A testbed for flexible and energy-efficient resource management with virtualized LTE-A nodes

Abstract

This paper describes the software architecture and the implementation of a fully operational testbed that demonstrates the benefits of flexible, dynamic resource allocation with virtualized LTE-A nodes. The testbed embodies and specializes the general software architecture devised within the Flex5Gware EU project, and focuses on two intelligent programs: the first one is a Global Scheduler, that coordinates radio resource allocation among interfering nodes; the second one is a Global Power Manager, which switches on/off nodes based on their expected and measured load over a period of minutes. The software framework is written using open-source software, and includes fast, scalable optimization algorithms at both components. Moreover, it supports virtualized BaseBand Units, implemented using OpenAir-Interface, that can run on physical and virtual machines. We present the results obtained via on-field measurements, that demonstrate the feasibility and benefits of our approach.

Practical feasibility, scalability and effectiveness of coordinated scheduling algorithms in cellular networks towards 5G

“…Hereafter, we describe several solutions, which strike different tradeoffs between the two. We refer the interested reader to [14] and [11] for more details on Flex5Gware's software framework.…”

“…Last, but not least, it is worth noting that our dynamic CS framework is fully compliant with the current 4G standard. In fact, it has been implemented and demonstrated in a live prototype of an LTE cellular network [13]- [14].…”

Minimizing Power Consumption in Virtualized Cellular Networks

“…The GPM contacts an LPM and instructs it to switch on/off the related Broadband Unit (BBU) and Remote Radio Head (RRH). The general architecture is described in [14]. Hereafter, we only recall aspects that are related to the GPM operation.…”

“…The intelligent program will then include the new nodes in their optimization cycles starting from their respective next period. In the testbed, the BBUs have been realized using OpenAirInterface [13], and RRHs are Ettus boards [14].…”

“…Live tests have been performed on the testbed to show the functionalities of the software modules, assess the communication overhead (which is small) and latency (which is tolerable, and mostly due to the OAI software). These tests are documented in [14], to which we refer the interested reader.…”