CN102591715A - Implementing method for optimizing network performance of virtual machine by using multiqueue technology - Google Patents
Implementing method for optimizing network performance of virtual machine by using multiqueue technology Download PDFInfo
- Publication number
- CN102591715A CN102591715A CN2012100021719A CN201210002171A CN102591715A CN 102591715 A CN102591715 A CN 102591715A CN 2012100021719 A CN2012100021719 A CN 2012100021719A CN 201210002171 A CN201210002171 A CN 201210002171A CN 102591715 A CN102591715 A CN 102591715A
- Authority
- CN
- China
- Prior art keywords
- vhost
- modification
- virtual
- queue
- data structure
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 29
- 238000005516 engineering process Methods 0.000 title claims abstract description 20
- 238000012986 modification Methods 0.000 claims abstract description 47
- 230000004048 modification Effects 0.000 claims abstract description 47
- 230000005540 biological transmission Effects 0.000 claims abstract description 19
- 230000015572 biosynthetic process Effects 0.000 claims description 90
- 238000005755 formation reaction Methods 0.000 claims description 89
- 238000005457 optimization Methods 0.000 claims description 8
- 230000008859 change Effects 0.000 claims description 7
- 238000004140 cleaning Methods 0.000 claims description 6
- 238000013461 design Methods 0.000 abstract description 8
- 230000006870 function Effects 0.000 description 13
- 238000004088 simulation Methods 0.000 description 7
- 230000008569 process Effects 0.000 description 5
- 238000012545 processing Methods 0.000 description 4
- 230000002093 peripheral effect Effects 0.000 description 3
- 238000013519 translation Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011833 dog model Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012806 monitoring device Methods 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Abstract
The invention relates to an implementing method for optimizing the network performance of a virtual machine by using a multiqueue technology, which comprises the following three steps of: a first step of modifying a network initialization part of QEMU and increasing support for multiqueue by modifying the QEMU; a step 2 of carrying out modification on the vhost multiqueue for supporting the QEMU to use the multiqueue and supporting a vhost-net multiqueue network card, which comprises modification of using one thread to carry out data transmission for one queue and modification of system call; and a step 3 of modifying the part, i.e. the vhost-net multiqueue network card, in a vhost module, which is related to the network, so that the virtual network card supports the multiqueue transmission. According to the invention, by designing and implementing a plurality of virtual queues from the virtual machine to a host, the aim of increasing the network data traffic and the throughput of the virtual machine is fulfilled. The method has ingenious, scientific and reasonable design and has high using value and wide application prospect in the technical field of computers.
Description
(1) technical field
The present invention adopts the many formations Intel Virtualization Technology based on virtio, and the IO that relates in the computer virtualized technology is virtual.Relate in particular to a kind of implementation method of using the virtual machine network performance optimization of many queue technologies, it is to adopt many formations method to realize mutual implementation method between virtual machine and the host in a kind of virtual network, belongs to field of computer technology.
(2) background technology
Just on large scale computer, realized commercial virtual machine as far back as eighties of last century IBM Corporation at the beginning of the seventies at the end of the sixties, many document descriptions the virtual machine that uses at OS 360 of IBM realization and classical virtual-machine architecture and described and utilized monitor of virtual machine to realize virtualized method.The researchist has summed up third-generation computer can virtual property (Virtualizability): point out and can virtual property mainly depend on the existing way that computer instruction is concentrated responsive instruction set.Summed up simultaneously can virtual property two important theorems:
Theorem one:,, just can on this computing machine, create a monitor of virtual machine if the responsive instruction set of this computing machine is the subclass of privileged instruction collection for the third-generation computer on any ordinary meaning.
Theorem two:,, just can on this computing machine, create the monitor of virtual machine of hybrid (hybrid) if the responsive instruction set of the user of this computing machine is the subclass of privileged instruction collection for the third-generation computer on any ordinary meaning.
Because most computers does not satisfy the requirement of theorem one, so some documents have loosened requirement: if meet the requirement of theorem two just can utilize be absorbed in-simulated mode (trap-and-emulate) completion is virtual.Can virtualized key be processing to the nonprivileged instruction of sensitivity.Several kinds of different virtual modes below the different disposal mode of this instruction caused.
1) based on fully virtualized (the Full Virtualization with Binary Translation) of binary translation.The dynamic scans binary code of this mode replaces with sensitivity instruction and jumps to corresponding simulation routine.Avoid the generation of responsive instruction in this way.VMware company can provide the commercialization solution based on this mode.This is that at present unique special hardware that under IA32/Intel 64 architectures, do not need is assisted the fully virtualized mode of realization.But also there is obvious defects in this scheme: owing to will monitor dynamically and the instant translation binary code, can lose on the performance.
2) auxiliary virtual (the Hardware Assisted Virtualization) of hardware.This mode is accomplished traditional virtual machine monitoring device (Virtual Machine Monitor, VMM) needed being absorbed in and the state switching through expansion instruction set through hardware.Can greatly simplify design of virtual machine in this way.But because hardware can be preserved the entire process device when state switches state and switch address space, the expense of bringing thus can not ignore, particularly under virtual machine and the frequent situation about switching of monitor of virtual machine.What patent of the present invention adopted is exactly that this mode realizes VMM.
3) accurate virtual (Para-virtualization).This mode shows cooperation way completion virtualization policy through static rewriting client operating system source code through client operating system and monitor of virtual machine.The monitor of virtual machine Xen of open source code just is based on this technology.This technology has been avoided the problem that traditional being absorbed in-analog form efficient is lower, but need have been made amendment to the client operating system source code through the perception of VM to VMM.
The IO device virtualization:
In virtual environment, the problem that I/O faces is: the peripheral hardware resource is limited in the reality, and in order to satisfy the demand of a plurality of Client OSs to peripheral access, VMM must pass through the multiplexing limited peripheral hardware resource of the virtualized mode of I/O.In the modern computer environment, the I/O device category is abundant, and the I/O equipment with different DLLs that different vendor produces makes the virtualized difficulty of I/O quite big.
According to the features and applications of different I/O equipment, the I/O equipment in the virtualization system can fall into 5 types:
● specialized equipment (Dedicated Devices)
● divisible equipment (Partitioned Devices)
● shared device (Shared Devices)
● pseudo-off-line equipment (Spooled Devices)
● the non-existent equipment of physics
The virtual important component part of I/O as system virtualizationization, equally can according to the realization technology be divided into that software I/O is fully virtualized, the I/O class is virtual and the auxiliary I/O of hardware fully virtualized.
In the virtualized environments such as virtual fully at software, type virtual and hardware is auxiliary fully virtualized; The virtual of I/O equipment all is not quite similar, and its core is deposit position and the VMM of the primary driving of I/O equipment (Legacy Driver) processing mode to I/O equipment.
During software is virtual fully, the necessary analog machine hardware of VMM, it is (for example, the network driver) of simulating in the minimum rank of session.Although simulation is very clean in this is abstract, it also is simultaneously a most inefficient, the most complicated.In class was virtual, Guest OS and VMM can coact, and made simulation more efficient.Along with the appearance of the auxiliary fully virtualized technology of hardware, new processor makes Guest OS efficient more to the switching of VMM through adding virtualized high level instructions.In addition, hardware is also updated along with the virtualized development of I/O.
1) I/O is fully virtualized
Under the complete virtualized environment of software, generally use the I/O simulated mode to come virtual I/O equipment.In I/O was virtual, device model was used to accomplish the simulation of target device, the processing of device request and response.The software interface of device model simulated target equipment makes the driving of the enough target devices of Guest OS ability drive virtual unit, and then pellucidly virtual unit is operated.
For simulated target device software interface, device model also need be realized the function of target device simultaneously.These functions are based on the software realization.So there is not direct related and corresponding relation in the target device that device model is simulated with the hardware of host, but is based upon on certain running environment.
For the virtual machine of different structure, the working method of its device model is similar: VMM catches the read-write operation in virtual unit address zone, catches the visit to equipment; When real hardware equipment need produce interruption, VMM also must be able to trigger the interruption of Guest OS.But for the virtual machine of different structure, the environment that device model moved can be different with the position.
VMware Workstation and KVM adopt the host operating system structure, and promptly their VMM runs on the legacy operating system.This type VMM can't obtain the control fully to hardware resource, therefore adopts the method for software simulation to come virtual i/o equipment usually.The I/O operation meeting of Guest OS is caught by VMM, and hands to user's attitude process of Host OS, and this process is come the behavior of analog machine through the system call to Host OS.
VMware ESX Server is to network interface card, and key equipments such as disk have adopted fully virtualized method.It adopts independent watch-dog model, and VMM directly runs on the hardware, the direct control hardware device.What Guest OS saw is one group of unified virtual i/o equipment.Each I/O operation to these virtual units all can be trapped among the VMM, by VMM I/O is instructed and resolves and be mapped to actual physical device, and directly control hardware is accomplished then.This method can obtain preferable performance, and is also transparent fully to client operating system.But because VMM needs the directly various I/O equipment of control, it is very complicated that its design can become, and can't tackle the fast updating of equipment.
The largest benefit of analogue technique is exactly need not revise customer equipment to drive, but that its shortcoming is efficient is very low, and the difficulty of simulation is big.The maximum performance cost of this method is that processor mode switches: comprise the switching of Guest OS to VMM, and simulate the switching of process from interior nucleus band VMM to user's attitude I/O.So optimizing upward, main policies reduces mode switch exactly.In order to reduce mode switch; Many I/O instructions among the VMware Workstation are directly handled in VMM; Only run into the operation (like read-write disk, network data pack receiving and transmitting) that really can change the physical equipment state, just can switch in the I/O simulation process and further handle.
2) the I/O class is virtual
In virtual fully, I/O drives the hardware driving that comprises Guest OS, and the real hardware among equipment simulating and the VMM drives.And type virtual three-decker that does not need this redundancy, I/O can realize to be higher than managing to find time of I/O instruction-level alternately.
Under the class virtualized environment; The kernel of Guest OS is modified; Primary device drives shifts out from Guest OS, has put into an equipment virtual machine (Device VM) of authorizing through VMM, and this equipment virtual machine processing is all transferred in the I/O of Guest OS request in all the other virtual machines.And inner at Guest OS, system is that each virtual i/o equipment is installed a special driving, and this drives the transmission of being responsible for the I/O request.The equipment virtual machine is resolved the I/O request that receives and is mapped to actual physical device after authorizing through VMM, gives its primary sound device driver at last and comes control hardware to accomplish.
At present, multiple main realization arranged based on the I/O of the virtual thought of class is virtual, as the I/O of Xen virtual with the Virtio framework.
3) the auxiliary I/O of hardware is fully virtualized
Hardware is auxiliary, and I/O is fully virtualized is widely used among the Xen.Xen is after having introduced the hardware virtualization technology, and its VMM can support the Guest OS of unmodified kernel.Although contain the driver of I/O equipment among the Guest OS of unmodified kernel, similar virtual the same, Xen does not allow Guest OS directly to visit actual physical hardware equally, but through the independent virtual machine dereference of authorizing of VMM.Adopting device model in the franchise virtual machine of Xen is that the Guest OS of unmodified kernel provides the abstract of various I/O equipment.Guest OS in the virtual machine accomplishes the visit of virtual i/o equipment by means of these virtual units, the mutual then device drives completion through franchise virtual machine of these virtual i/o device accesses and actual hardware.
The virtual framework of Virtio:
One group of general-purpose simulation equipment among the virtualized VMM of Virtio type of being abstract.Virtio is by Rusty Russell exploitation, and his purpose at that time is to support the virtualization solution lguest of oneself.Virtio allows VMM to derive one group of general analog machine, and lets them become available through a general API (API).
Had after type virtual VMM, Guest OS can realize one group of general interface, after one group of rear end driver, adopts specific equipment simulating.The rear end driver needs not be general, because they only realize the behavior that front end is required.
Virtio API relies on one and simply cushions abstract order and the data that Guest OS needs that encapsulate.Except front-end driven program (in Guest OS, realizing) and rear end driver (in VMM, realizing), Virtio has also defined two layers and has supported the communication of Guest OS to VMM.Top (being called Virtio) be the virtual queue interface, it appends to the rear end driver conceptive with the front-end driven program.Driver can use 0 or a plurality of formation, and concrete quantity depends on demand.For example, the Virtio network driver uses two virtual queues (is used for receiving, and another is used for sending), and Virtio piece driver only uses a virtual queue.In fact virtual queue is implemented as the linking point of crossing over Guest OS and VMM.But this can realize that prerequisite is that Guest OS and VMM realize it in an identical manner through any-mode.
(3) summary of the invention
1 purpose:
The invention provides a kind of implementation method of using the virtual machine network performance optimization of many queue technologies, it mainly adopts the Intel Virtualization Technology of the virtual formation of many formations, for the upper strata client operating system provides network service better more rapidly.
2 technical schemes:
2.1 method general introduction
Patent of the present invention is based on the virtual machine module kernel based virtual machine of kvm-linux operating system; Be characterized in making that through a plurality of virtual queues are changed in the single virtual queue communication between virtual machine and the host network data transmission between virtual machine and the host speeds up.
2.2 overall design
Many formations thinking is with transmission/reception formation parallelization, makes a plurality of formations can send/receive data simultaneously.Single formation is extended to many formations need be to qemu (relating to one of virtualized module of IO of kvm), vhost (relating to one of virtualized module of IO of kvm), and the virtio-net of client computer makes amendment.Because after initialization is accomplished, communicating by letter between vhost and the client computer is to carry out through the kvm kernel module, do not need qemu to carry out any operation, so the modification of qemu is only limited to vhost and virtio-net initialization, offers user's use interface.Vhost is carried out many formations etendue critical be the design data structure, relevant modification has comprised vhost system call, kernel thread establishment etc.Client computer virtio-net exists with the form that network interface card drives; Realize many formations, will realize that exactly many formations network interface card drives, the key of its modification also is in the design of data structure; And the formation choice mechanism, relevant modification comprises the detection of virtio-net equipment, formation selection etc.More than the modification of three parts do not isolate, must make as a whole consideration: qemu and need transmit whether launch information such as many formations, number of queues to vhost, virtio-net, vhost, virtio-net also need go to check these information; Formation among the vhost and the formation among the virtio-net also are one to one.
In sum, a kind of implementation method of using the virtual machine network performance optimization of many queue technologies of the present invention, these method concrete steps are following:
Step 1: the netinit to QEMU is partly made amendment, and through revising QEMU, increases the support for many formations.The modification of this part relates generally to following two parts:
1.QEMU TAP (in the qemu module relevant part) initialization with data transmission
Revise data structure and make it possible to support many formations, revised the filec descriptor of corresponding formation, increased the item that is used to indicate the descriptor number etc.
Be directed to the change of data structure, revise corresponding function to adapt to new data structure.
Increase and detect the function of whether launching many formations, when not launching many formations, support single formation transmission originally.
(2.Vhost_net in the vhost module relevant part) initialization with network
Revise data structure to support many formations.
Call the ioctl system call of the many formations of amended vhost, be used for creating and controlling queue.
Step 2:, be in order to support QEMU to the use of many formations and the support of many formations of vhost-net network interface card, mainly comprise that being directed to a formation uses a thread to carry out the modification of data transmission and the modification of system call to the modification of the many formations of vhost.
1. the modification of thread is used in relevant formation
Under non-many formations situation, the corresponding kernel thread of a vhost_dev (virtual unit in the virtual machine network), but to realize many formations now, just need a plurality of kernel threads.Therefore need the corresponding thread of a formation.
We mainly design again and revise following data structure: the modification of vhost_dev is primarily aimed at after kernel thread and formation are shone upon, and has some no longer items of needs; Need specified queue number among the vhost_virtqueue, and for the work queue at its place of each queue identity; Because the modification of preceding two data structures, relative data structure vhost_poll and vhost_net also will be directed to the change of front and make some modifications.
The function relevant with these data structures also need be adjusted according to the modification of data structure accordingly.
2. the modification of system call
Concrete modification comprises: the work of original initialize queue among the open is moved on among the ioctl, because need QEMU to transmit number of queues to kernel state, the open system call can't be satisfied time requirement, and ioctl can; Release is discharged function to be sent/receives formation by original release and be extended to and discharge specify number; Increase the work of initialize queue among the ioctl, the operation of cleaning formation is made amendment, will clear up single formation and become a plurality of formations of cleaning.
Step 3: the modification to vhost-net (in the vhost module part) relevant with network many formations network interface card makes Microsoft Loopback Adapter support the transmission of many formations.Concrete modification comprises: increase the support of feature; In the virtio_net_config data structure, increase the item that is used to explain many number of queues; With the reception virtual queue rvq in the virtnet_info data structure with send virtual queue svq separately as new data structure.
(3) advantage and effect
Through the technology based on many formations of virtio, the present invention has promoted the network transmission performance of virtual machine for the network of kvm virtual machine provides reliable optimization method.Through test, this optimization method can keep under the constant situation of average cpu busy percentage, and the average transmission handling capacity that promotes big packet is more than 20%.The present invention innovates on the existing mature technology basis, implements to be not difficult, and has good use and development prospect.
(4) description of drawings
Fig. 1 does not use the KVM network function and the module frame figure of many formations
Fig. 2 uses the KVM network function and the module frame figure of many formations
Fig. 3 FB(flow block) of the present invention
Symbol description is following among the figure:
Letter character content among Fig. 1,2 is the representative function title all.
(5) embodiment
See Fig. 1, Fig. 2, Fig. 3, a kind of implementation method of using the virtual machine network performance optimization of many queue technologies of the present invention, these method concrete steps are following:
Step 1: the netinit to QEMU is partly made amendment, and through revising QEMU, increases the support for many formations.The modification of this part relates generally to following two parts:
1, QEMU TAP (among the qemu relevant part) initialization with data transmission
Revise data structure and make it possible to support many formations, revised the filec descriptor of corresponding formation, increased the item that is used to indicate the descriptor number etc.
Be directed to the change of data structure, revise corresponding function to adapt to new data structure.
Increase and detect the function of whether launching many formations, when not launching many formations, support single formation transmission originally.
2, Vhost_net (in the vhost module relevant part) initialization with network
Revise data structure to support many formations.
Call the ioctl system call of the many formations of amended vhost, be used for creating and controlling queue.
Step 2:, be in order to support QEMU to the use of many formations and the support of many formations of vhost-net network interface card, mainly comprise that being directed to a formation uses a thread to carry out the modification of data transmission and the modification of system call to the modification of the many formations of vhost.
1, the modification of thread is used in relevant formation
Under non-many formations situation, the corresponding kernel thread of a vhost_dev (virtual unit in the virtual machine network), but to realize many formations now, just need a plurality of kernel threads.Therefore need the corresponding thread of a formation.
We mainly design again and revise following data structure: the modification of vhost_dev is primarily aimed at after kernel thread and formation are shone upon, and has some no longer items of needs; Need specified queue number among the vhost_virtqueue, and for the work queue at its place of each queue identity; Because the modification of preceding two data structures, relative data structure vhost_poll and vhost_net also will be directed to the change of front and make some modifications.
The function relevant with these data structures also need be adjusted according to the modification of data structure accordingly.
2, the modification of system call
Concrete modification comprises: the work of original initialize queue among the open is moved on among the ioctl, because need QEMU to transmit number of queues to kernel state, the open system call can't be satisfied time requirement, and ioctl can; Release is discharged function to be sent/receives formation by original release and be extended to and discharge specify number; Increase the work of initialize queue among the ioctl, the operation of cleaning formation is made amendment, will clear up single formation and become a plurality of formations of cleaning.
Step 3: the modification to vhost-net (in the vhost module part) relevant with network many formations network interface card makes Microsoft Loopback Adapter support the transmission of many formations.Concrete modification comprises: increase the support of feature; In the virtio_net_config data structure, increase the item that is used to explain many number of queues; With the reception virtual queue rvq in the virtnet_info data structure with send virtual queue svq separately as new data structure.
Claims (1)
1. implementation method of using the virtual machine network performance optimization of many queue technologies, it is characterized in that: these method concrete steps are following:
Step 1: the netinit to QEMU is partly made amendment, and through revising QEMU, increases the support for many formations; The modification of this part relates to following two parts:
1.qemu part---the QEMU TAP initialization relevant in the module with data transmission
Revise data structure and make it possible to support many formations, revised the filec descriptor of corresponding formation, increased the item that is used to indicate the descriptor number;
Be directed to the change of data structure, revise corresponding function to adapt to new data structure;
Increase and detect the function of whether launching many formations, when not launching many formations, support single formation transmission originally;
2.vhost part---the Vhost_net initialization relevant in the module with network
Revise data structure to support many formations;
Call the ioctl system call of the many formations of amended vhost, be used for creating and controlling queue;
Step 2:, be in order to support QEMU to the use of many formations and the support of many formations of vhost-net network interface card, comprise that being directed to a formation uses a thread to carry out the modification of data transmission and the modification of system call to the modification of the many formations of vhost;
1. the modification of thread is used in relevant formation
Under non-many formations situation, the corresponding kernel thread of the virtual unit vhost_dev in virtual machine network, but to realize many formations now, just need a plurality of kernel threads, therefore need the corresponding thread of a formation;
Following data structure is designed again and revises: the modification of vhost_dev is directed to after kernel thread and formation shine upon, and has some no longer items of needs; Need specified queue number among the vhost_virtqueue, and for the work queue at its place of each queue identity; Because the modification of preceding two data structures, relative data structure vhost_poll and vhost_net also will be directed to the change of front and make some modifications; The function relevant with these data structures also need be adjusted according to the modification of data structure accordingly;
2. the modification of system call
Concrete modification comprises: the work of original initialize queue among the open is moved on among the ioctl, because need QEMU to transmit number of queues to kernel state, the open system call can't be satisfied time requirement, and ioctl can; Release is discharged function to be sent/receives formation by original release and be extended to and discharge specify number; Increase the work of initialize queue among the ioctl, the operation of cleaning formation is made amendment, will clear up single formation and become a plurality of formations of cleaning;
Step 3: to part relevant with network in the vhost module---the modification of many formations of vhost-net network interface card makes Microsoft Loopback Adapter support the transmission of many formations; Concrete modification comprises: increase the support of feature; In the virtio_net_config data structure, increase the item that is used to explain many number of queues; With the reception virtual queue rvq in the virtnet_info data structure with send virtual queue svq separately as new data structure.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210002171.9A CN102591715B (en) | 2012-01-05 | 2012-01-05 | Implementing method for optimizing network performance of virtual machine by using multiqueue technology |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210002171.9A CN102591715B (en) | 2012-01-05 | 2012-01-05 | Implementing method for optimizing network performance of virtual machine by using multiqueue technology |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102591715A true CN102591715A (en) | 2012-07-18 |
| CN102591715B CN102591715B (en) | 2014-02-12 |
Family
ID=46480422
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210002171.9A Expired - Fee Related CN102591715B (en) | 2012-01-05 | 2012-01-05 | Implementing method for optimizing network performance of virtual machine by using multiqueue technology |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102591715B (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103236988A (en) * | 2013-05-15 | 2013-08-07 | 华为技术有限公司 | Method and device for allocating single root-input/output virtualization (SR-IOV) network card buffering resource |
| CN103353848A (en) * | 2013-06-28 | 2013-10-16 | 华为技术有限公司 | Multiple-queue or device selection method and computing node |
| WO2014023110A1 (en) * | 2012-08-07 | 2014-02-13 | 华为技术有限公司 | Network interface adapter registration method, driving device and server |
| CN104199718A (en) * | 2014-08-22 | 2014-12-10 | 上海交通大学 | Dispatching method of virtual processor based on NUMA high-performance network cache resource affinity |
| CN104618158A (en) * | 2015-01-28 | 2015-05-13 | 上海交通大学 | Embedded network virtualization environment VirtIO (virtual input and output) network virtualization working method |
| CN104750543A (en) * | 2013-12-26 | 2015-07-01 | 杭州华为数字技术有限公司 | Thread creation method, business request processing method and associated device |
| CN105094973A (en) * | 2015-07-31 | 2015-11-25 | 上海卫星工程研究所 | ADA program partition running supporting system based on sensitive instruction substitution |
| CN106250166A (en) * | 2015-05-21 | 2016-12-21 | 阿里巴巴集团控股有限公司 | A kind of half virtualization network interface card kernel accelerating module upgrade method and device |
| US9800523B2 (en) | 2014-08-22 | 2017-10-24 | Shanghai Jiao Tong University | Scheduling method for virtual processors based on the affinity of NUMA high-performance network buffer resources |
| CN107818056A (en) * | 2016-09-14 | 2018-03-20 | 杭州华为数字技术有限公司 | A kind of queue management method and device |
| CN108491278A (en) * | 2018-03-13 | 2018-09-04 | 网宿科技股份有限公司 | A kind of method and the network equipment of processing business data |
| CN108847972A (en) * | 2018-06-08 | 2018-11-20 | 山东超越数控电子股份有限公司 | The method of network data transmission rate is improved under a kind of virtual environment |
| CN109240802A (en) * | 2018-09-21 | 2019-01-18 | 北京百度网讯科技有限公司 | request processing method and device |
| CN111352697A (en) * | 2015-04-27 | 2020-06-30 | 阿尔特拉公司 | Flexible physical function and virtual function mapping |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2007090349A1 (en) * | 2006-02-08 | 2007-08-16 | Huawei Technologies Co., Ltd. | A method for multi-queue packet data transmission and a system thereof |
| CN101557420A (en) * | 2009-03-31 | 2009-10-14 | 北京航空航天大学 | Realization method of high-efficiency network communication of a virtual machine monitor |
| CN102073760A (en) * | 2010-12-31 | 2011-05-25 | 昆明理工大学 | Blast furnace-rotary furnace section scheduling system modeling method based on queuing theory and system simulation |
-
2012
- 2012-01-05 CN CN201210002171.9A patent/CN102591715B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2007090349A1 (en) * | 2006-02-08 | 2007-08-16 | Huawei Technologies Co., Ltd. | A method for multi-queue packet data transmission and a system thereof |
| CN101557420A (en) * | 2009-03-31 | 2009-10-14 | 北京航空航天大学 | Realization method of high-efficiency network communication of a virtual machine monitor |
| CN102073760A (en) * | 2010-12-31 | 2011-05-25 | 昆明理工大学 | Blast furnace-rotary furnace section scheduling system modeling method based on queuing theory and system simulation |
Cited By (27)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2014023110A1 (en) * | 2012-08-07 | 2014-02-13 | 华为技术有限公司 | Network interface adapter registration method, driving device and server |
| CN103236988B (en) * | 2013-05-15 | 2016-02-03 | 华为技术有限公司 | The distribution method of SR-IOV network interface card buffer resource and device |
| CN103236988A (en) * | 2013-05-15 | 2013-08-07 | 华为技术有限公司 | Method and device for allocating single root-input/output virtualization (SR-IOV) network card buffering resource |
| CN103353848A (en) * | 2013-06-28 | 2013-10-16 | 华为技术有限公司 | Multiple-queue or device selection method and computing node |
| CN103353848B (en) * | 2013-06-28 | 2017-04-26 | 华为技术有限公司 | Multiple-queue or device selection method and computing node |
| CN104750543B (en) * | 2013-12-26 | 2018-06-15 | 杭州华为数字技术有限公司 | Thread creation method, service request processing method and relevant device |
| CN104750543A (en) * | 2013-12-26 | 2015-07-01 | 杭州华为数字技术有限公司 | Thread creation method, business request processing method and associated device |
| CN104199718B (en) * | 2014-08-22 | 2017-08-11 | 上海交通大学 | A kind of dispatching method of the virtual processor based on NUMA high performance network cache resources affinity |
| US9800523B2 (en) | 2014-08-22 | 2017-10-24 | Shanghai Jiao Tong University | Scheduling method for virtual processors based on the affinity of NUMA high-performance network buffer resources |
| CN104199718A (en) * | 2014-08-22 | 2014-12-10 | 上海交通大学 | Dispatching method of virtual processor based on NUMA high-performance network cache resource affinity |
| CN104618158A (en) * | 2015-01-28 | 2015-05-13 | 上海交通大学 | Embedded network virtualization environment VirtIO (virtual input and output) network virtualization working method |
| CN104618158B (en) * | 2015-01-28 | 2018-05-04 | 上海交通大学 | VirtIO network virtualizations method of work in virtual embedded network environment |
| CN111352697B (en) * | 2015-04-27 | 2023-08-29 | 阿尔特拉公司 | Flexible physical function and virtual function mapping |
| US11675613B2 (en) | 2015-04-27 | 2023-06-13 | Altera Corporation | Flexible physical function and virtual function mapping |
| CN111352697A (en) * | 2015-04-27 | 2020-06-30 | 阿尔特拉公司 | Flexible physical function and virtual function mapping |
| CN106250166A (en) * | 2015-05-21 | 2016-12-21 | 阿里巴巴集团控股有限公司 | A kind of half virtualization network interface card kernel accelerating module upgrade method and device |
| US10521221B2 (en) | 2015-05-21 | 2019-12-31 | Alibaba Group Holding Limited | Method and apparatus for upgrading vhost-net module |
| US11055086B2 (en) | 2015-05-21 | 2021-07-06 | Alibaba Group Holding Limited | Method and apparatus for upgrading vhost-net module |
| CN106250166B (en) * | 2015-05-21 | 2019-09-20 | 阿里巴巴集团控股有限公司 | A kind of half virtualizes network interface card kernel accelerating module upgrade method and device |
| CN105094973A (en) * | 2015-07-31 | 2015-11-25 | 上海卫星工程研究所 | ADA program partition running supporting system based on sensitive instruction substitution |
| CN107818056A (en) * | 2016-09-14 | 2018-03-20 | 杭州华为数字技术有限公司 | A kind of queue management method and device |
| WO2019174074A1 (en) * | 2018-03-13 | 2019-09-19 | 网宿科技股份有限公司 | Method for processing service data, and network device |
| CN108491278B (en) * | 2018-03-13 | 2020-09-18 | 网宿科技股份有限公司 | Method and network device for processing service data |
| CN108491278A (en) * | 2018-03-13 | 2018-09-04 | 网宿科技股份有限公司 | A kind of method and the network equipment of processing business data |
| CN108847972A (en) * | 2018-06-08 | 2018-11-20 | 山东超越数控电子股份有限公司 | The method of network data transmission rate is improved under a kind of virtual environment |
| CN109240802A (en) * | 2018-09-21 | 2019-01-18 | 北京百度网讯科技有限公司 | request processing method and device |
| US11210127B2 (en) | 2018-09-21 | 2021-12-28 | Beijing Baidu Netcom Science And Technology Co., Ltd. | Method and apparatus for processing request |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102591715B (en) | 2014-02-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102591715B (en) | Implementing method for optimizing network performance of virtual machine by using multiqueue technology | |
| Ammons et al. | Libra: a library operating system for a jvm in a virtualized execution environment | |
| US11573714B2 (en) | Compressibility instrumented dynamic volume provisioning | |
| US8832688B2 (en) | Kernel bus system with a hyberbus and method therefor | |
| Hirofuchi et al. | Enabling instantaneous relocation of virtual machines with a lightweight vmm extension | |
| CN101271401B (en) | Server cluster unit system with single system image | |
| US20180157519A1 (en) | Consolidation of idle virtual machines | |
| Nussbaum et al. | Linux-based virtualization for HPC clusters | |
| KR101640769B1 (en) | Virtual system and instruction executing method thereof | |
| US20170102963A1 (en) | Method for Processing Interrupt by Virtualization Platform, and Related Device | |
| Xue et al. | {gScale}: Scaling up {GPU} Virtualization with Dynamic Sharing of Graphics Memory Space | |
| CN101620547A (en) | Virtual physical interrupt processing method of X86 computer | |
| Song et al. | Characterizing the performance and scalability of many-core applications on virtualized platforms | |
| WO2016101282A1 (en) | Method, device and system for processing i/o task | |
| Camargos et al. | Virtualization of Linux servers | |
| Liu et al. | Understanding the virtualization" Tax" of scale-out pass-through GPUs in GaaS clouds: An empirical study | |
| Masood et al. | Virtualization tools and techniques: Survey | |
| Bose et al. | Benchmarking database performance in a virtual environment | |
| Akshintala et al. | Trillium: The code is the IR | |
| CN103631648A (en) | Task processing method and system | |
| WO2021211241A1 (en) | Hypervisor hot restart | |
| Meier et al. | IBM systems virtualization: Servers, storage, and software | |
| De Gelas et al. | Hardware virtualization: the nuts and bolts | |
| Nguyen | Fast delivery of virtual machines and containers: understanding and optimizing the boot operation | |
| Carbone et al. | Windows Server 2008 Hyper-V Resource Kit |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20210218 Address after: 100089 West N3, Jinyu zhizaochang, Jiancai East Road, Haidian District, Beijing Patentee after: Zhongke Yucheng (Beijing) Technology Service Co.,Ltd. Address before: 100191 No. 37, Haidian District, Beijing, Xueyuan Road Patentee before: BEIHANG University |
|
| TR01 | Transfer of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20140212 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |