase and
some devices as cluster heads (CH) that relay data com‑ network‑assisted data communication phase for proxim‑
ing to and from the base station. Clustering methods pro‑ ity services. In the discovery phase, authentication and
posed in [24] and [25] are based on merge and split op‑ authorization are provided by the operator initially upon
erations and they can be implemented in a centralized D2D request. The authors also stated that the network‑
manner. Authors in [26] proposed a clustering scheme assisted D2D communication phase satis ies users’ pri‑
that is within the con ines of LTE standards. Users de‑ vacy and security concerns. Another hybrid scheme is
tect their neighbors and report list of their neighbors to presented in [36] that aims improvement both in energy
the BS. Then the BS forms clusters based on the neigh‑ consumption and latency compared to direct discovery.
borhood information. In [27, 28, 29] the authors propose In [37] an Evolved Packet Core (EPC) assisted discovery
centralized clustering methods that are based on social as approach has been considered, where users activate their
well as physical layer attributes. In [30] cluster heads are D2D capabilities in speci ic areas after invoked by BS to
used as relays of scheduling requests in order to allevi‑ increase the success rate and energy ef iciency. An eNB
ate the congestion at the physical random access channel assisted discovery and D2D link setup procedure is pro‑
(PRACH). Cluster head selection is randomized and clus‑ posed in [38] and performance is evaluated in terms of
ter formation is distributed. The work in [31] utilizes dis‑ energy consumption and delay. Centralized and semi‑
tributed clustering for vehicular networks. In the above‑ centralized D2D discovery algorithms are introduced and
mentioned works, caching was not considered. Recent compared in terms of required time slots to discover all
works such as [32] and [33] did not consider cache place‑ new D2D pairs in the cell in [39]. In [40] two device dis‑
ment, but proposed clustering schemes based on cached covery schemes have been proposed for centralized and
content, as well as social and physical layer attributes. autonomous D2D networks that aim to reduce power con‑
sumption and signaling overhead. The authors also claim
The above‑mentioned works reveal that clustering is a
that the proposed schemes are applicable for cache en‑
useful technique and it’s been utilized for various goals
abled D2D networks. Contrary, in [41] users can discover
in LTE‑based D2D transmission literature. The authors in
each other not only directly but also via forwarding nodes.
[34] were the irst ones to jointly address D2D clustering To overcome excessive energy consumption in direct dis‑
and cache placement. They proposed an algorithm that covery procedures of Wi‑Fi Direct Technology, authors
clusters helper (caching) nodes and then allocates con‑ proposed an LTE network assisted peer discovery [42].
tent in each cluster member in a way that maximizes of‑
loading and performs very close to the optimal solution. As mentioned formerly, LTE Direct is the only technology
However, none of the above‑mentioned considered cache that enables D2D communication in the cellular spectrum
placement jointly with the required content/user discov‑ (also known as in‑band D2D). Two discovery schemes
ery operations. have been de ined in the technical speci ication docu‑
ments of the LTE Direct [43]. In the Network Assisted
A crucial point in D2D communications is the discovery scheme, UserEquipments (UEs)that havebeen registered
operation which might denote device, content, and ser‑ for D2D communication are warned about proximate de‑
vice discovery. Regardless of the purpose of discovery, vices. On the other hand, in the Direct Discovery, UEs
these operations cause messaging overhead and interfer‑ send discovery beacons to discover other UE and proxim‑
ence induces outages in networks with high user den‑ ity services in an unabetted way.
sity. Therefore, discovery in D2D communication has to
be handled wisely [4]. Discovery operation could be done 1.1 Contributions
in a centralized or distributed manner. In the network‑
assisted discovery, the BS or another centralized entity Apart from the existing literature, in this paper a compre‑
that possesses the network information provides proxim‑ hensive study on D2D communication, which covers most
ity information, channel state information, resource allo‑ of the key aspects of D2D technology such as user cluster‑
cation information to the users. However, it requires net‑ ing, cache placement, discovery, advertisement, and re‑
work coverage and computational complexity along with source utilization are presented.

22 © International Telecommunication Union, 2021