(51) Int Cl.: G06F 9/455 ( ) G06F 9/50 ( )

Transcription

1 (19) TEPZZ 6987 B_T (11) EP B1 (12) EUROPEAN PATENT SPECIFICATION (4) Date of publication and mention of the grant of the patent: Bulletin 1/32 (21) Application number: (22) Date of filing: (1) Int Cl.: G06F 9/4 (06.01) G06F 9/0 (06.01) (86) International application number: PCT/CN11/07660 (87) International publication number: WO 12/3728 ( Gazette 12/32) (4) METHOD FOR DISPATCHING CENTRAL PROCESSING UNIT OF HOTSPOT DOMAIN VIRTUAL MACHINE AND VIRTUAL MACHINE SYSTEM VERFAHREN ZUR VERTEILUNG EINER ZENTRALPROZESSOREINHEIT EINER HOTSPOT VM UND VIRTUELLES MASCHINENSYSTEM PROCÉDÉ DE RÉPARTITION D UNITÉ CENTRALE DE MACHINE VIRTUELLE DE DOMAINE HOTSPOT ET SYSTÈME DE MACHINE VIRTUELLE (84) Designated Contracting States: AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR (43) Date of publication of application: Bulletin 14/08 (73) Proprietor: Huawei Technologies Co., Ltd. Longgang District Shenzhen, Guangdong (CN) (74) Representative: Epping - Hermann - Fischer Patentanwaltsgesellschaft mbh Schloßschmidstraße München (DE) (6) References cited: CN-A CN-A CN-A CN-A US-A US-A US-A US-A US-A (72) Inventor: HONG, Kaixing Shenzhen Guangdong (CN) EP B1 Note: Within nine months of the publication of the mention of the grant of the European patent in the European Patent Bulletin, any person may give notice to the European Patent Office of opposition to that patent, in accordance with the Implementing Regulations. Notice of opposition shall not be deemed to have been filed until the opposition fee has been paid. (Art. 99(1) European Patent Convention). Printed by Jouve, 7001 PARIS (FR)

2 1 EP B1 2 Description FIELD OF THE INVENTION [0001] The embodiments of the present invention relate to the computer technology, and particularly, to a method for dispatching CPUs of a hotspot domain virtual machine and a virtual machine system. BACKGROUND OF THE INVENTION [0002] The virtualization technology is a decoupling method for separating bottom hardware devices from an upper-layer operation system and applications. It introduces a Virtual Machine Monitor (VMM) layer to directly manage bottom hardware resources, and create a Virtual Machine (VM) independent from the bottom hardware for a use by the upper-layer operation system and applications. The VM may be classified into a hotspot domain VM and a common domain VM. The common domain VM is created by a user to operate general tasks. The hotspot domain VM plays a pivotal role in the system and is depended by the common domain VM. In various virtualization systems, the hotspot domain VM has different features. For example, in some systems there is a driver field virtualization, through which the common domain virtualization interacts with the physical hardware. [0003] In the VM system, the VM has no real physical resource. The CPU assigned to the VM is not a real physical CPU, but a Virtual Central Processing Unit (vcpu) assigned to the VM by the VMM. In the virtualization environment, dual-dispatching of the CPU resources exists: on one hand, an Operation System (OS) dispatcher inside the VM dispatches a thread to the vcpu; on the other hand, a VMM dispatcher dispatches the vcpu to the physical CPU. The implementation of the OS dispatcher inside the VM is dependent upon the specific OS and cannot be changed, thus the VMM dispatcher has a very important influence on the performance of the whole VM system. [0004] Currently, the VMM dispatcher usually adopts a dispatching strategy of equally dispatching the vcpus of the common domain VM and the hotspot domain VM, so the common domain VM may preempt the CPU time of the hotspot domain VM, which causes a poor real-time performance of the hotspot domain VM, and no response to the request from the common domain VM is made in time, particularly, when there are a large number of VMs, the whole performance may be sharply degraded. Another dispatching strategy is that the VMM dispatcher assigns a fixed physical CPU resource to the hotspot domain VM, and the number of the default vcpus is correlative to that of the real physical CPUs. When a large number of physical CPUs are available in the system, there are also many vcpus. On one hand, resource waste will be caused by a light system load. On the other hand, the system performance cannot be improved by increasing the vcpus in the hotspot domain VM in case of a heavy system load, especially when too many vcpus exist in the common domain. Therefore, the existing VM system has a low CPU resource utilization rate and a poor system performance. [000] US Patent Application US 09/0314 A1 discloses a method for scheduling true CPU resources for a virtual machine system. A virtual machine monitor (VMM) acquires information about loads on virtual CPUs (VCPUs), and the VMM assigns the true CPU resources to the VCPUs based on the information about the loads on the VCPUs in such a manner that more true CPU resources are assigned to a VCPU with a higher load. Specifically, the information about the loads is determined based on an interrupt response time or the number of interrupt requests to be processed in a VCPU, or based on an interrupt response time and the number of interrupt requests to be processed in a VCPU. SUMMARY OF THE INVENTION [0006] The embodiments of the present invention provide a method for dispatching CPUs of a hotspot domain VM and a VM system, so as to solve the deficiency of the prior art that the VM system has a low CPU resource utilization rate and a poor system performance. [0007] An embodiment of the present invention provides a method for dispatching CPUs of a hotspot domain VM, wherein each of vcpus of the hotspot domain VM is bound to a different physical CPU, the method including: determining a comprehensive utilization rate of the vcpus of the hotspot domain VM according to counted use information of the vcpus of the hotspot domain VM, wherein the use information of the vcpus includes a utilization rate of the vcpus of the hotspot domain VM and a use time occupancy rate of the vcpus of the hotspot domain VM, wherein the use time occupancy rate of the vcpus of the hotspot domain VM is a ratio of a sum of the operation time of all the vcpus of the hotspot domain VM to a total time, the total time is a sum of the operation time of all the vcpus and the wait time of all the vcpus, and the utilization rate of the vcpus of the hotspot domain VM is a ratio of a sum of the non-idle time of all the vcpus inside the hotspot domain VM to a sum of the operation time of all the online vcpus of the hotspot domain VM; determining to add or reduce at least one online vcpu for the hotspot domain VM, according to the comprehensive utilization rate of the vcpus and the maximum number of the vcpus assigned to the hotspot domain VM; after an online vcpu is added, adjusting the vcpus of a common domain VM in the running queue of a physical CPU bound to the online vcpu, so that the online vcpu can be dispatched on the physical CPU while the vcpus of the common domain VM cannot 2

3 3 EP B1 4 be dispatched on the physical CPU; or after an online vcpu is reduced, adjusting the vcpu of the hotspot domain VM in the running queue of a physical CPU bound to the offline vcpu, so that the vcpu of the common domain VM can be dispatched on the physical CPU. [0008] An embodiment of the present invention provides a VM system, including a hotspot domain VM, common domain VMs and a VMM, wherein: the VMM is configured to bind a different physical CPU to each of vcpus of the hotspot domain VM; the hotspot domain VM is configured to determine a comprehensive utilization rate of the vcpus of the hotspot domain VM according to counted use information of the vcpus of the hotspot domain VM, and add or reduce at least one online vcpu according to the comprehensive utilization rate of the vcpus and the maximum number of the vcpus assigned to the hotspot domain VM; wherein the use information of the vcpus includes a utilization rate of the vcpus of the hotspot domain VM and a use time occupancy rate of the vcpus of the hotspot domain VM, wherein the use time occupancy rate of the vc- PUs of the hotspot domain VM is a ratio of a sum of the operation time of all the vcpus of the hotspot domain VM to a total time, the total time is a sum of the operation time of all the vcpus and the wait time of all the vcpus, and the utilization rate of the vcpus of the hotspot domain VM is a ratio of a sum of the non-idle time of all the vcpus inside the hotspot domain VM to a sum of the operation time of all the online vcpus of the hotspot domain VM; the VMM is further configured to adjust the vcpus of a common domain VM in the running queue of a physical CPU bound to an online vcpu after the online vcpu is added, so that the online vcpu can be dispatched on the physical CPU while the vcpus of the common domain VM cannot be dispatched on the physical CPU; and the VMM is further configured to adjust the vcpu of the hotspot domain VM in the running queue of a physical CPU bound to an offline vcpu after the online vcpu is reduced, so that vcpus of the common domain VM can be dispatched on the physical CPU. [0009] With the method for dispatching CPUs of a hotspot domain VM CPU and the VM system according to the embodiments of the present invention, the hotspot domain VM dynamically adjusts the number of the online vcpus in the hotspot domain VM according to the use information of the vcpus. When the vcpus of the hotspot domain VM have a heavy load, an online vcpu is added to the hotspot domain VM, and the VMM enables the added online vcpu to exclusively occupy the bound physical CPU, so that the online vcpu of the common domain VM cannot be dispatched on the physical CPU Thus, the vcpu of the hotspot domain VM may preferably use the physical CPU prior to the vcpu of the common domain VM, so that a response to the request from the vcpu of the hotspot domain VM can be made in time, and the phenomenon that the vcpu of the common domain VM waits for the hotspot domain VM is avoided, thereby improving the whole performance of the virtual system. When the vcpus of the hotspot domain VM have a light load, the VMM deletes an online vcpu from the hotspot domain VM, and cancels the exclusivity of the vcpu of the hotspot domain VM to the physical CPU, so that the vcpu of the common domain VM can be dispatched on physical CPU, thereby improving the performance of the common domain VM, and preventing the waste of the CPU resources. BRIEF DESCRIPTION OF THE DRAWINGS [00] In order to more clearly describe the technical solutions of the embodiments of the present invention or the prior art, the drawings to be used in the descriptions of the embodiments or the prior art are briefly introduced as follows. It will be appreciated that the following drawings just illustrate some embodiments of the present invention, and a person skilled in the art can obtain other drawings from these drawings without paying a creative effort. Fig. 1A is a flowchart of a method for dispatching CPUs of a hotspot domain VM according to an embodiment of the present invention; Fig. 1B is a structure diagram of a VM according to an embodiment of the present invention; Fig. 2 is a flowchart of another method for dispatching CPUs of a hotspot domain VM according to an embodiment of the present invention; Fig. 3 is a structure diagram of a VM system according to an embodiment of the present invention; and Fig. 4 is a structure diagram of another VM system according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EM- BODIMENTS [0011] In order to make the objects, technical solutions and advantages of the present invention be clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described as follows with reference to the drawings. Apparently, the described embodiments are just a part of embodiments of the present invention rather than all the embodiments. Based on the embodiments of the present invention, any other embodiment obtained by a person skilled in the art without paying a creative effort will fall within the protection scope of the present invention. [0012] An embodiment of the present invention provides a method for dispatching CPUs of a hotspot domain 3

4 EP B1 6 VM to dynamically adjust the number of vcpus in the hotspot domain VM according to the system load, and assign exclusive physical CPU resources thereto. [0013] Fig. 1A is a flowchart of a method for dispatching CPUs of a hotspot domain VM according to an embodiment of the present invention, and Fig. 1B is a structure diagram of a VM according to an embodiment of the present invention. As illustrated in Fig. 1B, the VMM directly manages the bottom hardware resources, and creates a VM independent from the bottom hardware for a use by the upper-layer OS and applications. Each VM includes applications (APPS) and an OS dispatcher. Herein, the VM includes one hotspot domain VM and multiple common domain VMs each depending on the hotspot domain VM. The vcpu of the VM is in either an online state or an offline state. The online vcpu may be used by the VM OS while being dispatched by the VMM to the physical CPU. The online vcpu assigned with a thread is a non-idle online vcpu, otherwise it is an idle online vcpu. [0014] When creating vcpus for the hotspot domain VM, the embodiment of the present invention binds a different physical CPU to each vcpu of the hotspot domain VM, i.e., the vcpus of the hotspot domain VM are one-by-one corresponding to the physical CPUs. The vcpu of the hotspot domain VM can only operate in the bound physical CPU. In case of an offline vcpu of the hotspot domain VM, the VMM may temporarily assign the physical CPU corresponding to the offline vcpu to a vcpu of the common domain VM for a use. Once the vcpu becomes online (becoming online from offline), the vcpu of the common domain VM will not use the physical CPU corresponding to the online vcpu. [001] The embodiment is described hereinafter with reference to Fig. 1A. [0016] In step 11, the hotspot domain VM determines a comprehensive utilization rate of the vcpus of the hotspot domain VM according to counted use information of the vcpus of the hotspot domain VM, wherein the use information of the vcpus includes a utilization rate of the vcpus of the hotspot domain VM and a use time occupancy rate of the vcpus of the hotspot domain VM. [0017] The hotspot domain VM monitors the use information of the vcpus of the hotspot domain VM in real time, including: the utilization rate of the vcpus of the hotspot domain VM counted from the OS dispatcher of the hotspot domain VM and the use time occupancy rate of all the vcpus of the hotspot domain VM obtained from the VMM, and the hotspot domain VM determines the comprehensive utilization rate of the vcpus of the hotspot domain VM according to the utilization rate of the vcpus and the use time occupancy rate of the vcpus. [0018] Herein, the use time occupancy rate of the vc- PUs of the hotspot domain VM may be determined from a sum of the operation time and a sum of the wait time of all the vcpus of the hotspot domain VM. For example, the use time occupancy rate of the vcpus of the hotspot domain VM is a ratio of the sum of the operation time of all the vcpus of the hotspot domain VM to the total time, wherein the total time is a sum of the operation time of all the vcpus and the wait time of all the vcpus. [0019] Herein, the utilization rate of the vcpus of the hotspot domain VM is determined from the actual nonidle time of the vcpus inside the VM. The utilization rate of the vcpus of the hotspot domain VM is determined as a ratio of a sum of the non-idle time of all the vcpus inside the hotspot domain VM to a sum of the operation time of all the online vcpus of the hotspot domain VM. [00] In step 12, it is determined to add or reduce at least one online vcpu for the hotspot domain VM, according to the comprehensive utilization rate of the vc- PUs and the number of the vcpus assigned to the hotspot domain VM by the VMM. [0021] The comprehensive utilization rate of the vc- PUs of the hotspot domain VM may reflect the use condition of the vcpus by the hotspot domain VM. A high comprehensive utilization rate of the vcpus of the hotspot domain VM indicates a heavy load of the online vcpus of the hotspot domain VM, and the number of the online vcpus may be increased for the hotspot domain VM. A low comprehensive utilization rate of the vcpus of the hotspot domain VM indicates a light load of the online vcpus of the hotspot domain VM, and the number of the online vcpus may be decreased for the hotspot domain VM. During the increase of the online vcpu(s), online operation may be performed on one or more offline vcpus of the hotspot domain VM. During the decrease of one or more online vcpus, offline operation may be performed on one or more online vcpus of the hotspot domain VM. [0022] When the VM system is initialized, the VMM assigns a certain number of vcpus to the hotspot domain VM. Defaultly, the maximum number of the vcpus assigned to the hotspot domain VM by the VMM is the same as the number of the physical CPUs. When the hotspot domain VM has no enough vcpus, the OS dispatcher of the hotspot domain VM adds online vcpus to the hotspot domain VM, but the number of the online vcpus of the hotspot domain VM shall not exceed the number of the vcpus assigned to the hotspot domain VM by the VMM. That is, when the number of the online vcpus of the hotspot domain VM is equal to the maximum number of the vcpus initially assigned to the hotspot domain VM by the VMM, all the vcpus assigned to the hotspot domain VM by the VMM are in the online state, and no more online vcpu can be added to the hotspot domain VM. [0023] In step 13a, after an online vcpu is added, the VMM adjusts the vcpus of the common domain VM in the running queue of a physical CPU bound to the online vcpu, so that the online vcpu can be dispatched on the physical CPU while the vcpus of the common domain VM cannot be dispatched on the physical CPU. [0024] After adding the online vcpu to the hotspot domain VM, the VMM adjusts the vcpu of the common domain VM in the running queue of the physical CPU bound to the online vcpu, so that the online vcpu ex- 4

5 7 EP B1 8 clusively occupies the bound physical CPU. A specific method may include: the VMM creates a running queue for each physical CPU, and all the vcpus are added to different running queues to wait for a dispatching by the VMM. After adding the online vcpu to the hotspot domain VM, the VMM incorporates the online vcpu into the running queue of the physical CPU bound to the online vcpu while removing all the vcpus of the common domain VM from the running queue of the physical CPU, and records the removed vcpus in the affinity information of the physical CPU. [002] In step 13b, after an online vcpu is reduced, the VMM adjusts the vcpu of the hotspot domain VM in the running queue of a physical CPU bound to the offline vcpu, so that the vcpu of the common domain VM can be dispatched on the physical CPU. [0026] When the online vcpu of the hotspot domain VM is reduced, one or more online vcpus are changed to be offline, the VMM adjusts the vcpu of the hotspot domain VM in the running queue of the physical CPU bound to the offline vcpu, and cancels the exclusivity of the offline vcpu to the bound physical CPU, so that the vcpu of the common domain VM can be dispatched in the running queue of the bound physical CPU. A specific method may include: the VMM removes the offline vcpu from the running queue of the bound physical CPU, and incorporates the vcpu of the common domain VM recorded in the affinity information of the physical CPU bound to the offline vcpu into the running queue of the physical CPU bound to the offline vcpu. [0027] With the method for dispatching CPUs of a hotspot domain VM according to the embodiment of the present invention, the hotspot domain VM dynamically adjusts the number of the online vcpus in the hotspot domain VM according to the use information of the vc- PUs. When the vcpus of the hotspot domain VM have a heavy load, an online vcpu is added to the hotspot domain VM, and the VMM enables the added online vcpu to exclusively occupy the bound physical CPU, so that the online vcpu of the common domain VM cannot be dispatched on the physical CPU. Thus, the vcpu of the hotspot domain VM may preferably use the physical CPU prior to the vcpu of the common domain VM, so that a response to the request from the vcpu of the hotspot domain VM can be made in time, and the phenomenon that the vcpu of the common domain VM waits for the hotspot domain VM is avoided, thereby improving the whole performance of the virtual system. When the vcpus of the hotspot domain VM have a light load, the VMM deletes an online vcpu from the hotspot domain VM, and cancels the exclusivity of the vcpu of the hotspot domain VM to the physical CPU, so that the vcpu of the common domain VM can be dispatched on physical CPU, thereby improving the performance of the common domain VM, and preventing the waste of the CPU resources. [0028] Fig. 2 is a flowchart of another method for dispatching CPUs of a hotspot domain VM according to an embodiment of the present invention. The embodiment is described hereinafter with reference to Fig. 2. [0029] In step 21, the hotspot domain VM is started based on a configuration file, and the VMM creates N vcpus for the hotspot domain VM and binds a physical CPU to each vcpu. [00] During the initialization of the virtual system, the hotspot domain VM is started based on the configuration file specifying the maximum number N and the minimum number n of the vcpus that can be owned by the hotspot domain VM. After the hotspot domain VM is started, the number of the vcpus assigned to the hotspot domain VM by the VMM is N, which has to be less than or equal to the number of the actual physical CPUs. Defaultly, the maximum number of the vcpus is the same as the number of the physical CPUs. When being created, each vcpu of the hotspot domain VM is bound to one physical CPU, i.e., the corresponding vcpu can only be dispatched on the bound physical CPU, so as to prevent the vcpus of the hotspot domain VM from migrating to other physical CPU due to the optimization by the VMM dispatcher. [0031] In step 22, n online vcpus is reserved for the hotspot domain VM according to the configuration file, and other vcpus are changed to be offline. The reserved n online vcpus is enabled to exclusively occupy their bound physical CPUs. [0032] When the hotspot domain VM is started, the VMM creates N vcpus for the hotspot domain VM. When the system starts to run, only n online vcpus are reserved, and other N-n vcpus are changed to be offline. Subsequently, after the OS dispatcher adds or reduces the online vcpu for the hotspot domain VM, the number of the online vcpus of the hotspot domain VM varies between N and n. [0033] In step 23, the vcpu information of the hotspot domain VM and the information of the physical CPU of the system are obatained. [0034] The vcpu information of the hotspot domain VM obtained from the VMM includes not only the state (online and offline) of each vcpu of the hotspot domain VM and the number of the vcpus, but also the physical CPUs bound to the online vcpus. The CPU information of the physical system mainly includes the number of the physical CPUs. [003] In step 24, a performance statistics module obtains the utilization rate of the vcpus of the hotspot domain VM from the OS dispatcher of the hotspot domain VM, and obtains the use time occupancy rate of all the vcpus of the hotspot domain VM from the VMM. [0036] The OS dispatcher of the hotspot domain VM counts the utilization rate of the vcpus of the hotspot domain VM, and the VMM counts the use time occupancy rate of all the vcpus of the hotspot domain VM. The use time occupancy rate of the vcpus of the hotspot domain VM may be a ratio of a sum of the operation time of all the vcpus of the hotspot domain VM to the total time, wherein the total time is a sum of the operation time of

6 9 EP B1 all the vcpus and the wait time of all the vcpus. The utilization rate of the vcpus of the hotspot domain VM may be a ratio of a sum of the non-idle time of all the vcpus inside the hotspot domain VM to a sum of the operation time of all the online vcpus of the hotspot domain VM. [0037] In step 2, the performance statistics module determines a comprehensive utilization rate of the vc- PUs of the hotspot domain VM according to the utilization rate of the vcpus of the hotspot domain VM and the use time occupancy rate of all the vcpus of the hotspot domain VM. [0038] The comprehensive utilization rate of the vc- PUs of the hotspot domain VM may be a product of the utilization rate of the vcpus of the hotspot domain VM and the use time occupancy rate of all the vcpus of the hotspot domain VM. [0039] In step 26, an analysis and decision module judges whether the comprehensive utilization rate of the vcpus of the hotspot domain VM exceeds an upper limit of the utilization rate. If yes, steps 27a-27b are performed, otherwise step 28 is performed. [00] In step 27a, the OS dispatcher adds an online vcpu to the hotspot domain VM when the comprehensive utilization rate of the vcpus exceeds the upper limit of the utilization rate. [0041] When the comprehensive utilization rate of the vcpus exceeds the upper limit of the utilization rate, the analysis and decision module determines to add an online vcpu to the hotspot domain VM. When it is determined to add an online vcpu to the hotspot domain VM, the OS dispatcher judges whether the number of the online vcpus of the hotspot domain VM obtained in step 22 is less than the maximum number N of the vcpus assigned to the hotspot domain VM by the VMM; if yes, the OS dispatcher changes an offline vcpu to be online according to the state of each vcpu obtained in step 23. [0042] In step 27b, the VMM incorporates the added online vcpu into the running queue of a physical CPU bound to the online vcpu, removes all the vcpus of the common domain VM from the running queue of the physical CPU, and records the removed vcpus in the affinity information of the physical CPU. After step 27b, the flow returns to step 23 to continue the execution until the virtual system is started next time. [0043] After adding an online vcpu, the OS dispatcher notifies the VMM to perform an exclusive operation on the online vcpu. The VMM determines the physical CPU bound to the added online vcpu according to the vcpu information obtained in step 23, and incorporates the online vcpu into the running queue of the bound physical CPU. Next, the VMM traverses all the vcpus of the common domain VM. The VMM removes all the vcpus of the common domain VM from the running queue of the bound physical CPU, and records the removed vcpus in the affinity information of the physical CPU, so as to recover the vcpus of the common domain VM to the running queue of the physical CPU when the vcpus of the common domain VM are subsequently allowed to be dispatched on the physical CPU. [0044] In step 28, the analysis and decision module judges whether the comprehensive utilization rate of the vcpus of the hotspot domain VM is below a lower limit of the utilization rate; if yes, the flow goes to step 29a, otherwise the OS dispatcher maintains the number of the vcpus of the hotspot domain VM, and the flow returns to step 23 to continue the execution. [004] In step 29a, when the comprehensive utilization rate of the vcpus is below the lower limit of the utilization rate, the OS dispatcher reduces an online vcpu for the hotspot domain VM. [0046] When the comprehensive utilization rate of the vcpus is below the lower limit of the utilization rate, the analysis and decision module determines to reduce an online vcpu for the hotspot domain VM. According to the state of each vcpu obtained in step 23, the OS dispatcher may select an online vcpu having the maximum serial number and perform an offline operation thereon so that the online vcpu becomes offline. Herein, the upper and lower limits of the utilization rate may be empirical values set according to different application scenarios. [0047] In step 29b, the VMM removes the offline vcpu from the running queue of the bound physical CPU, and incorporates the vcpus of the common domain VM recorded in the affinity information of the physical CPU into the running queue of the physical CPU. [0048] After step 29b, the flow returns to step 23 to continue the execution until the virtual system is started next time. [0049] With the method for dispatching CPUs of a hotspot domain VM according to the embodiment of the present invention, after one or more online vcpus are added or reduced for the hotspot domain VM, the VMM adjusts the running queue of the physical CPU bound to the online vcpu. Thus when the vcpus of the hotspot domain VM have a heavy load, the added online vcpu exclusively occupies the bound physical CPU, and when the vcpus of the hotspot domain VM have a light load, the vcpus of the common domain VM are dispatched on the physical CPU. [000] Fig. 3 is a structure diagram of a VM system according to an embodiment of the present invention. As illustrated in Fig. 3, the VM system according to the embodiment includes: a hotspot domain VM 1, a plurality of common domain VMs 2, and a VM Monitor (VMM) 3. [001] The VMM 3 is configured to bind a different physical CPU to each of vcpus of the hotspot domain VM. [002] The hotspot domain VM 1 is configured to determine a comprehensive utilization rate of the vcpus of the hotspot domain VM according to counted use information of the vcpus of the hotspot domain VM, and add or reduce at least one online vcpu according to the comprehensive utilization rate of the vcpus and the maximum number of the vcpus assigned to the hotspot domain VM; wherein the use information of the vcpus in- 6

7 11 EP B1 12 cludes the utilization rate of the vcpus of the hotspot domain VM and the use time occupancy rate of the vc- PUs of the hotspot domain VM. [003] For example, the use time occupancy rate of the vcpus of the hotspot domain VM is a ratio of a sum of the operation time of all the vcpus of the hotspot domain VM to the total time, wherein the total time is a sum of the operation time of all the vcpus and the wait time of all the vcpus. The utilization rate of the vcpus of the hotspot domain VM may be a ratio of a sum of the nonidle time of all the vcpus inside the hotspot domain VM to a sum of the operation time of all the online vcpus of the hotspot domain VM. [004] The VMM 3 is further configured to adjust the vcpus of the common domain VM 2 in the running queue of a physical CPU bound to an online vcpu after the online vcpu is added, so that the online vcpu can be dispatched on the physical CPU while the vcpus of the common domain VM cannot be dispatched on the physical CPU. [00] The VMM 3 is further configured to adjust the vcpu of the hotspot domain VM in the running queue of a physical CPU bound to an offline vcpu after the online vcpu is reduced, so that vcpus of the common domain VM can be dispatched on the physical CPU. [006] Please refer to the descriptions in the embodiment corresponding to Fig. 1 for the functions realized by the above modules, and herein are omitted. [007] Particularly, the VMM 3 includes: a first adjustment unit and a second adjustment unit. [008] The first adjustment unit is configured to incorporate the online vcpu into the running queue of the physical CPU bound to the online vcpu, remove all the vcpus of the common domain VM from the running queue of the physical CPU, and record the removed vc- PUs in the affinity information of the physical CPU. [009] The second adjustment unit is configured to remove the offline vcpu from the running queue of the bound physical CPU, and incorporate the vcpu of the common domain VM recorded in the affinity information of the physical CPU into the running queue of the physical CPU. [0060] Please refer to the descriptions in the embodiment corresponding to Fig. 1 for the functions realized by the above units, and herein are omitted. [0061] With the VM system according to the embodiment of the present invention, the hotspot domain VM dynamically adjusts the number of the online vcpus in the hotspot domain VM according to the use information of the vcpus. When the vcpus of the hotspot domain VM have a heavy load, an online vcpu is added to the hotspot domain VM, and the VMM enables the added online vcpu to exclusively occupy the bound physical CPU, so that the online vcpu of the common domain VM cannot be dispatched on the physical CPU. Thus, the vcpu of the hotspot domain VM may preferably use the physical CPU prior to the vcpu of the common domain VM, so that a response to the request from the vcpu of the hotspot domain VM can be made in time, and the phenomenon that the vcpu of the common domain VM waits for the hotspot domain VM is avoided, thereby improving the whole performance of the virtual system. When the vcpus of the hotspot domain VM have a light load, the VMM deletes an online vcpu from the hotspot domain VM, and cancels the exclusivity of the vcpu of the hotspot domain VM to the physical CPU, so that the vcpu of the common domain VM can be dispatched on physical CPU, thereby improving the performance of the common domain VM, and preventing the waste of the CPU resources. [0062] Fig. 4 is a structure diagram of another VM system according to an embodiment of the present invention. As illustrated in Fig. 4, the hotspot domain VM 1 includes a performance statistics module 11, an analysis and decision module 12 and an OS dispatcher 13. [0063] The performance statistics module 11 is configured to count the use information of the vcpus of the hotspot domain VM, and determine a comprehensive utilization rate of the vcpus of the hotspot domain VM according to the monitored use information of the vcpus of the hotspot domain VM. [0064] The analysis and decision module 12 is configured to determine to add or reduce at least one online vcpu for the hotspot domain VM according to the comprehensive utilization rate of the vcpus determined by the performance statistics module 11 and the maximum number of the vcpus assigned to the hotspot domain VM by the VM dispatcher. [006] The OS dispatcher 13 is configured to add or reduce at least one online vcpu according to the decision of the analysis and decision module 12. [0066] Particularly, the performance statistics module 11 includes a first determination unit, a second determination unit and a third determination unit. [0067] The first determination unit is configured to determine a use time occupancy rate of the vcpus of the hotspot domain VM according to a sum of the operation time and a sum of the wait time of all the vcpus of the hotspot domain VM. The second determination unit is configured to determine a utilization rate of the vcpus of the hotspot domain VM according to a sum of the nonidle time of all the vcpus inside the hotspot domain VM and a sum of the operation time of all the online vcpus of the hotspot domain VM. The third determination unit is configured to determine a comprehensive utilization rate of the vcpus of the hotspot domain VM according to the use time occupancy rate of the vcpus of the hotspot domain VM and the utilization rate of the vcpus of the hotspot domain VM. [0068] Particularly, the analysis and decision module 12 includes an online determination unit and an offline determination unit. [0069] The online determination unit is configured to determine to add one or more online vcpus for the hotspot domain VM, when the comprehensive utilization rate of the vcpus determined by the third determination 7

8 13 EP B1 14 unit exceeds an upper limit of the utilization rate while the number of the vcpus of the hotspot domain VM is not more than the maximum number of the vcpus assigned to the hotspot domain VM by the VMM. The offline determination unit is configured to determine to reduce at least one online vcpu for the hotspot domain VM, when the comprehensive utilization rate of the vcpus determined by the third determination unit is below the lower limit of the utilization rate. [0070] Please refer to the descriptions in the embodiment corresponding to Fig. 2 for the functions realized by the above modules, and herein are omitted. [0071] With the VM system according to the embodiment of the present invention, after one or more online vcpus are added or reduced for the hotspot domain VM, the VMM adjusts the running queue of the physical CPU bound to the online vcpu. Thus when the vcpus of the hotspot domain VM have a heavy load, the added online vcpu exclusively occupies the bound physical CPU, and when the vcpus of the hotspot domain VM have a light load, the vcpus of the common domain VM are dispatched on the physical CPU. [0072] A person skilled in the art will appreciate that all or a part of steps for implementing the above method embodiments may be completed by instructing relevant hardware through a program that may be stored in a computer readable storage medium, and when being executed, the program performs the steps including the above method embodiments. The storage medium may include various mediums capable of storing program codes, such as ROM, RAM, magnetic disk and optical disk. [0073] Finally to be noted, the above embodiments are just used to describe, rather than limiting, the technical solutions of the present invention. Although the present invention is detailedly described with reference to the aforementioned embodiments, a person skilled in the art shall be appreciated that the technical solutions of the aforementioned embodiments can be modified, or some technical features thereof can be equivalently replaced. Those modifications or replacements do not make essences of corresponding technical solutions deviate from the scope of the technical solutions of each embodiment of the present invention. Claims 1. A method for dispatching Central Processing Units, CPUs, of a hotspot domain Virtual Machine, VM, wherein each of virtual CPUs, v CPUs, of the hotspot domain VM is bound to a different physical CPU, the method comprising: determining a comprehensive utilization rate of the vcpus of the hotspot domain VM according to counted use information of the vcpus of the hotspot domain VM, wherein the use information of the vcpus comprises a utilization rate of the vcpus of the hotspot domain VM and a use time occupancy rate of the vcpus of the hotspot domain VM, wherein the use time occupancy rate of the vcpus of the hotspot domain VM is a ratio of a sum of the operation time of all the vcpus of the hotspot domain VM to a total time, the total time is a sum of the operation time of all the vcpus and the wait time of all the vcpus, and the utilization rate of the vcpus of the hotspot domain VM is a ratio of a sum of the non-idle time of all the vcpus inside the hotspot domain VM to a sum of the operation time of all the online vcpus of the hotspot domain VM; determining to add or reduce at least one online vcpu for the hotspot domain VM, according to the comprehensive utilization rate of the vcpus and the maximum number of the vcpus assigned to the hotspot domain VM; after an online vcpu is added, adjusting the vc- PUs of a common domain VM in the running queue of a physical CPU bound to the online vcpu, so that the online vcpu can be dispatched on the physical CPU while the vcpus of the common domain VM cannot be dispatched on the physical CPU; or after an online vcpu is reduced, adjusting the vcpu of the hotspot domain VM in the running queue of a physical CPU bound to the offline vcpu, so that the vcpu of the common domain VM can be dispatched on the physical CPU. 2. The method according to claim 1, wherein adjusting the vcpus of the common domain VM in the running queue of the physical CPU bound to the online vcpu, so that the online vcpu can be dispatched on the physical CPU while the vcpus of the common domain VM cannot be dispatched on the physical CPU comprises: incorporating, by a VM dispatcher, the online vcpu into the running queue of the physical CPU bound to the online vcpu; and removing, by the VM dispatcher, all the vcpus of the common domain VM from the running queue of the physical CPU, and recording the removed vcpus in the affinity information of the physical CPU. 3. The method according to claim 1 or 2, wherein adjusting the vcpu of the hotspot domain VM in the running queue of the physical CPU bound to the offline vcpu, so that the vcpu of the common domain VM can be dispatched on the physical CPU comprises: removing, by a VM dispatcher, the offline vcpu from the running queue of the bound physical CPU; and 8

9 1 EP B1 16 incorporating, by the VM dispatcher, the vcpus of the common domain VM recorded in the affinity information of the physical CPU into the running queue of the physical CPU. 4. The method according to claim 1, 2 or 3, wherein determining the comprehensive utilization rate of the vcpus of the hotspot domain VM according to the counted use information of the vcpus of the hotspot domain VM comprises: determining the use time occupancy rate of the vcpus of the hotspot domain VM according to a sum of operation time and a sum of wait time of all the vcpus of the hotspot domain VM; determining the utilization rate of the vcpus of the hotspot domain VM according to a sum of non-idle time of all the vcpus inside the hotspot domain VM and a sum of operation time of all the online vcpus of the hotspot domain VM; and determining the comprehensive utilization rate of the vcpus of the hotspot domain VM according to the use time occupancy rate of the vcpus of the hotspot domain VM and the utilization rate of the vcpus of the hotspot domain VM.. The method according to claim 1, 2, 3 or 4, wherein determining to add or reduce at least one online vcpu for the hotspot domain VM, according to the comprehensive utilization rate of the vcpus and the maximum number of the vcpus assigned to the hotspot domain VM comprises: determining to add at least one online vcpu for the hotspot domain VM, when the comprehensive utilization rate of the vcpus exceeds an upper limit of the utilization rate while the number of the vcpus of the hotspot domain VM is not more than the maximum number of the vcpus assigned to the hotspot domain VM by a VM dispatcher; or determining to reduce at least one online vcpu for the hotspot domain VM, when the comprehensive utilization rate of the vcpus is below a lower limit of the utilization rate. 6. A Virtual Machine, VM, system, comprising a hotspot domain VM, a common domain VM and a VM Monitor, VMM, wherein: the VMM is configured to bind a different physical CPU to each of virtual CPUs, v CPUs, of the hotspot domain VM; the hotspot domain VM is configured to determine a comprehensive utilization rate of the vc- PUs of the hotspot domain VM according to counted use information of the vcpus of the hotspot domain VM, and determine to add or reduce at least one online vcpu for the hotspot domain VM according to the comprehensive utilization rate of the vcpus and the maximum number of the vcpus assigned to the hotspot domain VM; wherein the use information of the vcpus comprises a utilization rate of the vcpus of the hotspot domain VM and a use time occupancy rate of the vcpus of the hotspot domain VM, wherein the use time occupancy rate of the vcpus of the hotspot domain VM is a ratio of a sum of the operation time of all the vcpus of the hotspot domain VM to a total time, the total time is a sum of the operation time of all the vcpus and the wait time of all the vcpus, and the utilization rate of the vcpus of the hotspot domain VM is a ratio of a sum of the non-idle time of all the vcpus inside the hotspot domain VM to a sum of the operation time of all the online vcpus of the hotspot domain VM; the VMM is further configured to adjust the vc- PUs of the common domain VM in the running queue of a physical CPU bound to an online vcpu after the online vcpu is added, so that the online vcpu can be dispatched on the physical CPU while the vcpus of the common domain VM cannot be dispatched on the physical CPU; and the VMM is further configured to adjust the vcpu of the hotspot domain VM in the running queue of a physical CPU bound to an offline vcpu after the online vcpu is reduced, so that vcpus of the common domain VM can be dispatched on the physical CPU. 7. The system according to claim 6, wherein the hotspot domain VM comprises a performance statistics module, an analysis and decision module and an operation system, O S, dispatcher; wherein the performance statistics module is configured to count use information of the vcpus of the hotspot domain VM, and determine a comprehensive utilization rate of the vcpus of the hotspot domain VM according to the monitored use information of the vcpus of the hotspot domain VM; the analysis and decision module is configured to determine to add or reduce at least one online vcpu for the hotspot domain VM according to the comprehensive utilization rate of the vcpus and the maximum number of the vcpus assigned to the hotspot domain VM by a VM dispatcher; and the OS dispatcher is configured to add or reduce at least one online vcpu according to the decision of the analysis and decision module. 8. The system according to claim 7, wherein the performance statistics module comprises: a first determination unit configured to determine 9

10 17 EP B1 18 the use time occupancy rate of the vcpus of the hotspot domain VM according to a sum of operation time and a sum of wait time of all the vcpus of the hotspot domain VM; a second determination unit configured to determine the utilization rate of the vcpus of the hotspot domain VM according to a sum of nonidle time of all the vcpus inside the hotspot domain VM and a sum of operation time of all the online vcpus of the hotspot domain VM; and a third determination unit configured to determine the comprehensive utilization rate of the vcpus of the hotspot domain VM according to the use time occupancy rate of the vcpus of the hotspot domain VM and the utilization rate of the vcpus of the hotspot domain VM. 9. The system according to claim 7, wherein the analysis and decision module comprises: an online determination unit configured to determine to add one or more online vcpus for the hotspot domain VM, when the comprehensive utilization rate of the vcpus exceeds an upper limit of the utilization rate while the number of the vcpus of the hotspot domain VM is not more than the maximum number of the vcpus assigned to the hotspot domain VM by the VMM; and an offline determination unit configured to determine to reduce at least one online vcpu for the hotspot domain VM, when the comprehensive utilization rate of the vcpus is below a lower limit of the utilization rate.. The system according to claim 6 or 7, wherein the VMM comprises: a first adjustment unit configured to incorporate the online vcpu into the running queue of a physical CPU bound to the online vcpu, remove all the vcpus of the common domain VM from the running queue of the physical CPU, and record the removed vcpus in the affinity information of the physical CPU; and a second adjustment unit configured to remove the offline vcpu from the running queue of the bound physical CPU, and incorporate the vc- PUs of the common domain VM recorded in the affinity information of the physical CPU into the running queue of the physical CPU. Patentansprüche 1. Verfahren zum Verteilen von Zentraleinheiten (Central Processing Units, CPUs) einer virtuellen Maschine einer Hotspotdomäne, Hotspotdomänen-VM, wobei jede der virtuellen CPUs, vcpus, der Hotspotdomänen-VM an eine andere physikalische CPU gebunden ist, wobei das Verfahren Folgendes umfasst: Bestimmen einer Gesamtauslastungsrate der vcpus der Hotspotdomänen-VM gemäß gezählten Nutzungsinformationen der vcpus der Hotspotdomänen-VM, wobei die Nutzungsinformationen der vcpus eine Auslastungsrate der vcpus der Hotspotdomänen-VM und eine Nutzungszeitbelegungsrate der vcpus der Hotspotdomänen-VM umfassen, wobei die Nutzungszeitbelegungsrate der vc- PUs der Hotspotdomänen-VM ein Verhältnis einer Summe der Betriebszeit aller vcpus der Hotspotdomänen-VM zu einer Gesamtzeit ist, die Gesamtzeit eine Summe der Betriebszeit aller vcpus und der Wartezeit aller vcpus ist, und die Auslastungsrate der vcpus der Hotspotdomänen-VM ein Verhältnis einer Summe der Nichtleerlaufzeit aller vcpus innerhalb der Hotspotdomänen-VM zu einer Summe der Betriebszeit aller Online-vCPUs der Hotspotdomänen-VM ist; Bestimmen, wenigstens eine Online-vCPU für die Hotspotdomänen-VM hinzuzufügen oder zu reduzieren, gemäß der gesamten Auslastungsrate der vcpus und der maximalen Anzahl von vcpus, die der Hotspotdomänen-VM zugeordnet sind; nachdem eine Online-vCPU hinzugefügt worden ist, Anpassen der vcpus einer VM einer gemeinsamen Domäne in der laufenden Warteschlange einer physikalischen CPU, die an die Online-vCPU gebunden ist, so dass die OnlinevCPU auf der physikalischen CPU verteilt werden kann, während die vcpus der VM der gemeinsamen Domäne nicht auf der physikalischen CPU verteilt werden können; oder nachdem eine Online-vCPU reduziert worden ist, Anpassen der vcpu der Hotspotdomänen- VM in der laufenden Warteschlange einer physikalischen CPU, die an die Offline-vCPU gebunden ist, so dass die vcpu der VM der gemeinsamen Domäne auf der physikalischen CPU verteilt werden kann. 2. Verfahren nach Anspruch 1, wobei das Anpassen der vcpus der VM der gemeinsamen Domäne in der laufenden Warteschlange der physikalischen CPU, die an die Online-vCPU gebunden ist, so dass die Online-vCPU auf der physikalischen CPU verteilt werden kann, während die vcpus der VM der gemeinsamen Domäne nicht auf der physikalischen CPU verteilt werden können, Folgendes umfasst: Einfügen durch eine VM-Verteilungseinheit der Online-vCPU in die laufende Warteschlange der

11 19 EP B1 physikalischen CPU, die an die Online-vCPU gebunden ist; und Entfernen durch die VM-Verteilungseinheit aller vcpus der VM der gemeinsamen Domäne aus der laufenden Warteschlange der physikalischen CPU und Aufzeichnen der entfernten vc- PUs in den Affinitätsinformationen der physikalischen CPU. 3. Verfahren nach Anspruch 1 oder 2, wobei das Anpassen der vcpu der Hotspotdomänen-VM in der laufenden Warteschlange der physikalischen CPU, die an die Offline-vCPU gebunden ist, so dass die vcpu der VM der gemeinsamen Domäne auf der physikalischen CPU verteilt werden kann, Folgendes umfasst: Entfernen durch eine VM-Verteilungseinheit der Offline-vCPU aus der laufenden Warteschlange der gebundenen physikalischen CPU; und Einfügen durch die VM-Verteilungseinheit der vcpus der VM der gemeinsamen Domäne, die in den Affinitätsinformationen der physikalischen CPU aufgezeichnet sind, in die laufende Warteschlange der physikalischen CPU. 4. Verfahren nach Anspruch 1, 2 oder 3, wobei das Bestimmen der Gesamtauslastungsrate der vcpus der Hotspotdomänen-VM gemäß den gezählten Nutzungsinformationen der vcpus der Hotspotdomänen-VM Folgendes umfasst: Bestimmen der Nutzungszeitbelegungsrate der vcpus der Hotspotdomänen-VM gemäß einer Summe der Betriebszeit und einer Summe der Wartezeit aller vcpus der Hotspotdomänen- VM; Bestimmen der Auslastungsrate der vcpus der Hotspotdomänen-VM gemäß einer Summe der Nichtleerlaufzeit aller vcpus innerhalb der Hotspotdomänen-VM und einer Summe der Betriebszeit aller Online-vCPUs der Hotspotdomänen-VM; und Bestimmen der Gesamtauslastungsrate der vc- PUs der Hotspotdomänen-VM gemäß der Nutzungszeitbelegungsrate der vcpus der Hotspotdomänen-VM und der Auslastungsrate der vcpus der Hotspotdomänen-VM.. Verfahren nach Anspruch 1, 2, 3 oder 4, wobei das Bestimmen, wenigstens eine Online-vCPU für die Hotspotdomänen-VM hinzuzufügen oder zu reduzieren, gemäß der Gesamtauslastungsrate der vc- PUs und der maximalen Anzahl von vcpus, die der Hotspotdomänen-VM zugeordnet sind, Folgendes umfasst: Bestimmen, wenigstens eine Online-vCPU für die Hotspotdomänen-VM hinzuzufügen, wenn die Gesamtauslastungsrate der vcpus eine Obergrenze der Auslastungsrate übersteigt, während die Anzahl der vcpus der Hotspotdomänen-VM nicht größer ist als die maximale Anzahl der vcpus, die der Hotspotdomänen-VM durch eine VM-Verteilungseinheit zugewiesen sind; oder Bestimmen, wenigstens eine Online-vCPU für die Hotspotdomänen-VM zu reduzieren, wenn die Gesamtauslastungsrate der vcpus unterhalb einer Untergrenze der Auslastungsrate ist. 6. System virtueller Maschinen, VM-System, das eine Hotspotdomänen-VM, eine VM einer gemeinsamen Domäne und eine VM-Überwachungseinheit (VM Monitor, VMM) umfasst, wobei: die VMM konfiguriert ist, eine verschiedene physikalische CPU an jede der virtuellen CPUs, vc- PUs, der Hotspotdomänen-VM zu binden; die Hotspotdomänen-VM konfiguriert ist, eine Gesamtauslastungsrate der vcpus der Hotspotdomänen-VM gemäß gezählten Nutzungsinformationen der vcpus der Hotspotdomänen- VM zu bestimmen und zu bestimmen, wenigstens eine Online-vCPU für die Hotspotdomänen-VM gemäß der Gesamtauslastungsrate der vcpus und der maximalen Anzahl der vcpus, die der Hotspotdomänen-VM zugewiesen sind, hinzuzufügen oder zu reduzieren; wobei die Nutzungsinformationen der vcpus eine Auslastungsrate der vcpus der Hotspotdomänen-VM und eine Nutzungszeitbelegungsrate der vc- PUs der Hotspotdomänen-VM umfassen, wobei die Nutzungszeitbelegungsrate der vcpus der Hotspotdomänen-VM ein Verhältnis einer Summe der Betriebszeit aller vcpus der Hotspotdomänen-VM zu einer Gesamtzeit ist, die Gesamtzeit eine Summe der Betriebszeit aller vcpus und der Wartezeit aller vcpus ist, und die Auslastungsrate der vcpus der Hotspotdomänen- VM ein Verhältnis einer Summe der Nichtleerlaufzeit aller vcpus innerhalb der Hotspotdomänen-VM zu einer Summe der Betriebszeit aller Online-vCPUs der Hotspotdomänen-VM ist; die VMM ferner konfiguriert ist, die vcpus der VM der gemeinsamen Domäne in der laufenden Warteschlange einer physikalischen CPU, die an eine Online-vCPU gebunden ist, anzupassen, nachdem die vcpu hinzugefügt worden ist, so dass die Online-vCPU auf der physikalischen CPU verteilt werden kann, während die vcpus der VM der allgemeinen Domäne nicht auf der physikalischen CPU verteilt werden können; und die VMM ferner konfiguriert ist, die vcpu der 11

12 21 EP B1 22 Hotspotdomänen-VM in der laufenden Warteschlange einer physikalischen CPU, die an eine Offline-vCPU angebunden ist, anzupassen, nachdem die Online-vCPU reduziert worden ist, so dass vcpus der gemeinsamen Domänen- VM auf der physikalischen CPU verteilt werden können. 7. System nach Anspruch 6, wobei die Hotspotdomänen-VM ein Leistungsstatistikmodul, ein Analyseund Entscheidungsmodul und eine Betriebssystem- Verteilereinheit, OS-Verteilereinheit, umfasst; wobei das Leistungsstatistikmodul konfiguriert ist, die Nutzungsinformationen der vcpus der Hotspotdomänen-VM zu zählen und eine Gesamtauslastungsrate der vcpus der Hotspotdomänen-VM gemäß den überwachten Nutzungsinformationen der vcpus der Hotspotdomänen-VM zu bestimmen; das Analyse- und Entscheidungsmodul konfiguriert ist zu bestimmen, wenigstens eine Online-vCPU für die Hotspotdomänen-VM gemäß der Gesamtauslastungsrate der vcpus und der maximalen Anzahl der vcpus, die der Hotspotdomänen-VM durch eine VM-Verteilereinheit zugewiesen sind, hinzuzufügen oder zu reduzieren; und die OS-Verteilereinheit konfiguriert ist, wenigstens eine Online-vCPU gemäß der Entscheidung des Analyse- und Entscheidungsmoduls hinzuzufügen oder zu reduzieren. 8. System nach Anspruch 7, wobei das Leistungsstatistikmodul Folgendes umfasst: eine erste Bestimmungseinheit, die konfiguriert ist, die Nutzungszeitbelegungsrate der vcpus der Hotspotdomänen-VM gemäß einer Summe der Betriebszeit und einer Summe der Wartezeit aller vcpus der Hotspotdomänen-VM zu bestimmen; eine zweite Bestimmungseinheit, die konfiguriert ist, die Auslastungsrate der vcpus der Hotspotdomänen-VM gemäß einer Summe der Nichtleerlaufzeit aller vcpus innerhalb der Hotspotdomänen-VM und einer Summe der Betriebszeit aller Online-vCPUs der Hotspotdomänen-VM zu bestimmen; und eine dritte Bestimmungseinheit, die konfiguriert ist, die Gesamtauslastungsrate der vcpus der Hotspotdomänen-VM gemäß der Nutzungszeitbelegungsrate der vcpus der Hotspotdomänen-VM und der Auslastungsrate der vcpus der Hotspotdomänen-VM zu bestimmen. 9. System nach Anspruch 7, wobei das Analyse- und Entscheidungsmodul Folgendes umfasst: eine Online-Bestimmungseinheit, die konfiguriert ist zu bestimmen, eine oder mehrere Online-vCPUs für die Hotspotdomänen-VM hinzuzufügen, wenn die Gesamtauslastungsrate der vcpus eine Obergrenze der Auslastungsrate übersteigt, während die Anzahl der vcpus der Hotspotdomänen-VM nicht größer ist als die maximale Anzahl der vcpus, die der Hotspotdomänen-VM durch die VMM zugewiesen sind; und eine Offline-Bestimmungseinheit, die konfiguriert ist zu bestimmen, wenigstens eine OnlinevCPU für die Hotspotdomänen-VM zu reduzieren, wenn die Gesamtauslastungsrate der vc- PUs unterhalb einer Untergrenze der Auslastungsrate ist.. System nach Anspruch 6 oder 7, wobei die VMM Folgendes umfasst: eine erste Anpassungseinheit, die konfiguriert ist, die Online-vCPU in die laufende Warteschlange einer physikalischen CPU, die an die Online-vCPU gebunden ist, einzufügen, alle vc- PUs der gemeinsamen Domänen-VM aus der laufenden Warteschlange der physikalischen CPU zu entfernen und die entfernten vcpus in den Affinitätsinformationen der physikalischen CPU aufzuzeichnen; und eine zweite Anpassungseinheit, die konfiguriert ist, die Offline-vCPU aus der laufenden Warteschlange der gebundenen physikalischen CPU zu entfernen und die vcpus der gemeinsamen Domänen-VM, die in den Affinitätsinformationen der physikalischen CPU aufgezeichnet sind, in die laufende Warteschlange der physikalischen CPU einzufügen. Revendications 1. Procédé pour la distribution d Unités Centrales de Traitement, CPU (Central Processing Units), d une machine virtuelle, VM (Virtual Machine), de domaine de point d accès sans fil, dans lequel chacune des CPU virtuelles, vcpu (virtual CPU), de la VM de domaine de point d accès sans fil est liée à une CPU physique différente, le procédé comprenant : la détermination d un taux global d utilisation des vcpu de la VM de domaine de point d accès sans fil conformément à des informations d utilisation comptées des vcpu de la VM de domaine de point d accès sans fil, dans lequel les informations d utilisation des vcpu comprennent un taux d utilisation des vcpu de la VM de domaine de point d accès sans fil et un taux d occupation de temps d utilisation des vcpu de la VM de domaine de point d accès sans fil, dans lequel le taux d occupation de temps d utilisation 12

13 23 EP B1 24 des vcpu de la VM de domaine de point d accès sans fil est un rapport d une somme des temps de fonctionnement de toutes les vcpu de la VM de domaine d accès sans fil à un temps total, le temps total est une somme des temps de fonctionnement de toutes les vcpu et des temps d attente de toutes les vcpu, et le taux d utilisation des vcpu de la VM de domaine de point d accès sans fil est un rapport d une somme des temps de non-inactivité de toutes les vcpu se trouvant dans la VM de domaine de point d accès sans fil à une somme des temps de fonctionnement de toutes les vcpu en ligne de la VM de domaine de point d accès sans fil ; la détermination d ajouter ou de réduire au moins une vcpu en ligne pour la VM de domaine de point d accès sans fil, conformément au taux global d utilisation des vcpu et au nombre maximum des vcpu affectées à la VM de domaine de point d accès sans fil ; après qu une vcpu en ligne a été ajoutée, l ajustement des vcpu d une VM de domaine commun dans la file d attente d exécution d une CPU physique liée à la vcpu en ligne, de manière à ce que la vcpu en ligne puisse être distribuée sur la CPU physique tandis que les vcpu de la VM de domaine commun ne peuvent pas être distribuées sur la CPU physique ; ou après qu une vcpu en ligne a été réduite, l ajustement de la vcpu de la VM de domaine de point d accès sans fil dans la file d attente d exécution d une CPU physique liée à la vcpu hors ligne, de manière à ce que la vcpu de la VM de domaine commun puisse être distribuée sur la CPU physique. 2. Procédé selon la revendication 1, dans lequel l ajustement des vcpu de la VM de domaine commun dans la file d attente d exécution de la CPU physique liée à la vcpu en ligne, de manière à ce que la vcpu en ligne puisse être distribuée sur la CPU physique tandis que les vcpu de la VM de domaine commun ne peuvent pas être distribuées sur la CPU physique, comprend : l incorporation, par un distributeur de VM, de la vcpu en ligne dans la file d attente d exécution de la CPU physique liée à la vcpu en ligne ; et la suppression, par le distributeur de VM, de toutes les vcpu de la VM de domaine commun de la file d attente d exécution de la CPU physique, et l enregistrement des vcpu supprimées dans les informations d affinité de la CPU physique. 3. Procédé selon la revendication 1 ou 2, dans lequel l ajustement de la vcpu de la VM de domaine de point d accès sans fil dans la file d attente d exécution de la CPU physique liée à la vcpu hors ligne, de manière à ce que la vcpu de la VM de domaine commun puisse être distribuée sur la CPU physique, comprend : la suppression, par un distributeur de VM, des vcpu hors ligne de la file d attente d exécution de la CPU physique liée ; et l incorporation, par le distributeur de VM, des vcpu de la VM de domaine commun enregistrée dans les informations d affinité de la CPU physique dans la file d attente d exécution de la CPU physique. 4. Procédé selon la revendication 1, 2 ou 3, dans lequel la détermination du taux global d utilisation des vc- PU de la VM de domaine de point d accès sans fil conformément aux informations d utilisation comptées des vcpu de la VM de domaine de point d accès sans fil comprend : la détermination du taux d occupation de temps d utilisation des vcpu de la VM de domaine de point d accès sans fil conformément à une somme de temps de fonctionnement et à une somme de temps d attente de toutes les vcpu de la VM de domaine de point d accès sans fil ; la détermination du taux d utilisation des vcpu de la VM de domaine de point d accès sans fil conformément à une somme de temps de noninactivité de toutes les vcpu se trouvant dans la VM de domaine de point d accès sans fil et à une somme de temps de fonctionnement de toutes les vcpu en ligne de la VM de domaine de point d accès sans fil ; et la détermination du taux global d utilisation des vcpu de la VM de domaine de point d accès sans fil conformément au taux d occupation de temps d utilisation des vcpu de la VM de domaine de point d accès sans fil et au taux d utilisation des vcpu de la VM de domaine de point d accès sans fil.. Procédé selon la revendication 1, 2, 3 ou 4, dans lequel la détermination d ajouter ou de réduire au moins une vcpu en ligne pour la VM de domaine de point d accès sans fil, conformément au taux global d utilisation des vcpu et au nombre maximum des vcpu affectées à la VM de domaine de point d accès sans fil comprend : la détermination d ajouter au moins une vcpu en ligne pour la VM de domaine de point d accès sans fil, lorsque le taux global d utilisation des vcpu dépasse une limite supérieure du taux d utilisation alors que le nombre des vcpu de la VM de domaine de point d accès sans fil n est pas supérieur au nombre maximum des vcpu affectées à la VM de domaine de point d accès 13

14 2 EP B1 26 sans fil par un distributeur de VM ; ou la détermination de réduire au moins une vcpu en ligne pour la VM de domaine de point d accès sans fil, lorsque le taux global d utilisation des vcpu est inférieur à une limite inférieure du taux d utilisation. 6. Système de Machine Virtuelle, VM, comprenant une VM de domaine de point d accès sans fil, une VM de domaine commun et un moniteur de VM, VMM (Virtual Machine Monitor), dans lequel : le VMM est configuré pour lier une CPU physique différente à chacune des CPU virtuelles, vc- PU, de la VM de domaine de point d accès sans fil ; la VM de domaine de point d accès sans fil est configurée pour déterminer un taux global d utilisation des vcpu de la VM de domaine de point d accès sans fil conformément à des informations d utilisation comptées des vcpu de la VM de domaine de point d accès sans fil, et déterminer d ajouter ou de réduire au moins une vc- PU en ligne pour la VM de domaine de point d accès sans fil conformément au taux global d utilisation des vcpu et au nombre maximum des vcpu affectées à la VM de domaine de point d accès sans fil ; dans lequel les informations d utilisation des vc- PU comprennent un taux d utilisation des vcpu de la VM de domaine de point d accès sans fil et un taux d occupation de temps d utilisation des vcpu de la VM de domaine de point d accès sans fil, dans lequel le taux d occupation de temps d utilisation des vcpu de la VM de domaine de point d accès sans fil est un rapport d une somme des temps de fonctionnement de toutes les vcpu de la VM de domaine de point d accès sans fil à un temps total, le temps total est une somme des temps de fonctionnement de toutes les vcpu et des temps d attente de toutes les vcpu, et le taux d utilisation des vcpu de la VM de domaine de point d accès sans fil est un rapport d une somme des temps de non-inactivité de toutes les vcpu se trouvant dans la VM de domaine de point d accès sans fil à une somme des temps de fonctionnement de toutes les vc- PU en ligne de la VM de domaine de point d accès sans fil ; le VMM est en outre configuré pour ajuster les vcpu de la VM de domaine commun dans la file d attente d exécution d une CPU physique liée à une vcpu en ligne après que la vcpu en ligne a été ajoutée, de manière à ce que la vcpu en ligne puisse être distribuée sur la CPU physique tandis que les vcpu de la VM de domaine commun ne peuvent pas être distribuées sur la CPU physique ; et le VMM est en outre configuré pour ajuster la vcpu de la VM de domaine de point d accès sans fil dans la file d attente d exécution d une CPU physique liée à une vcpu hors ligne après que la vcpu en ligne a été réduite, de manière à ce que des vcpu de la VM de domaine commun puissent être distribuées sur la CPU physique. 7. Système selon la revendication 6, dans lequel la VM de domaine de point d accès sans fil comprend un module de statistiques de performances, un module d analyse et de décision et un distributeur de système d exploitation, OS (Operating System) ; dans lequel le module de statistiques de performances est configuré pour compter des informations d utilisation des vcpu de la VM de domaine de point d accès sans fil, et déterminer un taux global d utilisation des vcpu de la VM de domaine de point d accès sans fil conformément aux informations surveillées d utilisation des vcpu de la VM de domaine de point d accès sans fil; le module d analyse et de décision est configuré pour déterminer d ajouter ou de réduire au moins une vc- PU en ligne pour la VM de domaine de point d accès sans fil conformément au taux global d utilisation des vcpu et au nombre maximum des vcpu affectées à la VM de domaine de point d accès sans fil par un distributeur de VM ; et le répartiteur d OS est configuré pour ajouter ou réduire au moins une vcpu en ligne conformément à la décision du module d analyse et de décision. 8. Système selon la revendication 7, dans lequel le module de statistiques de performances comprend : une première unité de détermination configurée pour déterminer le taux d occupation de temps d utilisation des vcpu de la VM de domaine de point d accès sans fil conformément à une somme des temps de fonctionnement et à une somme des temps d attente de toutes les vcpu de la VM de domaine de point d accès sans fil ; une deuxième unité de détermination configurée pour déterminer le taux d utilisation des vc- PU de la VM de domaine de point d accès sans fil conformément à une somme des temps de non-inactivité de toutes les vcpu se trouvant dans la VM de domaine de point d accès sans fil et à une somme de temps de fonctionnement de toutes les vcpu en ligne de la VM de domaine de point d accès sans fil ; et une troisième unité de détermination configurée pour déterminer le taux global d utilisation des vcpu de la VM de domaine de point d accès sans fil conformément au taux d occupation de temps d utilisation des vcpu de la VM de do- 14

15 27 EP B1 28 maine de point d accès sans fil et au taux d utilisation des vcpu de la VM de domaine de point d accès sans fil. 9. Système selon la revendication 7, dans lequel le module d analyse et de décision comprend : une unité de détermination en ligne configurée pour déterminer d ajouter une ou plusieurs vc- PU en ligne pour la VM de domaine de point d accès sans fil, lorsque le taux global d utilisation des vcpu dépasse une limite supérieure du taux d utilisation tandis que le nombre des vcpu de la VM de domaine de point d accès sans fil n est pas supérieur au nombre maximum des vcpu affectées à la VM de domaine de point d accès sans fil par le VMM ; et une unité de détermination hors ligne configurée pour déterminer de réduire au moins une vcpu en ligne pour la VM de domaine de point d accès sans fil, lorsque le taux global d utilisation des vcpu est inférieur à une limite inférieure du taux d utilisation. 1. Système selon la revendication 6 ou 7, dans lequel le VMM comprend : 2 une première unité de réglage configurée pour incorporer la vcpu en ligne à la file d attente d exécution d une CPU physique liée à la vcpu en ligne, supprimer toutes les vcpu de la VM de domaine commun de la file d attente d exécution de la CPU physique, et enregistrer les vcpu supprimées dans les informations d affinité de la CPU physique ; et une deuxième unité de réglage configurée pour supprimer la vcpu hors ligne de la file d attente d exécution de la CPU physique liée, et incorporer les vcpu de la VM de domaine commun enregistrée dans les informations d affinité de la CPU physique dans la file d attente d exécution de la CPU physique

16 EP B1 16

17 EP B1 17

18 EP B1 18

19 EP B1 REFERENCES CITED IN THE DESCRIPTION This list of references cited by the applicant is for the reader s convenience only. It does not form part of the European patent document. Even though great care has been taken in compiling the references, errors or omissions cannot be excluded and the EPO disclaims all liability in this regard. Patent documents cited in the description US A1 [000] 19