Our most recent research findings produced artefacts and human resource development focused on Multimedia Delivery Architectures, Mobility Management, All-IP convergence, the Evolved Packet Core, Software-Defined Networks (SDN), Network Function Virtualization (NFV), Internet of Things (IoT), and Smart City Applications.
This experience places CRG in a strong position to investigate the future trends in internet research, namely the vision of highly performing networks to be deployed in the year 2020 and beyond. Largely branded as 5G, it will carry the increased traffic demands imposed by the huge number of always-on connections, when devices are expected to surpass tens of billions. 5G will likely drive the development of a new and improved radio interface, however this will not be the only improvement from 4G as an end-to-end system that streamlines the functionalities of the entire network needs to be designed. Due to the high performance requirements and expectations of the next generation mobile network architecture, the supporting infrastructure needs to allow for high flexibility, agility, energy efficiency, scalability and cost effectiveness especially as this architecture is likely to be the transport enabler for IoT devices and applications. The emerging paradigms of SDN, NFV, Cloud computing, All-IP, and supporting Application Programming Interfaces represent a first concrete step towards a new direction on the evolution of mobile network architectures.
An attempt has been made to split our main research areas into four sections. Moreover, these sections cut across various planes of a modern networking eco-system and are therefore not necessarily mutually exclusive.
CRG1: SDN, NFV and Cloud: Enablers for Mobile Network Architectures
The evolution of the mobile network architecture is an essential part of in the development process of 5G cellular mobile systems through the incorporation of advanced cloud techniques and network function virtualization. The emerging paradigms of Software Defined Networking (SDN) and Network Function Virtualization (NFV) is decoupling the software control plane from the hardware/software driven data plane and the virtualization of network functions on general purpose hardware. These architectural changes will permeate networks, ranging in size from the Internet core networks and enterprise networks, to wireless core networks and wireless radio access networks.
The design goals around 5G will be aimed at streamlining the performance and efficiency of the network to achieve optimal deployments. This will be seen as achieving greater throughput while lowering network latencies; enabling operations that are ultra-high reliable; higher connectivity densities; enhanced mobility features; and an overall enhanced performance. While performance is pushed to the extreme, efficiency is equally an important design aspect of 5G.
NFV, SDN and Cloud computing aims to transform the way that network operators design networks by evolving standard IT virtualization technology to consolidate many network equipment types onto industry standard high volume servers, switches and storage, which could be located in date center, network nodes and in the end user. It involves the implementation of network functions in software that can run on a range of industry standard server hardware, and that can be moved to, or instantiated in, various locations in the network as required, without the need for installation of new equipment.
Network operators’ networks are populated with a large and increasing variety of proprietary hardware appliances. To launch a new network service often requires yet another variety and finding the space and power to accommodate these boxes is becoming increasingly difficult; compounded by the increasing costs of energy, capital investment challenges and the rarity of skills necessary to design, integrate and operate increasingly complex hardware-based appliances. Moreover, hardware-based appliances rapidly reach end of life, requiring much of the procure-design-integrate-deploy cycle to be repeated with little or no revenue benefit. Worse, hardware lifecycles are becoming shorter as technology and services innovation accelerates, inhibiting the roll out of new revenue earning network services and constraining innovation in an increasingly network-centric connected world. We believe Network Functions Virtualization is applicable to any data plane packet processing and control plane function in fixed and mobile network infrastructures. Some of the main ideas to be investigated in this arena include:
- Infrastructure sharing amounts network operators to reduce CAPEX and OPEX costs
- Sophisticated service orchestration and mobility management to mobile end users in the evolving framework
- Mobility management issues in future networks and resource allocation in a network-shared NFV environment.
- Security of network functions operating in cloud computing environments
- Fog and Mobile Edge computing as cloud computing services are moved closer to the end user to reduce latencies as required for 5G networks
- Carrier grade resiliency and high availability of network functions deployed in NFV/SDN enabled clouds.
CRG2: Efficient Radio Resource Management in Next Generation Networks & Cognitive Networks
The exponential growth of data traffic resulting from proliferation of smart devices and mobile Internet has triggered the investigation of next generation network, which is expected to provide higher peak and user data rates, reduced latency, enhanced indoor coverage, and improved energy efficiency, when compared with the 4G network. Thus the next generation network is expected to support diverse set of devices, services, applications, and users with different requirements. In order to meet the requirements for the next generation wireless network, it is important to develop efficient radio resource management algorithms. The research on efficient radio resource management focuses on development of radio resource management algorithms and mobility management schemes for QoS provisioning, efficient resource allocation, and cost effectiveness in next generation wireless networks.
An important aspect of radio resource management to be addressed in this research is multipath packet transmission. Multipath packet transmission techniques offer network providers unique opportunity to scale up network capacity in order to serve emerging ‘bandwidth-hungry’ applications. Existing end user terminals are equipped with multiple interfaces to connect to more than one access technology simultaneously. The terminals can therefore be enable to transmit and receive data through multiple interfaces at the same time. The benefits that can be ripped by allowing IP traffic from the same application to be striped across multiple network paths include increased throughput and shorter delivery time. Error resilience can also be accomplished by duplicating IP packets and sending them through multiple paths concurrently. Moreover, the network load can be conveniently routed through the least loaded interfaces, thus enabling the network provider to balance traffic load efficiently.
Research on radio resource management also focuses on development fast and efficient vertical handover (VHO) schemes for communicating mobile nodes in heterogeneous wireless networks. This aspect deals with the selection of suitable access networks to ensure optimal quality of service provisioning for applications demanded by communicating mobile nodes. It also covers development of efficient spectrum handoff schemes for cognitive radio networks, and dynamic spectrum reallocation via autonomic management in next generation wireless networks. As more devices are connected via the mobile wireless networks, there is need for dynamic spectrum re-allocation, which allows changes to be implemented in allocated resources as the network environment changes. It also covers dynamic pricing for congestion control during the peak period is in next generation wireless network.
Another aspect of the research on radio resource management in next generation wireless network is interference management in small-cells, and development of effective backhaul solutions for densely deployed small cells in urban environments. There is need for new cost effective backhaul solutions that address the capacity and coverage problem in next generation wireless networks.
Research on Novel Approaches to Performance Evaluation and Benchmarking for Energy-Efficient Multicasting Over Coded Packet Wireless Networks is ongoing as energy efficient multicasting in wireless networks is becoming increasingly important in the field of information communication and technology. The core part of this research is based on performance measurement and benchmarking problems. We study existing methods of minimizing energy in wireless multicast networks and propose novel approaches that are based on Data Envelopment Analysis (DEA) method to further optimize energy consumption in wireless multicast networks.
In SDR we seek to customize the Ettus GNU Radio and similar platforms in order to perform physical laboratory investigations into bringing together concepts learnt in the disciplines of electromagnetics, RF and microwave engineering, digital signal processing, embedded systems, computer programming, and systems engineering to solve versatile communication requirements.
Recent research trends have identified Cognitive Radio (CR) has a viable solution to the challenge of spectrum scarcity. CR is a novel technology which improves spectrum utilization by seeking and opportunistically utilizing unused radio spectrum (White Space) in time, frequency and space domain on a real time basis. However, to achieve effective utilization of spectrum, CR must have the ability to sense its environment, perform channel state prediction, make decisions on channel allocation and select optimal transmission parameters. This research work intends to focus on the decision module in CR. This work will seek out the use of quick-to-converge optimization techniques to enhance the decision module of CR.
In addition a Cross Layer Approach to Antenna Selection in Cognitive Radio MIMO Systems is being investigated. When using antenna selection, the common practice has been to select the subset of transmit/receive antennas that maximise criteria at the physical layer. These algorithms do not exploit all the information available in specific system scenarios concerning the schemes and algorithms being used in upper layers. Within this framework therefore we propose the development of transmit AS algorithms in CR MIMO systems from a cross layer perspective. More precisely, we propose to focus on the interaction between the physical and the data link layers to study and design adaptive TAS algorithms aimed at maximising throughput at the data link layer while minimising interference at the primary user receiver.
Improving cognitive radio dynamic and spectral efficiency by using game theory enhanced distributed cooperative sensing (Spectral efficiency in CRN) is being studied.
The need for the next generation of telecommunications that include converged and omnipresent networks brings about many challenges for wireless networks both in the backbone as well as in the access network for Telecommunications companies. One of the main problems is spectrum limitation, and Cognitive Radio (CR) technology is a proposed solution to the challenge imposed by legacy static spectrum allocation. The challenge of developing network protocols, standards and technologies that should be used for the successful implementation of Cognitive Radio Networks (CRNs) is of interest to researchers.
Energy Efficient Interference Management in Heterogeneous Networks (HetNets) which incorporates macro cells, pico-cells, femto-cells and Wi-Fi are seen as the future of wireless communication considering the increasing capacity demand. Developing efficient ways to coordinate and manage this interference is a key issue if operators are to remain competitive and offer good user experience to subscribers.
CRG3: Future Internet Research and Experimentation – IoT/M2M
With the Internet of Things (IoT) standards undergoing major evolutions, cellular network standards are now adding techniques to improve network performance to address traffic patterns generated by an ever increasing number of IoT devices. Another reason behind this shift is related to the IoT service platforms as standardized by oneM2M and IoT applications.
This is continued research to improve Future Internet Research Experimentation testbed capabilities in Europe and in South Africa, demonstrated through experimental research linking smart and green technological and social innovation, essential to address economic, social, and environmental challenges due to the increase in urbanization, requiring informed decisions based on Internet of Things generated data. A particular challenge is the unstable power supply of cities in developing countries (e.g. South Africa), thus requiring smart energy management. Future handling of grid overload in South Africa involves demand-response mechanisms, installing small devices at the end-user, communicating over different network technologies to a central controller, allowing loads to be measured and limited if necessary. Further challenges include the deployment of affordable smart sensors (e.g. air sensors) as well as gathering information from nodes with limited power access.
In scenarios from energy consumption to waste bin levels, data is either sent over IP networks (which delivers data immediately) or collected in a delay tolerant mode by mobile devices of individuals or crowds.
In delay tolerant mode the data is stored locally, to be delivered when a suitable network is reached. In cases of open data collection the devices in this Future Internet realm are targets of security attacks and might be vendor-locked with proprietary software stacks.
Our approach to address these issues is to interweave sophisticated Smart City platforms and an ETSI/oneM2M compliant Machine-to-Machine (M2M) communication framework (TUB/FOKUS OpenMTC). We emphasize secure identification and authentication of sensors and users as well as store and forward functionality.
The integration of several frameworks and associated applications involve two testbeds for smart cities M2M communication: one at the Technical University of Berlin (TUB), Germany, and another at the University of Cape Town (UCT), South Africa.
At University of Cape Town, South Africa:
- Experimental Machine to Machine testbed, based on OpenStack, a cloud computing software and on the FOKUS OpenMTC middleware framework.
- Connected via VPN to the TUB testbed and integrated in the control flow of FITeagle and OpenSDNCore
- Hosts devices and gateways for aggregating and exchanging data.
- Used for prototyping IoT applications in the areas of Smart Home, Smart Energy, eHealth
- Expands on developed applications by making use of OpenStack virtualization to
investigate scalability-driven evaluations
- Allows for resources to be used as “Education Experiments” for students with access to the testbeds.
The outcomes encourage the development of affordable technologies for future Internet, research activities on delay tolerant networks and opportunistic communications, as well as developments supporting innovative applications for social integration, improving the capabilities of testbeds on Future Internet technologies in Europe and in South Africa.
The development of the technology (testbeds and platforms) can be used as best practice to guide the research, development and implementation of similar technologies for application in South Africa and Worldwide.
CRG4: Smart Cities
By the year 2050, it is estimated that more than two thirds of the global population will inhabit cities. It is expected that most urban growth will occur in less developed countries during the next decades.
Many cities are currently developing strategies towards becoming “smarter cities”. In Europe, several cities challenge to compete to be the best models of sustainable urban development. Similar initiatives are going on in Africa to develop sustainable energy service, as well as other smarter services motivated by the emergence of many rapidly growing economies.
Mobile connectivity has transformed daily life across the globe, enabling the potential improvement in various services. A recent survey by Analysis Mason found that 87% of the respondents across Africa indicated that mobile devices were the main means through which they connected to the Internet. On average mobile subscription penetration has reached 72% across Africa with varying country penetration rates and is expected to continue growing until 2017 at which point Africa will have an estimated subscription penetration of 97%.
The need of cost efficient, secure and user-friendly solutions is common within African and European cities. However, in several African countries, urban expansion has often been characterized by informality and unplanned settlements. Those countries also often have to face the increase in slum dwellers with a challenge in optimizing resource management and enhancing different kind of services.
The lack of reliable networking infrastructure in some areas is one of the main challenges that could prevent the usefulness of Smart City solutions and likely common around the global. Research work is on-going to resolve such problem of unreliable connectivity using various approaches. However, the need to understand the special requirements, limitations of existing system/infrastructure and available opportunities in each city is important to adapt the technology and develop the optimum, sustainable and efficient solution. The environment conditions such as humidity, temperature, salinity, wind, and sunlight levels are important factors that could influence the implementation of related systems.
In Africa, the municipality’s limited resources add more challenges to the administration and management process. Therefore, more than anywhere any Smart City solution must encourage the inclusion of inhabitants in order to insure the sustainability of the system. For example, in many African cities regular public transportation tends to rely on small collective and uncontrolled means of transportation (i.e. vans or small buses). The implementation of a smart transportation service to improve the exiting service requires the cooperation of buses owner/drivers and end-users engagement. Lessons to be learned from the implementation in one area shall be considered, as slimier problems are likely to be faced globally.
A citizens-based smart city model could be more efficient than a municipality-centric smart city model, therefore the designed solution need to add value upon existing services for the benefits of all citizens and stakeholders.