Adaptive live migration of a virtual machine for a physical storage device controller
Disaggregated computing infrastructure with dynamic resource allocation addresses inefficiencies in data centers by optimizing resource utilization and reducing costs through flexible allocation and high-speed interconnects.
Patent Information
- Application Number
- DE112022007644
- Authority / Receiving Office
- DE · DE
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2022-11-08
- Publication Date
- 2025-07-31
AI Technical Summary
Existing data centers face inefficiencies in resource utilization and management due to the integration of compute, storage, and network resources in hyperconverged servers, leading to suboptimal upgrade and utilization rates, and the need for improved network speed and latency in cloud computing environments.
Implementing disaggregated computing infrastructure with high-speed interconnects to aggregate computing, storage, and network fabric resources, allowing for flexible resource allocation and management through an orchestrator server that dynamically allocates resources based on workload requirements, enhancing resource utilization and reducing operational costs.
This approach improves resource utilization and reduces operational costs by enabling independent upgrades and optimal resource allocation, while maintaining high-speed network connectivity and efficient workload execution.
Smart Images
Figure 00000000_0000_ABST
Abstract
Claims
[1] Data center, comprising: a source network device; and a destination network device, wherein the destination network device is to receive a migration message and perform a migration of a virtual machine for a first physical storage device in a source network device to a second physical storage device in the destination network device and perform a peer-to-peer copy between a first virtual Non-Volatile Memory Express (NVMe) namespace for the first physical storage device in the source network device to a second NVMe namespace for the second physical storage device in the destination network device without stopping the virtual machine. [2] The data center of claim 1, further comprising: a hypervisor to send the migration message using a remote procedure call. [3] The data center of claim 1, further comprising: a hypervisor to send the migration message using a socket communication channel. [4] The data center of claim 1, wherein the target network device comprises: a copy engine, where the copy engine has to perform the peer-to-peer copy. [5] The data center of claim 4, wherein the source network device is to process input / output (I / O) commands for the first physical storage device while the copy engine performs the peer-to-peer copy. [6] The data center of claim 4, wherein the copy engine comprises: a logical block address bitmap for a range of logical block addresses used by the virtual machine, the logical block address bitmap to record a synchronization status of the range of logical block addresses between the first physical storage device and the second physical storage device during the peer-to-peer copy. [7] The data center of claim 1, wherein the destination network device is an infrastructure processing unit and the source network device is an infrastructure processing unit. [8] A network device, the network device comprising: a network interface controller for communicating with a source network device; and a physical storage device, wherein the network device is to receive a migration message and perform a migration of a virtual machine for a first physical storage device in the source network device to the physical storage device and perform a peer-to-peer copy between a first virtual Non-Volatile Memory Express (NVMe) namespace for a first physical storage device in a source network to a second NVMe namespace for the physical storage device without stopping the virtual machine. [9] The network device of claim 8, wherein a hypervisor is to send the migration message using a remote procedure call. [10] The network device of claim 8, wherein a hypervisor is to send the migration message using a socket communication channel. [11] The network device of claim 8, further comprising: a copy engine, where the copy engine has to perform the peer-to-peer copy. [12] The network device of claim 11, wherein the source network device is to process input / output (I / O) commands for the first physical storage device while the copy engine performs the peer-to-peer copy. [13] The network device of claim 11, wherein the copy engine comprises: a logical block address bitmap for a range of logical block addresses used by the virtual machine, the logical block address bitmap to record a synchronization status of the range of logical block addresses between the first physical storage device and a second physical storage device during the peer-to-peer copy. [14] The network device of claim 8, wherein the network device is an infrastructure processing unit and the source network device is an infrastructure processing unit. [15] One or more non-transitory machine-readable media comprising a plurality of instructions stored thereon that, in response to their execution, cause a network device to: performs a migration of a virtual machine for a first physical storage device in a source network device to a second physical storage device in the network device; and performs a peer-to-peer copy between a first Non-Volatile Memory Express (NVMe) namespace in the source network device and a second NVMe namespace for the second physical storage device in the network device without stopping the virtual machine. [16] The one or more non-transitory machine-readable media of claim 15, wherein a hypervisor is to send a migration message using a remote procedure call. [17] The one or more non-transitory machine-readable media of claim 15, wherein a hypervisor is to send a migration message using a socket communication channel. [18] The one or more non-transitory machine-readable media of claim 15, wherein the source network device is to process input / output (I / O) commands for the first physical storage device while the network device is performing the peer-to-peer copy. [19] The non-transitory machine-readable medium or multiple non-transitory machine-readable media of claim 18, wherein the network device further comprises: store a logical block address bitmap for a range of logical block addresses used by the virtual machine; and record in the logical block address bitmap a synchronization status of the range of logical block addresses between the first physical storage device and the second physical storage device during the peer-to-peer copy. [20] The one or more non-transitory machine-readable media of claim 15, wherein the network device is an infrastructure processing device and the source network device is an infrastructure processing device.