D2D
In voice-centric systems it was implicitly accepted that two parties willing to establish a call
would not be in close proximity. In the age of data, this premise might no longer hold, and it
could be common to have situations where several co-located devices would like to wireless
share content (e.g., digital pictures) or interact (e.g., video gaming or social networking). Handling these communication scenarios via simply connecting through the network involves gross
inefficiencies at various levels:
• Multiple wireless hops are utilized to achieve what requires, fundamentally, a single hop.
This entails a multi fold waste of signaling resources as well as higher latency.
• Transmit powers of a fraction of a Watt (in the up link) and several Watts (in the down link)
are consumed to achieve what requires, fundamentally, a few milliWatts. This, in turn, entails unnecessary levels of battery drain and interference to all other devices occupying the same signaling resources elsewhere.
• Given that the path losses to possibly distant base stations are much stronger than direct link ones, the corresponding spectral efficiency are also lower. While it is clear that D2D has the potential to handle local communication more efficiently,local high-data-rate exchanges could also be handled by other radio access technologies such as Bluetooth or WiFi direct. Use cases requiring a mixture of local and non local content or a mixture of low-latency and high-data-rate constraints (e.g., interaction between users via augmented reality) could represent more compelling reasons for the use of D2D. In particular, we envision D2D as an important enabler for applications requiring low latency,4 especially in future network deployments utilizing base band centralization and radio virtualization. From a research perspective, D2D communication presents relevant challenges:
• Quantification of the real opportunities for D2D. How often does local communication
occur? What is the main use case for D2D: fast local exchanges, low-latency applications, or energy saving?
• Integration of a D2D mode with the up link/down link duplexing structure.
• Design of D2D-enabled devices, from both a hardware and a protocol perspective, by providing
the needed flexibility at both the PHYS and medium access control (MAC) layers.
• Assessing the true net gains associated with having a D2D mode, accounting for possible
extra overheads for control and channel estimation. Finally, note that while D2D is already being studied in the Third Generation Partnership Project (3GPP) as a 4G add-on (refer to footnote 2), the main focus of current studies is proximity detection for public safety [15]. What we are discussing here is having a D2D dimension native supported in 5G.