Node and network acceleration method

The method applies vDPA technology to Kubernetes-based Containerized VMs by defining a vDPA network interface, enhancing network performance and portability, enabling efficient network acceleration and VM migration.

KR103046719B1Active Publication Date: 2026-11-10SK TELECOM CO LTD
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
KR1020240070092
Authority / Receiving Office
KR · KR
Patent Type
Patents
Current Assignee / Owner
Filing Date
2024-05-29
Publication Date
2026-11-10
Estimated Expiration
2044-05-29

AI Technical Summary

Technical Problem

Existing vDPA technology is designed for VMs running directly on Bare Metal and does not support application to containerized VMs in Kubernetes environments, limiting network performance improvement in such setups.

Method used

A method is proposed to apply vDPA technology to Kubernetes-based Containerized VMs by defining a virtual machine based on a vDPA network, using a vDPA-CNI to create a vDPA interface directly connecting to a vDPA hardware device, involving a compute container and a vdpa-binding-sidecar container to provide vDPA interface information, and generating a VM definition document with injected vDPA interface details.

Benefits of technology

Enables network acceleration for containerized VMs, ensuring high-speed network performance and portability, allowing non-disruptive VM migration across environments with SR-IOV-level transmission speeds.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 112024058321267-PAT00006_ABST
    Figure 112024058321267-PAT00006_ABST
Patent Text Reader

Abstract

The present invention relates to a method for applying vDPA (virtio Data Path Acceleration) technology to Kubernetes-based Containerized VMs (Virtual Machines).
Need to check novelty before this filing date? Find Prior Art

Description

Technology Field

[0001] The present invention relates to a method for applying vDPA (virtio Data Path Acceleration) technology to Kubernetes-based Containerized VMs (Virtual Machines). Background Technology

[0002] vDPA (virtio Data Path Acceleration) is a technology designed to improve the network performance of VMs (Virtual Machines) and can maintain the portability of VMs while guaranteeing SR (Single Root)-IOV (I / O Virtualization) level transmission speeds.

[0003] vDPA technology was designed and developed with the case of running a Virtual Machine (VM) directly on a Bare Metal (BM) in mind.

[0004] Therefore, due to this limited design, while application cases for VMs or containers exist for vDPA technology, application cases for containerized VMs cannot be found.

[0005] In this regard, VMs provisioned in Kubernetes take the form of Containerized VMs in which VMs are deployed inside a POD.

[0006] Therefore, in order to expect improved network performance in a Kubernetes environment, specific measures need to be established to apply vDPA technology to Containerized VMs.

[0007] Accordingly, the present invention proposes a new method that enables network acceleration for containerized VMs, just as network acceleration is performed for general VMs. Meanwhile, the technology forming the background of the present invention is disclosed in 'Korean Published Patent Application No. 10-2019-0107978 (September 23, 2019)'. The problem to be solved

[0008] The present invention was created in consideration of the aforementioned circumstances, and the objective of the present invention is to propose a specific method for applying vDPA (virtio Data Path Acceleration) technology to Kubernetes-based Containerized VMs (Virtual Machines). means of solving the problem

[0009] A node according to an embodiment of the present invention for achieving the above objective is characterized by comprising: a memory containing instructions; and a processor that defines a virtual machine (VM) based on a vDPA (virtio Data Path Acceleration) network by executing the instructions, and calls a vDPA-CNI (Container Network Interface) when a POD in which the VM is deployed is created to generate a vDPA interface for directly connecting the vDPA network to a vDPA hardware device.

[0010] Specifically, the POD may include a compute container, which is a base container for running the VM, and a vdpa-binding-sidecar container, which is an auxiliary container that provides information about the vDPA interface.

[0011] Specifically, the processor generates Domain.XML, which is a VM definition document for running the VM, and the vdpa-binding-sidecar container can inject information about the vDPA interface into the Domain.XML.

[0012] Specifically, the vDPA-CNI includes a network-device-info-map, which is information about the vDPA HW device, as an annotation within the POD, and the vdpa-binding-sidecar container can reference the network-device-info-map from a volume that is created separately from the annotation within the POD and mounted to the container.

[0013] Specifically, the processor can drive the compute container based on Domain. XML containing information about the vDPA interface, so that the VM is created and executed by the compute container.

[0014] A network acceleration method performed at a node according to an embodiment of the present invention for achieving the above objective is characterized by comprising: a virtualization definition step of defining a virtual machine (VM) based on a vDPA (virtio Data Path Acceleration) network; and an interface creation step of creating a vDPA interface for directly connecting the vDPA network to a vDPA HW device by calling a vDPA-CNI (Container Network Interface) when a POD in which the VM is deployed is created.

[0015] Specifically, the POD may include a compute container, which is a base container for running the VM, and a vdpa-binding-sidecar container, which is an auxiliary container that provides information about the vDPA interface.

[0016] Specifically, the above method further includes a virtualization document definition step of generating Domain.XML, which is a VM definition document for running the VM, and the vdpa-binding-sidecar container can inject information about the vDPA interface into the Domain.XML.

[0017] Specifically, the vDPA-CNI includes a network-device-info-map, which is information about the vDPA HW device, as an annotation within the POD, and the vdpa-binding-sidecar container can reference the network-device-info-map from a volume that is created separately from the annotation within the POD and mounted to the container.

[0018] Specifically, the above method can drive the compute container based on a Domain.XML containing information about the vDPA interface, so that the VM is created and executed by the compute container. Effects of the invention

[0019] Accordingly, according to the node and network acceleration method of the present invention, a virtual machine (VM) based on a vDPA network is defined, and when a POD in which the VM is deployed is created, a vDPA-CNI (Container Network Interface) is called to create a vDPA interface for directly connecting the vDPA network to a vDPA hardware device, thereby enabling network acceleration for containerized VMs just like for general VMs. Brief explanation of the drawing

[0020] Figures 1 and 2 are exemplary diagrams illustrating conventional vDPA (virtio Data Path Acceleration) technology. FIG. 3 is an exemplary diagram illustrating a Kubernetes environment according to an embodiment of the present invention. FIGS. 4 and FIGS. 5 are a configuration diagram and an example diagram for explaining a node according to an embodiment of the present invention. FIG. 6 is a flowchart illustrating a network acceleration method according to an embodiment of the present invention. Specific details for implementing the invention

[0021] Hereinafter, preferred embodiments of the present invention will be described with reference to the attached drawings.

[0022] In one embodiment of the present invention, vDPA (virtio Data Path Acceleration) technology is used.

[0023] Generally, PCI-PT (Pass Through) or SR (Single Root)-IOV (I / O Virtualization), which directly connects the NIC (Network Interface Controller) to the Guest VM, are mainly used to improve the network performance of a VM (Virtual Machine).

[0024] In this case, similar to the virtio (Virtual I / O) architecture, the communication channel with the NIC can be managed by separating it into a data plane and a control plane.

[0025] Here, the control plane is used for configuration changes and function negotiations between the NIC and the VM to establish and terminate the data plane, and the data plane is used to transmit actual data (packets) between the NIC and the VM.

[0026] However, in order to connect the NIC directly to the VM, the NIC needs to support a virtio ring layout.

[0027] In this regard, vDPA technology can maintain the portability of VMs by ensuring SR-IOV-level transmission speeds without exposing vendor-specific interfaces to VMs.

[0028] vDPA can essentially be understood as an approach that standardizes the NIC SR-IOV data plane using a virtio layout and deploys a single standard virtio driver on VMs isolated from vendor implementations, while adding a common control plane and SW infrastructure to support it.

[0029] In vDPA technology, the data plane uses a virtio ring layout, similar to virtio full hardware offloading, so that the NIC and VM are directly connected.

[0030] However, unlike virtio full HW offloading, vDPA technology adds a general-purpose vDPA driver (VDPA parent driver) to the kernel (Host kernel), as shown in Fig. 1, to automatically translate the driver and control plane of the vendor NIC (vDPA HW device) into the virtio control plane.

[0031] This allows NIC vendors to minimize the dependency between their own drivers (lightweight vDPA add-ons) and the VM's virtio control plane, thereby enabling VMs to perform network acceleration more universally regardless of the NIC vendor.

[0032] As such, vDPA enables NIC vendors to achieve line-rate performance in the data plane while maintaining VM portability with much less effort by implementing virtio's data plane inside the NIC much more flexibly than Virtio Full HW offloading and adding a translation layer between the vendor NIC driver and virtio's control plane.

[0033] Meanwhile, conventional vDPA technology was designed and developed with consideration for cases where a VM (Virtual Machine) is placed directly on a BM (Bare-Metal).

[0034] To apply vDPA technology to a VM, for example, as shown in Fig. 2, the vhost interface within the VM is linked with the vhost-vdpa interface, and through this, the VM can be directly connected to the VFs (Virtual Functions) within the NIC.

[0035] The vDPA Framework abstracts common properties of vDPA hardware and supports the use of existing vhost or virtio (Virtual I / O) subsystems, thereby hiding the complexity of hardware vDPA implementations and providing a secure and integrated interface that can be used by both Kernelspace and userspace subsystems.

[0036] From a hardware perspective, vDPA can be implemented by different types of H / W interfaces, and if implemented via a PCI interface, it can be a Physical Function (PF), a VF, or a Scalable Function (SF), which is another slicing technology of a PF.

[0037] Of course, vDPA can be implemented through a non-PCI interface or through software emulation.

[0038] Meanwhile, as mentioned earlier, this vDPA technology is designed to consider only cases where VMs are placed directly on the BM, so while there are application cases for VMs or containers, there are no application cases for containerized VMs.

[0039] In this regard, VMs provisioned in Kubernetes take the form of Containerized VMs in which VMs are deployed inside a POD.

[0040] Therefore, in order to expect improved network performance in such a Kubernetes environment, specific measures need to be established to apply vDPA technology to Containerized VMs.

[0041] Accordingly, in one embodiment of the present invention, a specific method for applying vDPA technology to a Containerized VM is proposed.

[0042] FIG. 3 illustrates an exemplary Kubernetes environment according to an embodiment of the present invention.

[0043] As illustrated in FIG. 3, a Kubernetes environment according to one embodiment of the present invention includes a node (100) and is divided into a userspace, which is a memory space used for a program to operate, and a kernelspace, which is a memory space required to run an operating system.

[0044] VMs are deployed in userspace, taking the form of containerized VMs where VMs are deployed inside a POD.

[0045] Such containerized VMs can be created using a combination of Kubernetes, which manages the operation of PODs in kernelspace, and KubeVirt, which supports the virtualization of VMs within PODs.

[0046] In particular, in one embodiment of the present invention, network acceleration can be supported for a Containerized VM in userspace, which is made possible through vDPA-CNI (Container Network Interface) that supports a direct connection between the VM and the NIC.

[0047] In summary, in a Kubernetes environment according to one embodiment of the present invention, network acceleration can be supported by applying vDPA technology to a Containerized VM through the aforementioned configuration. Below, the configuration of a node (100) for realizing this will be described in more detail.

[0048] FIG. 4 schematically shows the configuration of a node (100) according to one embodiment of the present invention.

[0049] As illustrated in FIG. 4, a node (100) according to one embodiment of the present invention may be configured to include a memory containing instructions and a processor that executes instructions within the memory.

[0050] In particular, in the case of a processor according to one embodiment of the present invention, it may have a functional configuration including a virtualization unit (110) and an operation unit (120) according to an implementation function according to the execution of instructions.

[0051] The virtualization unit (110) can be understood as a function of KubeVirt that supports virtualization in userspace.

[0052] In the following, the implementation functions of this virtualization unit (110) will be named by classifying them into VM Controller, VMI Controller, and virt-handler according to their detailed functions.

[0053] The operation department (120) can be understood as a Kubernetes function that manages the operation of the POD in userspace.

[0054] Above, the node (100) according to the embodiment of the present invention can support network acceleration by applying vDPA technology to a Containerized VM through the functional configuration of the processor described above. Below, each functional configuration of the processor to realize this will be explained with reference to the contents of FIG. 5.

[0055] The virtualization unit (110) is responsible for defining VMs based on the vDPA network.

[0056] More specifically, the virtualization unit (110) defines a VM that utilizes the vDPA network in response to a request from a user.

[0057] At this time, when the VM Controller receives a VM CR (Custom Resource) with a vDPA network attached from the user, it creates a VMI (Virtual Machine Instance) CR based on the entered VM CR (Create VMI CR FROM VM).

[0058] And, when a VMI CR is created, the VMI Controller generates a manifest, which is metadata for creating a POD, based on the VMI CR (Create POD manifest from VMI).

[0059] Meanwhile, the VMI CR can be referenced to generate documentation for running the VM.

[0060] In this regard, virt-handler generates Domain.XML, a VM definition document for running the VM, by referencing the VMI CR (Define & manage the VM Life-Cycle).

[0061] The operation department (120) is responsible for the function of creating PODs.

[0062] More specifically, the operations department (120) creates a POD in which a VM is placed by referencing the POD's manifest.

[0063] To this end, the operation unit (120) calls vDPA-CNI (Container Network Interface) for the POD where the VM is placed to create a vDPA interface for directly connecting the vDPA network to a vDPA HW device (e.g., NIC).

[0064] At this time, when a node available for the vDPA data plane is designated, the vDPA CNI generates network-device-info-map information, which is information about the vDPA HW device (PF, VF, or SF) to be connected to the data plane of that node, on-demand and injects it into the POD as an annotation (inject vDPA device-info into the POD as an annotation).

[0065] Meanwhile, a POD according to one embodiment of the present invention is created as a virt-launcher POD to include a compute container, which is a basic container for running a VM, and a vdpa-binding-sidecar container, which is an auxiliary container that provides information about the vDPA interface.

[0066] In this regard, the aforementioned virt-handler does not provide detailed information about the vDPA interface.

[0067] Therefore, in one embodiment of the present invention, a vdpa-binding-sidecar container that serves the role of writing the relevant information is created separately.

[0068] However, as the vdpa-binding-sidecar container operates at the same hierarchy as the compute container, it cannot directly reference network-device-info-map information written as annotations within the POD.

[0069] Accordingly, the operation unit (120) creates a volume for network-device-info-map information separately from the annotation within the POD and mounts the information so that it can be referenced in the vdpa-binding-sidecar container (load the annotation into a volume / refer to the device-info by mounting the volume).

[0070] In response to this, the vdpa-binding-sidecar container generates information about vDPA interfaces in XML format by referencing the network-device-info-map information as it is mounted as a volume, and injects the generated information about vDPA interfaces into the Domain.XML, which is the VM definition document generated by the virt-handler to run the VM.

[0071] To aid in understanding the explanation, the process of generating information about the vDPA interface in XML format with reference to the contents of Figure 5 is examined in more detail as follows.

[0072] In other words, vdpa-binding-sidecar generates XML based on the content defined in network-device-info-map. By default, it checks the type within device-info to determine whether it is vDPA, and if the result is vDPA, it creates an interface with type vDPA.

[0073] The vhost is automatically entered as the driver name in device-info, the path content within the vdpa area is copied to Source dev, and ua-${device-name} is entered as the alias name.

[0074] Finally, the PCI address to be used as the data plane is written to address, and this information is generated based on the pci-address content.

[0075] Here, the first four digits are domain values ​​in hex format, the last two digits are bus values ​​in hex format, the next two digits are slot values ​​in hex format, and the last digit is function values, which are automatically generated when using VF or SF, and are also copied and written to the function area.

[0076] Thus, when information about the vDPA interface generated in XML format by the vdpa-binding-sidecar container is injected into the Domain.XML generated by the virt-handler, the virt-handler starts the compute container based on the Domain.XML, causing the compute container to create and execute a VM within the POD.

[0077] As described above, according to the configuration of a node (100) in one embodiment of the present invention, a VM based on a vDPA network is defined, and when a POD in which the VM is deployed is created, a vDPA-CNI (Container Network Interface) is called to create a vDPA interface for directly connecting the vDPA network to a vDPA HW device, thereby enabling network acceleration for a Containerized VM just like for a general VM. Accordingly, a high-speed network stack can be used in a Containerized VM, allowing for enhanced network performance even in an environment where VMF and CMF are mixed. Furthermore, it provides an environment that guarantees VM portability, such as non-disruptive VM migration (Live Migration) from one BM to another, while simultaneously guaranteeing SR-IOV-level transmission speeds (whereas Live Migration is impossible with SR-IOV due to dependencies between the VM and NIC vendors, Live Migration is possible with vDPA).

[0078] Hereinafter, a network acceleration method according to an embodiment of the present invention will be described with reference to FIG. 6.

[0079] For convenience of explanation, the node (100) described above with reference to FIGS. 4 and FIG. 5 will be referred to as the entity performing the network acceleration method described below.

[0080] First, the node (100) defines a VM that utilizes the vDPA network in response to a request from a user.

[0081] At this time, when the VM Controller receives a VM CR (Custom Resource) with a vDPA network attached from the user, it creates a VMI (Virtual Machine Instance) CR based on the received VM CR (S110-S120).

[0082] And, when a VMI CR is created, the VMI Controller creates a manifest, which is metadata for creating a POD, based on the VMI CR (S130).

[0083] Meanwhile, the VMI CR can be referenced to generate documentation for running the VM.

[0084] In this regard, virt-handler generates Domain.XML, which is a VM definition document for running the VM, by referencing the VMI CR (S140).

[0085] Furthermore, the node (100) creates a POD in which a VM is placed by referencing the POD's manifest.

[0086] To this end, the node (100) calls vDPA-CNI (Container Network Interface) for the POD where the VM is placed to create a vDPA interface for directly connecting the vDPA network to a vDPA HW device (e.g., NIC) (S150).

[0087] At this time, when a node available for the vDPA data plane is designated, the vDPA CNI generates network-device-info-map information, which is information about the vDPA HW device (PF, VF, or SF) to be connected to the data plane of that node, on-demand and writes it in the form of annotations within the POD.

[0088] Meanwhile, a POD according to one embodiment of the present invention is created as a virt-launcher POD to include a compute container, which is a basic container for running a VM, and a vdpa-binding-sidecar container, which is an auxiliary container that provides information about the vDPA interface.

[0089] In this regard, the aforementioned virt-handler does not provide detailed information about the vDPA interface.

[0090] Therefore, in one embodiment of the present invention, a vdpa-binding-sidecar container that serves the role of writing the relevant information is created separately.

[0091] However, as the vdpa-binding-sidecar container operates at the same hierarchy as the compute container, it cannot directly reference network-device-info-map information written as annotations within the POD.

[0092] Accordingly, the node (100) creates a volume for network-device-info-map information separately from the annotation within the POD and mounts the information so that it can be referenced in the vdpa-binding-sidecar container (S160).

[0093] In this regard, the vdpa-binding-sidecar container generates information about vDPA interfaces in XML format by referencing the network-device-info-map information as it is mounted as a volume, and injects the generated information about vDPA interfaces into Domain.XML, which is a VM definition document generated by virt-handler to run the VM (S170).

[0094] To aid in understanding the explanation, let's look more specifically at the process of generating information about the vDPA interface in XML format, referring to the content of Figure 5 shown earlier.

[0095] In other words, vdpa-binding-sidecar generates XML based on the content defined in network-device-info-map. By default, it checks the type within device-info to determine whether it is vDPA, and if the result is vDPA, it creates an interface with type vDPA.

[0096] The vhost is automatically entered as the driver name in device-info, the path content within the vdpa area is copied to Source dev, and ua-${device-name} is entered as the alias name.

[0097] Finally, the PCI address to be used as the data plane is written to address, and this information is generated based on the pci-address content.

[0098] Here, the first four digits are domain values ​​in hex format, the last two digits are bus values ​​in hex format, the next two digits are slot values ​​in hex format, and the last digit is function values, which are automatically generated when using VF or SF, and are also copied and written to the function area.

[0099] Subsequently, when information about the vDPA interface generated in XML format in the vdpa-binding-sidecar container is injected into the Domain.XML generated by the virt-handler, the virt-handler starts the compute container based on the Domain.XML, causing a VM to be created and executed within the POD by the compute container (S180).

[0100] As described above, according to a network acceleration method of an embodiment of the present invention, a VM based on a vDPA network is defined, and when a POD in which the VM is deployed is created, a vDPA-CNI (Container Network Interface) is called to create a vDPA interface for directly connecting the vDPA network to a vDPA HW device, thereby enabling network acceleration for Containerized VMs just like for general VMs. Accordingly, a high-speed network stack can be used in Containerized VMs, allowing for enhanced network performance even in environments where VMF and CMF are mixed. Furthermore, it provides an environment that guarantees VM portability, such as non-disruptive VM migration (Live Migration) from one BM to another, while simultaneously guaranteeing SR-IOV-level transmission speeds (whereas Live Migration is impossible with SR-IOV due to dependencies between VM and NIC vendors, Live Migration is possible with vDPA).

[0101] Meanwhile, a network acceleration method according to one embodiment of the present invention may be implemented in the form of program instructions that can be executed through various computer means and recorded on a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc., either individually or in combination. The program instructions recorded on the medium may be those specifically designed and configured for the present invention, or those known and available to those skilled in the art of computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes; optical recording media such as CD-ROMs and DVDs; magneto-optical media such as floptical disks; and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, and flash memory. Examples of program instructions include machine code, such as that generated by a compiler, as well as high-level language code that can be executed by a computer using an interpreter, etc. The hardware devices described above may be configured to operate as one or more software modules to perform the operation of the present invention, and vice versa.

[0102] Although the present invention has been described in detail with reference to preferred embodiments, the present invention is not limited to the above-described embodiments, and the technical concept of the present invention extends to the scope in which various modifications or alterations are possible by anyone with ordinary knowledge in the technical field to which the present invention belongs, without departing from the gist of the present invention as claimed in the following claims. Industrial applicability

[0103] According to the node and network acceleration method of the present invention, the limitations of existing technologies are overcome in that vDPA (virtio Data Path Acceleration) technology can be applied to Kubernetes-based Containerized VMs (Virtual Machines). As such, the invention is industrially applicable because it not only offers sufficient potential for the commercialization or sales of the applied device, but is also practically and clearly implementable. Explanation of the symbols

[0104] 100: Node 110: Verification Unit 120: Scheduling Unit

Claims

Claim 1 A node comprising: a memory containing instructions; and a processor that defines a virtual machine (VM) based on a vDPA (virtio Data Path Acceleration) network by executing the instructions, and, when creating a POD on which the VM is placed, calls a vDPA-CNI (Container Network Interface) for the POD on which the VM is placed to create a vDPA interface for directly connecting the vDPA network to a vDPA hardware device, wherein the POD includes a compute container, which is a basic container for running the VM, and a vdpa-binding-sidecar container, which is an auxiliary container that provides information about the vDPA interface. Claim 2 delete Claim 3 A node according to claim 1, wherein the processor generates Domain.XML, which is a VM definition document for running the VM, and the vdpa-binding-sidecar container injects information about the vDPA interface into the Domain.XML. Claim 4 A node according to claim 1, wherein the vDPA-CNI writes a network-device-info-map, which is information about the vDPA HW device, as an annotation within the POD, and the vdpa-binding-sidecar container references the network-device-info-map from a volume that is created separately from the annotation within the POD and mounted to the container. Claim 5 A node according to claim 3, wherein the processor drives the compute container based on Domain. XML containing information about the vDPA interface, thereby causing the VM to be created and executed by the compute container. Claim 6 A network acceleration method performed at a node, comprising: a virtualization definition step for defining a virtual machine (VM) based on a vDPA (virtio Data Path Acceleration) network; and an interface creation step for creating a vDPA interface for directly connecting the vDPA network to a vDPA hardware device by calling vDPA-CNI (Container Network Interface) when creating a POD on which the VM is deployed, wherein the POD includes a compute container which is a basic container for running the VM, and a vdpa-binding-sidecar container which is an auxiliary container that provides information about the vDPA interface. Claim 7 delete Claim 8 A network acceleration method according to claim 6, wherein the method further includes a virtualization document definition step of generating Domain.XML, which is a VM definition document for running the VM, and wherein the vdpa-binding-sidecar container injects information about the vDPA interface into the Domain.XML. Claim 9 A network acceleration method according to claim 6, wherein the vDPA-CNI writes a network-device-info-map, which is information about the vDPA HW device, as an annotation within the POD, and the vdpa-binding-sidecar container references the network-device-info-map from a volume that is created separately from the annotation within the POD and mounted to the container. Claim 10 In claim 8, the network acceleration method is characterized by driving the compute container based on a Domain. XML containing information about the vDPA interface, so that the VM is created and executed by the compute container. Claim 11 A computer program stored on a computer-readable recording medium in combination with hardware to execute the method of any one of claims 6, 8 through 10.

Citation Information

Patent Citations

  • Node for operating container group, and management system and method for container group

    CN115705198A

  • Configurable device interface

    US20210232528A1

  • Method and application for realizing virtual network data packet forwarding based on vDPA

    CN113472624A

  • Method and device for creating virtual machine with vDPA equipment, equipment and storage medium

    CN117742891A