Hyunwoo Joe

Smart phone power model generation using use pattern analysis

This paper presents an automated power model generation method for smart phone. Accurate power model can be achieved from usage pattern stored on the phone. This method can be applied to any mobile devices regardless of the manufacturer and model. The generated power model can be used to estimate energy consumption of the mobile device.

Automated power model generation method for smartphones

This paper presents an automated power model generation method for smartphones. Having an accurate power model of mobile devices can help identify the sources of fast discharge of the battery and provide opportunities for optimizing battery use. However, current methods rely on either external measurement equipment or internal current sensor support. This study presents a novel method for power modeling that exploits the usage patterns stored on the phone. This method can be applied to any mobile device regardless of the manufacturer and the model of the device. A power consumption analysis tool, and an artificial usage pattern generator tool have been developed for power modeling. The proposed method is able to generate an accurate power model of various smartphones even without detailed knowledge of the phone hardware.

Full virtualizing micro hypervisor for spacecraft flight computer

Recently partitioning and virtualization techniques for Integrated Modular Avionics (IMA) of aeronautics sector are proposed as the candidate architecture for safety-critical space applications. However, spacecraft software has subtle difference from aeronautic applications. Radiation particles in space environment cause various faults on the spacecraft computer. Requirement for autonomous operation with constrained resources is more stringent in space missions. Once it is launched, no way to refurbish the spacecraft without tremendous cost. These extra properties on top of those of regular aeronautic systems cause large software development cost in spacecraft projects. Summing up, spacecraft software should have real-time property, fault tolerance and efficient resource usage. In this paper, we introduce a hypervisor for spacecraft computer to improve reusability of inherited flight software from previous missions without redevelopment cycle. Fault tolerance is designed into the hypervisor to provide autonomous operation in space. We designed a prototype which is Type-II full virtualized hypervisor with kernel-level ARINC 653 partitioning on a dual-core LEON4-based flight computer for spacecraft. As the guest system, RTEMS-based flight software running on ERC32 flight computer is chosen because it has been used for many recent space missions and its flight software is likely to be reused when multicore LEON4 becomes widely available.

O-Sleep: Smartphones' Output-Oriented Power Saving Mode

Fast battery discharge is the most concerned issue for smart phone users. This is the result of the smart phones convergence trend and increasing complexity of its applications. Energy management techniques to reduce unnecessary battery usage are being studied. In this paper, we propose a new paradigm-shifting power saving mode for smart phones. O-Sleep is a smart phones output-oriented power saving mode. Rather than entering processor sleep mode when there is no user input, o-sleep puts a smart phones user interface into sleep mode when there is no meaningful output from the phone. While a device is processing a users request, output from the device may require processing time. We consider the situation as the devices output idle time and put the phones user interfaces into sleep mode while maintaining other subsystems in active state. To prove our concept, we have applied our technique onto various smart phone applications with varying operation environment. From the experiment, we have found we can save 37% of power consumption during up to 58% of its active usage with our test scenarios.

Output-oriented power saving mode for mobile devices

Abstract Fast battery discharge is still the most nerve wracking issue for smartphone users. Though many energy saving methods have been studied, still users are not satisfied with their phone’s battery. Power management system provides low power mode when the phone is not in use for a long time. When the user is interacting with the phone, current system assumes the user is interactive and should keep the device in active mode. However, this is not true. After the user’s interaction, the device processes the request and displays the result on the smartphone’s output device. During this period, the user cannot see any meaningful information from the phone. In this paper, we propose a new low power mode where we put smartphone’s output device into low power mode while phone is preparing result for the user. We named this as o-sleep, an output-oriented power saving mode. While a device is processing a user’s request, output from the device may require preparation time. We consider the situation as the device’s output idle time and put the phone’s user interfaces into sleep mode while maintaining other subsystems in active state. To prove our concept, we have applied our technique onto various smartphone applications with varying operation environment. From the experiment, we found that the smartphone entered the o-sleep mode up to 58% of its total usage time in various test scenarios. Usability study supported feasibility of our proposed method.

Effects of dynamic isolation for full virtualized RTOS and GPOS guests

Industrial systems currently include not only control processing with real-time operating system (RTOS) but also information processing with general-purpose operating system (GPOS). Multicore-based virtualization is an attractive option to provide consolidated environment when GPOS and RTOS are put in service on a single hardware platform. Researches on this technology have predominantly focused on the schedulability of RTOS virtual machines (VMs) by completely dedicated physical-CPUs (pCPUs) but have rarely considered parallelism or the throughput of the GPOS. However, it is also important that the multicore-based hypervisor adaptively selects pCPU assignment policy to efficiently manage resources in modern industrial systems. In this paper, we report our study on the effects of dynamic isolation when two mixed criticality systems are working on one platform. Based on our investigation of mutual interferences between RTOS VMs and GPOS VMs, we found explicit effects of dynamic isolation by special events. While maintaining low RTOS VMs scheduling latency, a hypervisor should manage pCPUs assignment by event-driven and threshold-based strategies to improve the throughput of GPOS VMs. Furthermore, we deal with implicit negative effects of dynamic isolation caused by the synchronization inside a GPOS VM, then propose a process of urgent boosting with dynamic isolation. All our methods are implemented in a real hypervisor, KVM. In experimental evaluation with benchmarks and an automotive digital cluster application, we analyzed that proposed dynamic isolation guarantees soft real-time operations for RTOS tasks while improving the throughput of GPOS tasks on a virtualized multicore system. We examine and analyze how a RTOS VM and a GPOS VM interact and influence each other.We analyze the explicit and implicit effects of dynamic isolation for vCPUs.The dynamic isolation shows low scheduling delay of RTOS and high throughput of GPOS.All of the proposed concepts are implemented on a full-fledged hypervisor.

Lua-Based Virtual Machine Platform for Spacecraft On-Board Control Software

Mission critical embedded software for autonomous operation requires high development cost due to its long development cycle. One of the potential solutions for reducing the cost is to reuse the software developed at previous missions. Virtual machine platform such as JVM is a good example to provide code portability across various missions. Flight software in aerospace field is adopting this concept to improve reusability and eventually to reduce development cost. In this paper, we propose a Lua-based virtualization environment for spacecraft flight software. Flight software for spacecraft control consists of a few tasks that are highly autonomous. Lua is chosen as the script language for programming the control tasks. Though Lua was designed with simplicity and portability, it only supports multithreading with collaborative coroutines. To support preemptive multitasking, we implement time slicing coroutines as spacecraft control processes. New coroutine scheduler is devised and time slicing functionality is added into the scheduler. Scheduler locking and message passing with external flight software are also implemented. Instead of modifying the Lua interpreter, we have exploited the debug support APIs for our implementation. For evaluation, we have implemented the flight software virtualization environment on the flight computer. Accuracy of the time slicing scheduler is also analyzed.

Linux-based memory efficient ARINC 653 partition scheduler

Development of the space industry has led to a diversity of specialized electronics-units implemented on space system. However, this provoked new problems such as increased systems complexity and the size. As a solution to the problem, the Integrated Modular Avionics (IMA) architecture and ARINC 653 standard have been suggested. This paper introduces a Linux-based memory efficient ARINC 653 partition scheduler, which employs fixed-priority preemptive scheduling algorithm for partitions scheduling. The implemented partition scheduler has small memory footprint for each partition and produces low partition switching overhead. The prototype was executed on a LEON4 processor, which is the European Space Agency (ESA) Next Generation Multicore Processor (NGMP) in the space sector. In evaluation, we analyzed execution trace, memory footprint and scheduling overhead.

Automatic Power Model Generation for Sensor Network Simulator

Energy consumption estimation in sensor network is a critical process for network lifetime estimation before actual deployment. Energy consumption can be estimated by simulating the sensor network with a power model. Power model is the key component for the accurate estimation. However, the power model is not publicly accessible and it is not easy to generate accurate fine-grain power model. In this paper we proposed a simplified but yet accurate power model for AVR-based sensor nodes. Also, we developed an automated power model generation tool. The tool generates an instruction-level power model that can be connected to sensor network simulators. We model the current consumption of ATmega128 instruction set which is the most widely used processor in sensor node. Loading, execution, and control of the measurement framework are managed by the tool. Using the tool, we can generate power models of various sensor nodes using ATmega128 as their processor. Experimental result shows that our tool successfully generated accurate power models of various sensor nodes including Mica2.

Dual display of virtual machines for automotive infotainment systems

We introduce GPU sharing between virtual machines on an embedded hypervisor. To achieve GPU virtualization, we apply an API remoting method which requires low overhead to OpenGL ES. For multimedia application, we target in-vehicle infotainment system and configure two guest virtual machines. The digital cluster is implemented on a bare-metal virtual machine. Furthermore, for multimedia application, we implement a linux-compatible 3D navigation. The framework can display both guest virtual machines on a single display panel so that output could be allocated separately on the panel. With this work, we explored how to virtualize GPU on a hypervisor for automotive systems.