Scalable Signaling Support
Despite the huge success of cellular networks and the Long Term Evolution (LTE), there are challenges ahead. The performance requirements of LTE yield massive signaling traffic, imposing further demands on already strained networks. In an effort to manage the growing volume of data in cellular networks, there has been an increasing interest in deploying relays, distributed antennas, and smallcell access points (e.g., pico- and femtocells) in residential areas, subways and offices.
These network architectures – heterogeneous networks or HetNets – are being deployed to offload existing cells, and to increase the quality of experience for subscribers. However, the deployment of HetNets results in a significant increase of transport connections. Moreover, HetNets increase the overall signaling load on the network due to the signaling required to simply administer the small cells, as well as the additional load that results from frequent handovers in these environments. Also, recently, SDN has been introduced into the cellular network environment, something which will change the way signalling will be performed in the future.
The SCTP transport protocol plays a key role in current and future LTE cellular networks. SCTP is employed in the transportation of signaling traffic between the Radio Access Network (RAN) and the Evolved Packet Core (EPC), i.e., the signaling and data traffic aggregation network of LTE. It is also used within the IP Multimedia Subsystem (IMS), i.e., the subsystem within LTE responsible for multimedia services. As one way to handle an increasing signaling load in future LTE networks, we will investigate how to support low-latency communication over multiple paths in SCTP. There are currently ongoing efforts within the Internet Engineering Task Force (IETF) to standardize the concurrent multipath transfer extensions to SCTP proposed by Iyengar et al. Concurrent Multipath Transfer SCTP (CMT-SCTP). We intend to further this work by extending CMT-SCTP with novel solutions for low-latency communication. To avoid having interference between independent traffic flows, SCTP supports multiple, logical streams; messages that are lost in one stream do not affect the delivery of messages belonging to other streams. We will build on the stream concept, adding intelligent stream and path scheduling schemes in combination with appropriate congestion control for low-latency communication over CMT-SCTP.
To handle the increasing load from data and signaling in a cost-efficient and manageable way, there are several ongoing efforts to migrate EPC/IMS to a data center cloud. The current efforts in Network Function Virtualisation (NVF), and the integration with SDN enables much more flexible network deployments that are able to react during runtime on rapid scalability changes. This project aims to further this work by investigating cloud-based architectures for EPC/IMS. The implications and requirements on latency will be analyzed, and the performance of different transport services will be evaluated, including our CMT-SCTP solutions. In addition, the impact of SDN-based architectures (e.g., EPC) on the performance of different transport services will be analyzed, especially when virtualized network components are integrated with SDN-based approaches.
The increase in control signaling traffic is not only a problem in the core of LTE, but also in the LTE radio access parts – particularly in rapidly moving environments. When mobile terminals move frequently, location registrations or handovers are made, something which in turn generates a large volume of signaling traffic. As part of this project, we intend to study ways to optimize the signaling traffic management in these harsh environments.