Being cheap and easy to deploy, dense WLANs are becoming the most popular solution to providing Internet access in locations where the population of users is large, such as on campuses, large enterprises, etc. The large density of access points (APs) comes from the need to have enough capacity to carry the traffic generated at peak hours although, in these scenarios, traffic varies a lot on a daily, weekly, or seasonal basis. During low or no traffic periods, APs are underutilized, even if they are consuming energy almost in the same amount as if they were fully loaded. Promising solutions to reducing this form of energy waste consist of activating only the number of APs that is strictly needed to carry the actual traffic; in other words, to make capacity dynamically adaptive through resource-on-demand (RoD) strategies. In this paper, we investigate the case of a portion of the dense WLAN on our campus. Through real trace analysis, we investigate users' behavior in accessing the WLAN and formulate a stochastic characterization of it. We propose a simple model that describes RoD strategies and use it to study the system performance that is evaluated in terms of AP activity and inactivity periods, AP switching frequency, and energy saving. Finally, we present some results obtained by experimenting with RoD strategies in a portion of the WLAN. Our results show that RoD strategies for dense WLANs are feasible and effective in trading-off the opposite needs to save some energy and to guarantee a smooth network operation and high quality of service.

The power consumption in chips, in general, and in crossbars switching fabrics, in particular, grows with the maximum sustainable throughput. Due to the fast increasing traffic demands, the performance scalability of crossbars is severely limited by the capability of cooling the hardware devices. Hence, reducing the power consumption is an important design question to improve the crossbar switching performance. We propose to leverage Dynamic Voltage and Frequency Scaling (DVFS) hardware technique for the switching fabric. The main idea is to exploit temporary underloaded conditions to decrease the crossbar transmission rate while preserving maximum throughput. Differently from previous works, we consider a scenario in which the arrival rates are unknown in advance. Our proposed architecture is based on a power controller which runs periodically and independently of the packet scheduler, and whose decisions are based on the real time estimation of the arrival rates. We discuss the performance tradeoff in terms of throughput, delays and power, and show the relevant performance gain due to the use of DVFS in controlling the crossbar.

We consider an N × N Input-Queued (IQ) switch with a crossbar-based switching fabric implemented on a single chip. The power consumption produced by the crossbar chip, due to the data transfer, grows as NR3, where R is the maximum bit rate. Thus, at increasing bit rate, power dissipation is becoming more and more challenging, limiting the crossbar scalability for high-performance switches. We propose to exploit Dynamic Voltage and Frequency Scaling (DVFS) techniques to control packet transmissions through each crosspoint of the switching fabric. Our power control operates independently of the packet scheduler and exploits the knowledge of a traffic matrix obtained by online measurements. We propose a family of control algorithms to reduce the power consumption. The algorithms are particularly efficient in nonoverloaded conditions. The actual potential of the proposed approach is also evaluated on a real design case synthesized on a 90 nm CMOS technology.

We consider an N×N input-queued switch based on a crossbar switching fabric implemented on a single chip. The thermal power produced by the crossbar chip grows as N R3, where R is the maximum bit rate. Power dissipation is becoming more and more challenging, limiting the crossbar scalability for high performance switches. We propose to exploit Dynamic Voltage and Frequency Scaling (DVFS) techniques, quite commonly used in integrated circuit design, to control packet transmissions through each crosspoint of the switching fabric. Our thermal control operates independently of the packet scheduler and it is based on short-term traffic measurements. We propose a family of control algorithms to reduce the thermal power dissipation in non-overloaded conditions.

We consider a slotted input-queued switch with a crossbar-like switching fabric. In each time-slot, a centralized scheduler determines a switching fabric configuration to transfer packets. We consider the energy consumption needed to configure the switching fabric and we assume that the energy depends on the number of modifications in the switching configuration in two consecutive time-slots. We address the problem of scheduling a set of packets to minimize the required energy while preserving high throughput. We reduce the overall problem to the combination of two different optimization problems. We propose a family of algorithms to solve the problem and we discuss their energy-throughput performance.