End: 31/12/2017
Funding: National
Status: Completed
Mobile Networks (MONET)
Acronym: 5GNORM
Call ID: TEC2014-60491-R
Code: TEC2014-60491-R
The forecasted traffic volumes in 5G networks along with declining ARPUs mandate for a fundamental rethinking of network architecture and deployment towards more efficient management and operation.
Therefore, the move from 4G to 5G will not only impact the radio access network (e.g., by improving its link rate), but it will also impact the way the whole network is managed and operated. This is why 5G research brings a fresh and global view to network architecture reasoning in two main directions. First, it triggers multiple research efforts on the automated management of network functions that have traditionally been manually operated.
In this respect, let us highlight the work undertaken by 3GPP since Release 10 on self-organizing network functions and their coordination. And second, it has recently triggered a number of efforts not only on improving the operation logic and performance of the network functions per se (i.e., their internals), but also on building frameworks for virtualizing networks and functions to offer an unprecedented degree of network programmability (e.g., to quickly scale up/down the network without deploying new hardware). Software Defined Networking (SDN) introducing control and data plane decomposition, and Network Functions Virtualization (NFV) proposing the operation of network functions as software instances fall under this latter umbrella.
In this context, project 5G Network ORchestration and Managament (5GNORM) aims at moving traditional networks (e.g., cellular, IP-based) to the next level of efficiency in terms of network resource consumption and sharing by: 1) proposing an architecture of dynamically managed and operated network functions that integrates and exploits the best of SDN and NFV, and 2) demonstrating the operation of such architectural framework for specific virtual functions (e.g., virtual base stations, virtual self-organized networking functions).
These architectural framework and functions will be evaluated in two representative network sharing scenarios: 1) optimization of the resources of an infrastructure provider with multiple tenants, and 2) optimization of the network resources under the exclusive control of one of these tenants.
Moreover, 5GNORM aims to assign a starring role to experimental research and validation of the theoretical concepts under study. It is for this reason that 5GNORM will build on top of two of the tools developed by the group to generate the required framework for evaluating the 5GNORM network architecture. This will include: 1) the addition of the required components (hardware and software) to the EXTREME Testbed to build a flexible SDN/NFV experimental research framework, based on, for instance, OpenStack cloud and OpenDaylight controller platforms, and 2) the extension of the ns-3 LENA simulator/emulator, with particular emphasis on improving its performance in emulated mode, which will enable cross-validation of simulation and experimental results over the testbed by using the same code.
We believe that the concepts under study in 5GNORM, as well as its experimental framework will help in achieving high impact in the framework of future industrial contracts or EU projects, and also in academia, through the publication of our research work in renowned conferences and journals. For the same reason, we also believe, 5GNORM will become a remarkable training framework on 5G networks not only for PhD students, but also for the team as a whole.
Coordinator