Interference Management for 5G A METIS-II Perspective E. Pateromichelakis 1 , H. Celik 2 , R. Fantini 3 , Ö. Bulakci 1 , L. M. Campoy 4 , A. M. Ibrahim 5 , F.J. Lorca 4 , M. Shariat 6 , M. Tesanovic 6 , Y. Yang 5 1 Huawei Technologies ERC, Germany; 2 KTH, Sweden; 3 Telecom Italia, Italy; 4 Telefónica, Spain; 5 Intel Deutschland GmbH, Germany; 6 Samsung Electronics R&D Institute, UK Abstract—This paper provides highlights on the envisioned interference management framework for 5G RAN, which is pursued by 5G PPP METIS-II project. In this framework, factors like the expected high density of access nodes reusing the same spectrum, the diverse sources of interference from heterogeneous access technologies and the consideration of multiple 5G services with different KPIs can have strong impact on the user’s performance. To this end, three key interference management drivers are discussed, which aim improving users’ performance in terms of cell-edge throughput, provide energy- efficiency aware resource management and minimize the signalling overhead using BS clustering and context-awareness. Keywords—5G; Interference Management; Energy Efficiency; METIS-II I. INTRODUCTION In fifth generation (5G) radio access networks (RANs), new challenges in terms of interference will arise mainly due to the employment of ultra-dense networks (UDNs). Furthermore, numerous other factors will strongly affect the way interference management is handled, e.g. the wide usage of beam-forming, the uplink/downlink (UL/DL) cross-interference in case of dynamic time-division duplex (TDD), novel modes of communication (e.g., self-backhauling, and cellular assisted device-to-device, D2D), and more diverse and stringent application requirements, e.g., latency-critical applications. It is, thus, required to develop an interference management functionality block that natively supports the new communication variants and effectively satisfies very different and demanding performance requirements. Interference management in current cellular networks has been extensively studied in literature [1]. Various power control techniques have also been developed to provide enhanced performance within the network. In Long-Term Evolution- Advanced (LTE-Advanced) networks, such mechanisms were mainly studied and standardized from an UL perspective with a particular focus on keeping the receiver dynamic range below a pre-determined level [1] [2]. Nevertheless, in the case of ultra- dense deployments of access nodes in 5G networks, interference management schemes are gaining even more relevance and shall be tailored for the dynamic operation envisioned in such networks [3]. In addition, 5G RAN is expected to operate on various bands (below and above 6 GHz) and support various 5G services with wide range of requirements. Further, RAN moderation will imply dynamic radio topologies [4], e.g., activation/deactivation of nomadic access nodes (NNs) to attain on-demand network densification for coverage and capacity enhancement [5]. To this end, special focus of interference management will be on reducing the overhead and signaling required (e.g., with context aware mechanisms), while keeping energy efficiency high (e.g. by dynamically switching off small cells). II. ENVISIONED INTERFERENCE MANAGEMENT FRAMEWORK A. Overview of technologies As mentioned above, there are many challenges for a future 5G RAN with respect to interference avoidance and management and the coordination of multiple transmit and receive points. For example, new traffic types for machine-type communications (MTC) will generate short to very short bursts of data and will require novel concepts and mechanisms for the interference management. In addition, towards the dynamic small cell deployment, moving relays are assumed considering the possibility of implementing full-duplex operation using different frequency bands for the backhaul link and access link, for relaxing the scheduling burden for certain delay-sensitive services. In 5G RAN, one key RAN functionality framework, which is aimed to construct the agile resource management (RM) framework [3], is the Interference Management functional block. The agile RM framework aims at the efficient and effective mapping of resources to diverse services and applications in the complex 5G landscape, which includes different deployments, use cases and novel communication modes, e.g., various D2D communication modes. The highlights of the envisioned agile RM framework will also be discussed in the EuCNC presentation. Interference Management in this paper can be grouped in three different classes, given the different objectives they aim to solve. That is,  Enhancing cell-edge throughput by either using cooperative nomadic nodes in hotspot areas or by creating “interference-free” zones to enhance cell edge performance.  Enhancing energy efficiency by dynamically switching on/off small cells, while providing coordinated multi-