The EMBRACE project addresses the topic of designing efficient solutions for 5G networks taking into account human behavior, uncertainty, and heterogeneity of networking resources.
The divergence of mobile data traffic and data revenue growth is increasing the pressure on mobile network operators to find effective and efficient ways to handle the data traffic that is swamping their networks. Associated to this pressure, the planning and operation of mobile networks have been traditionally based on oversimplified models of demand. Thus, there is the concern that such demand models are not able to resist to the natural changes on the user behaviors along the time. The associated consequence is the development of models that underestimate the opportunities for cost minimization or maximization of the user satisfaction. In addition, resource allocation models and services orchestration models designed for the planning and operation of mobile networks have also made optimistic assumptions about the resource availability and synergies between the services, e.g., by assuming that the resources and operational services have low uncertainty or a simple average can properly describe them. While this could be a reasonable approach in the traditional mobile networks, it is inappropriate in the context of 5G networks.
All these factors are putting the human element in the foreground, making essential (1) to center the big data analytics around individuals and (2) to bring human’s behavior awareness to the design of resource management solutions, while coping with environmental uncertainties, the rapidly growing capacity requirements, and new pervasive demands.
The purpose of the EMBRACE Associated Team is to develop a coherent and feasible framework that will advance over the standard unrealistic resource allocation approach in a number of ways:
- First, the EMBRACE framework will highly rely on user context information. Context here includes information about the habits of a mobile user as well as about its current state, in terms of positioning data, content demand he/she generates, and wireless interactions. Moreover, the solutions defined in the project will make use of both individual context data and aggregate community context data.
- Second, the EMBRACE framework will properly represent uncertainty by employing more realistic assumptions on the resource availability. This will be achieved by input models that describe the user behavior and stochastic optimization models that deals with uncertainty.
- Third, the EMBRACE framework will be designed so as to comply with the existing mobile network heterogeneity, and consequently, to support realistic network requirements from the user side in terms of resource demands and realistic resource availability from the operator network side.
The original framework developed in the project is expected to significantly improve the utilization of resources and synergies between networking services at the core and edge of mobile wireless network, and also end-user device. The novelty brought by the EMBRACE framework lies in the fact that designed algorithms will leverage knowledge resulted from users behaviour, network heterogeneity, and uncertainty. The evaluation by simulation and practical experimentation will also be one of the key themes of EMBRACE.