Mobile Operating Systems Laboratory Dankook University


Mobile OS laboratory has been established since 2014, aiming on incubating diverse research projects.

Recently, we are working on three important themes:

  • Trustworthy execution environment with Linux
    • An ultimate goal of security in computing is to making computing systems trustworthy.
    • Integrity measurement, remote attestation using tamper-proof hardware are one approaches towards trusted computing. We are working on the bootloader, which is the origin of information flow, and a key parts in establishing root-of-trust.
    • UEFI secure boot, TrustedGRUB2 are common approaches to enhance security, relating trusted computing. We make Linux booting process harden using both booting so that we can lock down a hardware platform with tight integrity guarantee.
  • Mobile/IoT systems (Power-performance scaling of single-ISA heterogeneous multicore)
    • big.LITTLE cores in ARM presents very unique power-performance scaling. To get more performance with less energy, traditional DVFS, multicore are popularly used. We further investigates on asymmetric big.LITTLE cores, maximizing throughput, minimizing latency, lowering average power, and energy consumption.
    • Potential use cases include parallel processing applications (such as UHD video, some parallel gaming workload)
    • As a first step, integrated analysis for DVFS+multicore has conducted.
    • Cost-benefit analysis for latency-power consumption is currently undergoing.
  • Working with many cores (2K+) for dummies
    • AI (artificial intelligence), machine learning demands huge computing power. Many cores will definitely help what we can do more with computers. One of hardware implementation uses Xeon-Phi, that equips 50+ cores/card. We are looking into what we can do more with more cores.
    • Compared against GPU, and FPGA, Xeon-Phi provides more general, more IA-similar ISA, which means you can run more general applications than GPU. Note that GPGPU needs special support at layers of software stacks such as CUDA, OpenCL, that would not be necessary to Xeon-Phi.
    • Potential use case include massively parallel application that can do some complex security operations.
    • We are considering this platform, in relation with cloud computing environment. Running containers on Many core will open up new challenges as well.
  • More projects/ideas are coming, and welcomed. Send email to the prof. YOO for further questions.