Kernel Bus System to Build Virtual Machine Monitor and the Performance Service Framework and Method Therefor

• US 20110296411A1
• Filed: 05/25/2011
• Published: 12/01/2011
• Est. Priority Date: 05/25/2010
• Status: Active Grant

• Current Assignee: Transoft
• Original Assignee: Transoft
• Inventors: Tang, Chang Bin, Wang, Yanxin
• Granted Patent: US 8,832,688 B2
• Time in Patent Office: Days
• US Class Current: 718/1
• CPC Class Codes: G06F 9/45545 Guest-host, i.e. hypervisor is an application program itself, e.g. VirtualBox

Abstract

Some embodiments concern a kernel bus system for building at least one virtual machine monitor. The kernel bus system is based on kernel-based virtual machine. The kernel bus system is configured to run on a host computer. The host computer comprising one or more processors, one or more hardware devices, and memory. The kernel bus system can include: (a) a hyperbus; (b) one or more user space components; (c) one or more guest space components configured to interact with the one or more user space components via the hyperbus; (d) one or more VMM components having one or more frontend devices configure to perform I/O operations with the one or more hardware devices of the host computer using a zero-copy method or non-pass-thru method; (e) one or more para-virtualization components having (1) a virtual interrupt configured to use one or more processor instructions to swap the one or more processors of the host computer between a kernel space and a guest space; and (2) a virtual I/O driver configured to enable synchronous I/O signaling, asynchronous I/O signaling and payload delivery, and pass-through delivery independent an QEMU emulation; and (f) one or more KVM components. The hyperbus, the one or more user space components, the one or more guest space components, the one or more VMM components, the one or more para-virtualization components, and the one or more KVM components are configured to run on the one or more processors of the host computer. Other embodiments are disclosed.

Claims

1. A kernel bus system for building at least one virtual machine monitor, the kernel bus system is based on a kernel-based virtual machine, the kernel bus system is configured to run on a host computer, the host computer comprising one or more processors, one or more hardware devices, and memory, the kernel bus system comprising:

a hyperbus;

one or more user space components;

one or more guest space components configured to interact with the one or more user space components via the hyperbus;

one or more virtual machine monitor components comprising;

one or more frontend devices configured to perform one or more first input/output operations with the one or more hardware devices of the host computer using a zero-copy method or a non-pass-thru method; and

one or more para-virtualization components comprising;

a virtual interrupt configured to use one or more processor instructions to swap the one or more processors of the host computer between a kernel space and a guest space; and

a virtual input/output driver configured to enable synchronous input/output signaling, asynchronous input/output signaling, payload delivery, and pass-through delivery substantially independent of an QEMU process emulation,

wherein;

the hyperbus, the one or more user space components, the one or more guest space components, the one or more virtual machine monitor components, and the one or more para-virtualization components are configured to run on the one or more processors of the host computer.

2. The kernel bus system of claim 1, further comprising:

one or more kernel-based virtual machine components, wherein;

the one or more guest space components comprise one or more virtual machine operating system kernels;

the hyperbus comprises one or more kernel-based virtual machine backend components; and

the one or more kernel-based virtual machine backend components of the hyperbus are configured to determine one or more code paths for the one or more virtual machine operating system kernels of the one or more guest space components; and

the one or more kernel-based virtual machine components are configured to run on the one or more processors of the host computer.

3. The kernel bus system of claim 2, wherein:

the one or more guest space components comprise a virtual machine;

the virtual machine of the one or more guest space components is configured to execute the one or more first input/output operations;

the virtual machine comprises a virtual machine kernel;

the hyperbus comprises the user-defined kernel entity;

the hyperbus is configured such that when the virtual machine of the one or more guest space components begins execution the one or more first input/output operations, the hyperbus delivers a signal to the user-defined kernel entity of the hyperbus, wherein the signal informs the user-defined kernel entity of the hyperbus that the virtual machine is executing the one or more first input/output operations; and

the user-defined kernel entity is configured to access at least one of the one or more virtual machine monitor components or the one or more kernel-based virtual machine components to complete the one or more first input/output operations.

4. The kernel bus system of claim 1, wherein:

the one or more guest space components comprise at least one virtual machine;

the at least one virtual machine comprises a user space application;

the hyperbus comprises one or more hyperbus components;

the at least one virtual machine is configured such that when the user space application of the at least one virtual machine executes one or more second input/output operations, the at least one virtual machine interacts with the one or more hyperbus components via a trap interrupt; and

the one or more hyperbus components is configured to switch the one or more second input/output operations of the user space application of the at least one virtual machine to a user space when the at least one virtual machine interacts with the one or more hyperbus components via the trap interrupt.

5. The kernel bus system of claim 1, wherein:

the one or more guest space components comprise one or more virtual machines; and

the hyperbus is configured to simultaneously operate on the one or more virtual machines of the one or more guest space components using one or more local kernel calls and shared memory of the memory.

6. The kernel bus system of claim 1, further comprising:

a ISA/ABI module,

wherein;

the one or more guest space components comprise one or more virtual machines;

the one or more user space components comprise a device manager;

the device manager of the one or more user space components is configured to manage the ISA/ABI module using one or more device management functions;

the one or more device management functions comprise discovering, assigning, traffic-controlling, and band-binding; and

the hyperbus is configured to broadcast the device management functions to the one or more virtual machines of one or more guest space components.

7. The kernel bus system of claim 1, wherein:

the hyperbus comprises one or more policy decision function modules and one or more policy execution point modules;

the one or more user space components comprise a policy rule base and a rule base user interface;

the one or more policy decision function modules of the hyperbus are configured to receive one or more policies from the rule base user interface of the one or more user space components and make decisions regarding the one or more policies using one or more rules of the policy rule base of the one or more user space components; and

the one or more policy execution point modules are configured to execute the policy passed from a policy decision function module.

8. The kernel bus system of claim 1, wherein:

the one or more guest space components comprise one or more virtual machines;

the one or more user space components comprise a device manager, a policy rule base, and a rule base user interface;

the hyperbus comprises;

a shadow ISA/ABI; and

one or more policy decision function modules;

the rule base user interface is configured to allow an administrator to set one or more policies of the one or more policy decision function modules of the hyperbus;

the device manager of the one or more user space components is configured to send one or more device management functions to the hyperbus;

the one or more device management functions comprise discovering, assigning, traffic-controlling, and band-binding; and

the shadow ISA/ABI of the hyperbus is configured to broadcast the one or more device management functions to the one or more virtual machines of the one or more guest space components.

9. The kernel bus system of claim 1, wherein:

the hyperbus comprises a software hash calculator and a hyperbus TCS agent;

the software hash calculator is configured to calculate a signed hash value;

the hyperbus TCS agent is configured to obtain a stored hash value from at least one of a peripheral device module or a storage module; and

the hyperbus TCS agent is further configured to stop execution of the kernel bus system if the stored hash value does not match the signed hash value.

10. The kernel bus system of claim 1, wherein:

the one or more user space components comprise a performance database, a bottleneck identifier, and a workload mediator;

the hyperbus comprises one or more policy execution point modules;

the one or more policy execution point modules are configured to provide first performance data to the bottleneck identifier;

the performance database is configured to store historical performance data and statistical performance data;

the bottleneck identifier is configured to at least partially determine bottleneck information using the first performance data, the historical performance data, and the statistical performance data; and

the workload mediator is configured as a policy decision function to reconfigure a task time slice based on the bottleneck information provided by the bottleneck identifier.

11. The kernel bus system of claim 12, wherein:

the hyperbus further comprises a radix tree traverser;

the radix tree traverser is configured to determine a radix tree; and

the radix tree is configured to be used to address the first performance data.

12. The kernel bus system of claim 1, wherein:

the hyperbus comprises;

a local hyperbus configured to run the one or more processors of the host computer; and

a networked hyperbus configured to run one or more processors of a remote computer.

13. The kernel bus system of claim 1, wherein:

the hyperbus comprises;

a performance service module configured to enable an administrator to adjust a performance of the host computer; and

the one or more user space components comprise;

a policy manager; and

a user space application library configured to control the performance service module of the hyperbus via the policy manager.

14. The kernel bus system of claim 1, wherein:

the one or more processor instructions used to swap between the kernel space and the guest space comprise VMExit and VMEntry.

15. A method of building at least one virtual machine monitor via a kernel bus system on a host computer, the method comprising:

providing a hyperbus configured to run on the host computer;

providing one or more user space components configured to run on the host computer;

providing one or more guest space components configured to run on the host computer and further configured to interact with the one or more user space components via the hyperbus;

providing one or more virtual machine monitor components configured to run on the host computer and comprising;

one or more frontend devices configured to perform one or more input/output operations with one or more hardware devices of the host computer using a zero-copy procedure or a non-pass-thru procedure;

providing one or more para-virtualization components configured to run on the host computer and comprising;

a virtual interrupt configured to use one or more processor instructions to swap one or more processors of the host computer between a kernel space and a guest space; and

a virtual input/output driver configured to enable synchronous input/output signaling, asynchronous input/output signaling and payload delivery, and pass-through delivery substantially independent of a QEMU process emulation.

16. A computer system configured to run on a host computer, the host computer comprising one or more processors, one or more hardware devices, and memory, the computer system comprising:

a hyperbus comprising;

one or more hyperbus service modules configured to run on the host computer; and

one or more hyperbus backend components configured to run on the host computer; and

one or more guest space components configured to run on the host computer and comprising;

one or more virtual machines comprising one or more virtual machine applications, the one or more virtual machine applications configured to perform one or more input/output operations;

one or more virtual machine operating system kernels comprising one or more drivers, the one or more drivers are configured to interact with the one or more hyperbus service modules to facilitate the one or more virtual machines performing the one or more input/output operations; and

one or more VIRTIO devices,

wherein;

the one or more virtual machine operating system kernels of the one or more guest space components are configured to enter the one or more hyperbus backend components using at least one trap instruction; and

the one or more hyperbus backend components are configured to choose a code path from one or more code paths to interact with the one or more VIRTIO devices.

17. The computer system of claim 16, wherein:

the one more hyperbus service modules comprises an AIO service agent.

18. The computer system of claim 16, wherein:

a QEMU module comprises the one or more VIRTIO devices.

19. The computer system of claim 16, further comprising:

one or more user space components,

wherein;

the one or more user space components comprise a performance database, a bottleneck identifier, and a workload mediator;

the hyperbus comprises two or more policy execution point modules;

the two or more policy execution point modules are configured to provide first performance data to the bottleneck identifier;

the performance database is configured to store historical performance data and statistical performance data;

the bottleneck identifier is configured to at least partially determine bottleneck information using the first performance data, the historical performance data, and the statistical performance data; and

the workload mediator is configured to reconfigure a task time slice based on the bottleneck information provided by the bottleneck identifier.

20. The computer system of claim 17, wherein:

the hyperbus further comprises a radix tree traverser;

the radix tree traverser is configured to determine a radix tree; and

the first performance data comprise the radix tree.