1 / 17

Project Overview

Project Overview. 2014/05/05. Current Project. “Research on Embedded Hypervisor Scheduler Techniques” Design an energy-efficient scheduling mechanism in virtualized environment. Assume the hardware platform is asymmetric multi-core . big.LITTLE core architecture.

elsa
Télécharger la présentation

Project Overview

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Project Overview 2014/05/05

  2. Current Project • “Research on Embedded Hypervisor Scheduler Techniques” • Design an energy-efficient scheduling mechanism in virtualized environment. • Assume the hardware platform is asymmetric multi-core. • big.LITTLE core architecture.

  3. big.LITTLE Core Architecture • Developed by ARM in Oct. 2011. • Combine two kinds of architecturally compatible cores. • To create a multi-core processor that can adjust better to dynamic computing needs and use less power than clock scaling alone. • big cores are more powerful but power-hungry, while LITTLE cores are low-power but (relatively) slower.

  4. Three Types of Models • Cluster migration • CPU migration(In-Kernel Switcher) • Heterogeneous multi-processing (global task scheduling)

  5. Type 1: Cluster Migration • Either big or LITTLE cores are used simultaneously.

  6. Type 2: CPU Migration • Logical CPU: a pair of big and LITTLE core. • Only one of the two cores in a pair is powered up and processing tasks at a time.

  7. Type 3: HMP • All the big and LITTLE cores can be used at the same time.

  8. Task Scheduling in Asymmetric Multi-cores • Task should be distributed unevenly. • Only critical tasks should execute on big cores in order to minimize power consumption. • Different from traditional SMP scheduling. • Needs to design new scheduling algorithms.

  9. Asymmetric Multi-core in Datacenter • Asymmetric multi-core is not limited to hand-held devices or embedded systems. • There will be servers with asymmetric cores in a datacenter. • Virtualized environment in datacenter.

  10. Current Hypervisor Architecture and Problem Low computing resource requirement High computing resource requirement GUEST1 GUEST2 GUEST2 Task 3 Task 1 Task 4 Task 2 Android Framework Android Framework Android Framework If Guest OS scheduler is not big.LITTLE-aware, it will assign tasks to vCPUs evenly in order to achieve load balancing. OSKernel OS Kernel OSKernel Scheduler Scheduler Scheduler VCPU VCPU VCPU VCPU VCPU VCPU Hypervisor vCPUscheduler will assign vCPUs evenly to physical ARM cores since it is not big.LITTLE-aware. Hypervisor vCPU Scheduler Cannot take advantage on big.LITTLEcore architeture ARM Cortex-A15 ARM Cortex-A7 Performance Power-saving

  11. Current Hypervisor Architecture and Problem(Cont.) GUEST1 GUEST2 Android Framework Android Framework If Guest OS scheduler is not big.LITTLE-aware, it will assign tasks to vCPUs evenly in order to achieve load balancing. OS Kernel OS Kernel Scheduler Scheduler Task 4 Task 3 VCPU VCPU VCPU VCPU Even if hypervisor vCPU scheduler is big.LITTLE-aware, it will schedule these vCPUs to either big cores or LITTLE cores since they have the same loading. Task 2 Task 1 Hypervisor b-L vCPU Scheduler Both on big core Cannot take advantage on big.LITTLEcore architeture ARM Cortex-A15 ARM Cortex-A7 Ot both on LITTLE core Performance Power-saving

  12. Challenge • The hypervisor scheduler cannot take advantage of big.LITTLE architecture if the scheduler inside guest OS is not big.LITTLE aware.

  13. Our Solution • Apply VM introspection(VMI) to retrieve the process list in a VM. • VMI is a technique that allows the hypervisor to inspect the contents of the VM in real-time. • Modify the CPU masks of tasks in the VM in order to create an illusion of “big vCPU” and “LITTLE vCPU”. • Hypervisor scheduler can assign the vCPU to corresponding big or LITTLE cores.

  14. Hypervisor Architecture with VMI Low computing resource requirement High computing resource requirement GUEST1 GUEST2 GUEST2 [1|0] [0|1] Task 3 Task 1 [1|0] [0|1] Task 4 Task 2 Android Framework Android Framework Android Framework Linaro Linux Kernel OS Kernel OS Kernel Scheduler Scheduler Scheduler VM Introspector gathers task information from Guest OS Treat thisvCPU as LITTLE core since tasks with low computing requirement are scheduled here. VCPU VCPU VCPU VCPU VCPU VCPU VCPU Hypervisor Task-to-vCPUMapper Hypervisor vCPUscheduler will schedule big vCPUto A15, and LITTLE vCPU to A7. Modify the CPU mask of each task according to the task information from VMI VM Introspector b-L vCPU Scheduler ARM Cortex-A15 ARM Cortex-A7 Performance Power-saving

  15. Hypervisor Architecture with VMI(Cont.) Guest OS 1has two task with high computing requirement and two task with low computing requirement Guest OS 2has two task with low computing requirement GUEST1 GUEST2 [1|0] [0|1] Task 1 Task 2 Task 3 Task 1 [1|1] [1|1] [1|0] [0|1] Task 4 Task 2 Android Framework Android Framework OS Kernel OS Kernel Scheduler Scheduler VM Introspector gathers task information from Guest OS Treat this vCPU as LITTLE core since tasks with low computing requirement are scheduled here. VCPU VCPU VCPU VCPU VCPU VCPU VCPU Hypervisor Task-to-vCPUMapper Hypervisor vCPUscheduler will schedule big vCPUto A15, and LITTLE vCPU to A7. Modify the CPU mask of each task according to the task information from VMI VM Introspector b-L vCPU Scheduler ARM Cortex-A15 ARM Cortex-A7 Performance Power-saving

  16. Current Status • Implementing the proposed architecture. • Trying to enable VMI on Xen 64 bit. • Developing new scheduling algorithms for big.LITTLE core.

  17. Q&A

More Related