Hypervisors are an essential part of our computing infrastructure, yet ensuring their correctness remains a significant challenge for the community. While several hypervisors have been formally verified using traditional methods, they have typically required a huge effort and significant input from verification experts. With the increasing diversity of hypervisors, driven by open hardware and custom ISAs, there is a growing need for more accessible approaches that can be used by non-experts.

While Function as a Service (FaaS) platforms can initialize function sandboxes on worker nodes in 10-100s of milliseconds, the latency to schedule functions in real FaaS clusters can be orders of magnitude higher. The current approach of building FaaS cluster managers on top of legacy orchestration systems (e.g., Kubernetes) leads to high scheduling delays when clusters experience high sandbox churn, which is common for FaaS. Generic cluster managers use many hierarchical abstractions and internal components to manage and reconcile cluster state with frequent persistent updates. This becomes a bottleneck for FaaS since the cluster state frequently changes as sandboxes are created on the critical path of requests. Based on our root cause analysis of performance issues in existing FaaS cluster managers, we propose Dirigent, a clean-slate system architecture for FaaS orchestration with three key principles. First, Dirigent optimizes internal cluster manager abstractions to simplify state management. Second, it eliminates persistent state updates on the critical path of function invocations, leveraging the fact that FaaS abstracts sandbox locations from users to relax exact state reconstruction guarantees. Finally, Dirigent runs monolithic control and data planes to minimize internal communication overheads and maximize throughput. We compare Dirigent to state-of-the-art FaaS platforms and show that Dirigent reduces 99th percentile per-function scheduling latency for a production workload by 2.79× compared to AWS Lambda. Dirigent can spin up 2500 sandboxes per second at low latency, which is 1250× more than Knative.

Piezoelectric actuators are used successfully to enable locomotion of micro-robot such as micro air vehicle with flapping wings. However, the use of piezoelectric actuators presents a major challenge for power electronic design: the input capacitance of such devices makes their use delicate because the instantaneous power may be much greater than the average effective power. Due to this challenge, conventional drive circuits, especially inductor, become too bulky or inefficient in low mass applications. This work describes a converter suitable to drive this kind of actuators with low inductor. Moreover the converter provides for recovery of the energy stored on the actuator capacitance back to the primary power supply when the actuator is not supply. The proposed converter is a DC/AC structure, which is capable of producing a square unipolar AC voltage. In our design, an auxiliary shunt circuits are connected to the actuator. Signal flow graph modeling, analysis and design of such a scheme is presented. Experimental validations have demonstrated the effectiveness of the proposed technique for recovery energy.

Powering from ambient energies various electric accessories on bicycles like lamps, on-board computer, communication devices, or even wireless sensor nodes will avoid the use of batteries and improves the availability of these devices when the bicycle is not used for a long time. This paper reports on vibration resources detected in a bicycle as a potential energy source for supplying these on-board devices. The measurements showed that the bandwidth of the energy vibration is reduced with speed. For this application, a piezoelectric vibration harvester was designed and equipped with a self-powered voltage switching interface circuit. Within the field test, it is demonstrated that sufficient energy is harvested.

In this paper, a piezoelectric energy harvesting device with bistable non-linear vibration combined with a switching-type interface (SSHI) are compared to the traditional linear technique. The voltage switching techniques are used to increase the output power of energy harvesting devices for low-coupled structures. However, these devices are efficient only at the resonance. The bistable non-linear technique can broaden the available bandwidth. We used the power flow and work-cycles to simplify the analysis of the non-linear vibration with switching techniques. Numerical simulations of the equivalent circuit of bistable device are shown. Finally, experimental results presented for the SSHI circuit combined with the bistable technique show more energy can be harvested and larger bandwith.

Proven by both academia and industry, silicon carbide (SiC) semiconductors become ready to replace their silicon (Si) counterparts in power electronics. However, since they operate under high electrical constraints and at high switching frequencies, electromagnetic interference (EMI) level becomes higher. As a consequence, developing new modeling methods that require low time consumption, able to reconstruct accurately the transient behavior of a switching cell becomes a necessity. In this paper, a new modeling method based on configurable partial transfer functions (CPTFs) is proposed. This method allows an ultrafast accurate reconstruction of transient waveforms and an accurate EMI sources identification in both temporal and frequency domains. The parameters used to fill these CPTFs are identified by measurement or directly from the manufacturer technical files. The obtained results are verified by comparing simulations with experimental switching waveforms.

In this paper a novel strategy to increase the power of piezoelectric transformer (PT) by using a cooling system is presented. A thermal pad is directly attached to a PT to dissipate the heat and enhance the power capacity. In a piezoelectric ceramic the heat increase quickly when the vibration velocity become too large. Therefore there is a limit in vibration velocity magnitude. To explain the relationship between the vibration velocity, the output current and the output power we propose a theoretical-phenomenological model based on nonlinear equivalent circuit. It will be shown that the vibration velocity and temperature influence the characteristics of the piezoelectric transformer significantly. A large vibration velocity may lead that the internally thermo-physical feedback loop of the PT enters into an unstable state. In this paper, a PT with the cooling system was implemented in a DC/DC converter. A zero voltage switching (ZVS) half-bridge is used to drive the PT and a full-wave rectifier is used to obtain the DC load voltage. As a result, the maximum power of the PT based DC/DC converter can be increased from 4.41 W to 10.2 W at specific temperature. The study comprises of a theoretical part and experimental proof-of-concept demonstration of the proposed system.