Presentation on theme: "Chapter X: Towards Energy Efficiency in Next Generation Green Mobile Networks: a Queueing Theory Approach Glaucio H. S. Carvalho 1, Isaac Woungang 2 and."— Presentation transcript:
Chapter X: Towards Energy Efficiency in Next Generation Green Mobile Networks: a Queueing Theory Approach Glaucio H. S. Carvalho 1, Isaac Woungang 2 and Alagan Anpalagan 3 1 Faculty of Computation, Federal University of Pará, Brazil 2 Dept. of Computer Science, Ryerson University, Canada 3 Dept. of Electrical Engineering, Ryerson University, Canada HANDBOOK ON GREEN INFORMATION AND COMMUNICATION SYSTEMS
Outline Introduction Heterogeneous Wireless Networks (HetNets) Diverse Types of HetNets Radio Resource Management (RRM) RRM in HetNets Queueing theory Green Approaches in HetNets Green JRRM in HetNets Conclusion
Introduction The astronomical growth in the demand for wireless access has urgently required from Mobile Network Operators (MNOs) an expansion of their access networks; To achieve this objective, MNOs have started a dense deployment of new Radio Base Stations (RBSs); Concern ICT sector contributes with 2% of the GHC emissions per annum; The major responsible for this environmental threat is the huge amount of energy demanded to keep the sector in operation; MNOs represent 0.2% of the total energy consumed and with the required expansion, it is foreseen that it grows even more.
Introduction (contd.) Understanding why the combination between energy consumption and network expansion is a real concern. SourcePercentage % Radio Base Station (RBS)57 Mobile Network Exchange (MTX)20 Core network15 Data Center6 Retail2 Sources: Z. Hasan et al., Green Cellular Networks: A Survey, Some Research Issues and Challenges, IEEE Communications Surveys & Tutorials, vol.13, no.4, pp , Fourth Quarter T. Chen et al., Network Energy Saving Technologies for Green Wireless Access Networks, IEEE Wireless Communications, vol.18, no.5, pp , October 2011.
Introduction (contd.) BTS drains a significant portion of the energy consumed on network; BTS is the lead actor in the network expansion; With the expansion, MNOs run the risk of seeing their electricity bills biting a huge amount of their profits; Green network design is mandatory! Benefits of adopting it: Contribute to the world preservation; Economical reasons.
Introduction (contd.) Some initiatives world-wide. Energy Aware Radio and Network Technologies (EARTH); https://www.ict-earth.eu/ Towards Real Energy Efficiency Network Design (TREND); Cognitive Radio and Cooperative Strategies for Power Saving in Multi-Standard Wireless Devices (C2Power) ; Optimizing Power Efficiency in Mobile Radio Networks (OPERA-NET);
Introduction (contd.) Some promising results Development of new power amplifiers (PAs), based on Doherty-architectures and Aluminium Gallium nitrite (GaN), reduces the energy consumption in BTS; PAs consume the major portion of the BTS, so turning them off in BTS radio transceivers is an interesting solution. Example: LTE RBSs are able to turn PAs off during idle times by employing power saving protocols as discontinuous reception (DRX) and transmission (DTX);
Introduction (contd.) Some promising results Employment of eco-friendly green energy sources is a viable alternative to reduce deployment costs; In many sites the payback period for green power technology is less than three years; Cognitive and cooperative wireless technologies are now in embryonated stage; The inclusion of green design in their development agenda can result in significant gains in terms of power consumption reduction at medium to long term.
Introduction (contd.) Present moment Huge deployment of BTSs, From different manufactures, belonging to different owners, with different technologies, making part of different design strategies covering almost all parts of the globe; The coexistence of these networks defines what is called heterogeneous wireless networks (HetNets); To achieve energy efficiency in HetNets environment, MNOs require an immediate solution.
Introduction (contd.) Present moment A step in this direction is to dynamically manage the coverage according to traffic load fluctuations; Our approach is to switch off BTSs in such a way that their power consumption could be translated into energy saving! Threshold based Green Joint Radio Resource Management Apply a dynamic coverage management (DCM) algorithm to decide, based on the network load condition the value of the threshold, when some cells can be turned off or on; ;We use the framework of queueing theory for performance evaluation;
Heterogeneous Wireless Networks (HetNets) HetNets are formed by multiple heterogeneous wireless networks; Each one has its own characteristic such as: capacity, access technology, security, power consumption, delay, coverage, access cost; Interesting feature: Some wireless access networks are overlaid by others in such a way that a multi-layer structure is naturally built.
Heterogeneous Wireless Networks (HetNets) Overlay networks can be suitably explored to match multiple design: Boost the system capacity; Increase system coverage; Enhance user satisfaction; Offer different access costs for end-user;
Diverse Types of HetNets A taxonomy based on the coverage area Wireless Wide Area Network (WWAN):cellular mobile networks 2G: GSM and TDMA/ G: GPRS 2.75G: EDGE (EGPRS) 3G: WCDMA, CDMA2000 High data rates as well as mobility. Wireless Metropolitan Area Networks (WWAN): IEEE standard. WiMaX; Up to 70Mbps in the range of 50km; Support to QoS.
Diverse Types of HetNets A taxonomy based on the coverage area Wireless Local Area Network (WLAN) : standard Low cost and operation in an unlicensed frequency band; Higher data rates ( Mbps) within a short coverage (100m) Found in places like coffee shops, airports, offices, etc e standard came to support real time services Wireless Rural Area Network (WRAN): IEEE standard The first cognitive radio standard. Use TV broadcast bands by means of cognitive radio technology as long as they do not cause harmful interference to incumbent TV receivers to deliver broadband services;
Diverse Types of HetNets A taxonomy based on the coverage area Personal Wireless Area Networks (WPANs): standard Piconets; Shorter coverage area of until 10m; Supporting up to eight devices; Possibility to rise the delivery of medical services to another level as, for instance, anytime and anywhere care;
Diverse Types of HetNets A taxonomy based on the coverage area New Technologies: cognitive and cooperative Cognitive and Dynamic Spectrum Access (DSA): –Allow unlicensed or secondary users (SU) equipped with cognitive radio to use idle frequency bands in the licensed spectrum as long as they do not harmfully interfere with the licensed users or primary users (PU) operation; Cooperative network: –Use relay nodes to offer alternatives paths from BTS till mobile users. –Improvement in the system performance in many dimensions: Reliability, coverage, spatial frequency reuse;
Radio Resource Management (RRM) RRM in wireless networks is more challenge because: Radio resources are extremely expensive and scarce; The actual offered traffic load is huge; The QoS requirements of the applications are quite diverse and often conflicting; To cope with all of these objectives, RRM uses traffic management mechanisms: Call admission control (CAC) ; Scheduling;
Radio Resource Management (RRM) Call admission control (CAC) Decides whether of not an incoming service request is accepted; Takes into account the network load and the QoS profile of the incoming service request; Generally designed to handle calls at connection level, but it must also cope with packet level;
Radio Resource Management (RRM) Scheduling Schedules which data packet from which data session will be transmitted; Examples: FIFO –data packet that comes first will be first served; Priority queueing –Data packets are stored in different priority buffers. The scheduler schedules the traffic classes based on the occupancy of the higher priority buffers; Weighted Fair Queueing (WFQ) –Provides fairness among data flows by using weights to prevent monopolization of the bandwidth by some flows;
RRM in HetNets In overlay homogeneous cellular mobile networks, each layer operates independently of the others; RRM algorithms have to handle the incoming service requests offered in each layer; New challenges in HetNets: The variations on traffic load in space and time can result in underutilization and overutilization of radio resources Some cell can experience blocking while others are quite idle;
RRM in HetNets Solution Joint RRM (JRRM) or Common RRM (CRRM) ; Advantages: Whole vision of all layers (and cells inside them) Cope with the following tasks: –Deciding whether an incoming service request should be accepted or blocked (like homogeneous networks); – Selecting in which of the cells, an incoming service request has to be accommodated.
RRM in HetNets JRRM functionalities can be split into three procedures: Resource monitoring: Acquiring information about users preferences (cost, QoS requirement, wireless technology, etc.) and networks state; Decision making: Most important step; Types of decision: Network selection and bandwidth allocation; User centric-approach, network-centric, or both; Techniques to support decision: Markov Decision Process (MDP), Fuzzy logic, game theory, TOPSIS,etc. Decision enforcement: Execution of decision;
Queueing theory Mathematical branch of computer science and electrical engineering used for building performance models; Supports the traffic theory; Traffic models: Arrival process: Poisson process, Markov Modulated Poisson Process (MMPP), Interrupted Poisson Process (IPP); Service process: Channel holding time, Th=1/ h; Exponentially distributed and given by the combination of two exponential distribution: Call time duration (Td=1/ d) and stay time (Ts=1/ s);
Queueing theory Specification of system resources: Number of radio resource channels Buffer System capacity is translated as effective or equivalent bandwidth, regardless of the multiple access technology.
Green Approaches in HetNets Overlay structure presents an immense potential to reduce the energy usage by applying the dynamic coverage management (DCM) algorithms; With DCM, the JRRM scheme can determine when and which cells inside the covered geographical area can be turned off or on; DCM updates the role of the JRRM to not only balance the traffic load to achieve a suitable QoS, but also save energy
Green Approaches in HetNets Application of DCM algorithms is based on the traffic load fluctuations in time and space; Uses the combinations of cells with different sizes and capacities to cope with these variations; Challenge on HetNets Find the optimal or at least a good combination; Design using only macrocells: Copes with mobility but is inefficient due to the high data rates required for actual data services; Design using only smaller cells: Deals with capacity, but can suffer due to user mobility; Good strategy Combination of macrocells, microcells, picocells, etc. Switch off some cells if their traffic load is low; From energy efficiency perspective, we are still crept!
Green Approaches in HetNets Standalone approach Power consumption reduction that is the network selection: Goal: –Choose the network that best balance performance and energy consumption; –Energy reduction due to network scanning; Self-Optimization, Self-Organization, Self-Healing (SON) algorithms: Son has been employed for handoff optimization, antenna tilt optimization, interference minimization; Can contribute with power consumption reduction if it comes to be used as a goal. For example: –Optimize the Frequency Reuse to dynamically adapt it to match the spatio-temporal variations of the offered traffic;
Green Approaches in HetNets Cross-layer The problem of energy consumption spreads over all layers of protocol stack, Cross layer design can be applied to match this objective; Promising line of research Example: Rethinking MAC schemes to cope with energy savings; –Determining the optimal size of a variable length TDMA, which in turn leads to the determination of the optimal constellation size; Devices can prolong their sleep mode as long as possible in order to reduce power consumption; –Mobile users should determine when wake up to receive or transmit data packets;
Green JRRM in HetNets Proposal Threshold based Green JRRM in HetNets. Explore the differences in coverage areas in overlay wireless networks to balance performance and energy consumption. The core idea behind the scheme is the DCM algorithm.
Green JRRM in HetNets Architecture of the proposed green JRRM DCM Load control Monitors an informs the network load; Radio Channels Macrocell (CM) and microcell (Cm) Threshold in macrocell (K);
Green JRRM in HetNets Proposed DCM algorithm Characteristics: Assume macrocell always on; Macrocell planned to cover the entire geographical area; Hold the offered traffic load on macrocell whenever possible; Thus the microcell can be kept off saving energy;
Green JRRM in HetNets Queueing Model and Analysis 2D Continuous Time Markov Chain (CTMC) model whose state is: Small scale system
Green JRRM in HetNets Queueing Model and Analysis Performance measures: Blocking probabilities and utilization in macrocell and microcell Blocking probability Quantifies the chance of an incoming call makes an access request and finds all radio resources busy
Green JRRM in HetNets Queueing Model and Analysis Performance measures: Blocking probabilities and utilization in macrocell and microcell Utilization Ratio between the occupied radio resources and the total available radio resources
Green JRRM in HetNets Queueing Model and Analysis How to measure greenness; The definition of green metrics is an open discussion in industry and academia Green opportunity All the changes that a MNO has to reduce the energy consumption. In our model it is given by the probability of finding the microcell off.
Green JRRM in HetNets Applications Dense deployment of femtocells is a hot research topic in wireless community; Challenges Deal with the high handoff rates due to the small size of cells; An efficient radio resource management strategy must be designed to support a fast and real time call switching among cells in order to keep the service continuity and the QoS provisioning; SON algorithms might tackle such situation and aid to reduce latency and network overhead in future; To cope with mobility, we envision apply our threshold based green JRRM in current scenario of HetNets; Possible applications: –Users with high mobility profile;
Conclusion We discussed a threshold based green JRRM for HetNets Use queueing model to perform different analysis For instance: How is the Green JRRM affected by the choice of the threshold value? How much is the proposed JRRM green? Huge research potential to develop green HetNets that can also provide QoS and mobility support in addition to high data rates.