F. Giannone

Virtualized eNB latency limits

In flexible functional split, functions of a virtualized evolved NodeB (eNB) can be disaggregated in distributed computational resources. One of the main constraints for their placement is the latency experienced by the communication between the Virtual Machines (VM) hosting the functions. This paper evaluates experimentally the latency limits for different functional splits providing insights on flexible functional split implementation.

Impact of CoMP VNF placement on 5G Coordinated Scheduling performance

To address demanding requirements in terms of expected throughput, latency and scalability, 5G networks will offer high capacity to support huge volumes of traffic generated by heterogeneous services. Dense deployment of small cells can provide a valid solution but are prone to high levels of interference especially at the cell-edge. However, to reduce inter-cell interference and improve cell-edge throughput, a set of techniques known as Coordinated Multipoint (CoMP) has been introduced. Coordinated Scheduling (CS) is a CoMP technique that assigns resources to mobile users to avoid interference between users that are assigned within the same Physical Resource Blocks (PRBs). On the other hand, Software Defined Mobile Networking (SDMN) and Network Function Virtualization (NFV) represent two key technologies to enhance flexibility and efficiency of resource usage within the Radio Access Network (RAN). However, the implementation of CoMP CS techniques on NFV architecture in a dense small cell scenario have not been analyzed yet. In this paper, we propose the joint use of CoMP CS and NFV by studying the implications of different deployment strategies, as constrained by the physical topology of the underlying RAN. The performance of both distributed and centralized CoMP CS are compared in terms of convergence delay and traffic overhead. Guidelines for the optimal design are provided.

Requirements for 5G fronthaul

This paper first overviews possible evolved NodeB (eNB) functional splits. Then it illustrates which requirements the Radio Access Network (RAN) shall satisfy to effectively transport 5G protocol data units (PDUs) as a function of the considered functional split.

Performance evaluation of CoMP coordinated scheduling over different backhaul infrastructures: A real use case scenario

This work studies how cell information distribution for Coordinated Scheduling in Cooperative Multi Point Transmission might be affected by the backhaul infrastructure. Several CoMP schemes are investigated and a novel OSPF based advertisement protocol for CoMP information distribution is proposed. Star topology with centralization at macro cell is shown to be the most convenient choice in terms of both convergence delay and deployment costs. The proposed protocol represents a feasible solution preserving network configurability and management.

Orchestrating Lightpath Recovery and Flexible Functional Split to Preserve Virtualized RAN Connectivity

In the next-generation radio access network (NG RAN), the next-generation evolved NodeBs (gNBs) will be, likely, split into virtualized central units (CUs) and distributed units (DUs) interconnected by a fronthaul network. Because of fronthaul latency and capacity requirements, optical metro-ring networks are among the main candidates for supporting converged 5G and non-5G services. In this scenario, a degradation in the quality of transmission of the lightpaths connecting DU and CU can be revealed (or anticipated) based on monitoring techniques. Thus, the lightpath transmission parameters can be adapted to maintain the required bit error rate (BER). However, in specific cases, the original requested capacity between DU and CU could be not guaranteed, thus impacting the service. In this case, another DU-CU connectivity should be considered, relying on a change of the so-called functional split. This study proposes a two-step recovery scheme orchestrating lightpath transmission adaptation and functional split reconfiguration to guarantee the requested connectivity in a virtualized RAN fronthaul. Results show that, for the connections that cannot be transported by the original lightpath, a graceful degradation followed by a recovery is possible within tens of seconds.

How much is fronthaul latency budget impacted by RAN virtualisation

In the New Radio Access Network architecture (New RAN), currently envisioned by 3GPP, the evolved NodeB (eNB) functions can be split between a Distributed Unit (DU) and Central Unit (CU). Furthermore, as per the Virtual RAN (VRAN) approach, such functions can be virtualised (e.g., in simple terms, deployed in virtual machines). In such scenario, the fronthaul network connecting DU and CU must fulfill different latency and capacity requirements based on the selected functional split. This study experimentally evaluates in a federated testbed how the fronthaul latency budget (i.e., the latency requirement of the fronthaul network connecting DU and CU), specified by Standard Developing Organisations (SDO) (3GPP in this specific case), is impacted by virtualising some of the RAN functions. In particular, Option 7-1 (i.e., intra-PHY split) and different virtualisation methods are considered for the CU. Furthermore, it evaluates how jitter (i.e., delay variation) impacts the DU-CU communication. The obtained results show that light virtualisation methods (e.g., Docker) impact less the fronthaul latency budget than heavy virtualisation methods (e.g., VirtualBox). In addition, a maximum jitter of about 40s can be tolerated in the fronthaul.