Virtualization method for audio-frequency integrated network resources

Abstract

The invention discloses a virtualization method for audio frequency integrated network resources, which belongs to the technical field of audio frequency network resource management, and comprises the following steps of: constructing a distributed virtual switch layer covering a subnet, embedding a virtual switch instance into a kernel mode network stack of each physical host in a cluster, dynamically mapping the cloud hosts in the subnet to a physical host group through a consistent Hash algorithm; when a cloud host sends an audio data packet, a source physical host virtual switch instance executes the following steps: extracting a DSCP field and marking the DSCP field as EF, and through GRE-over-UDP tunnel packaging, replacing a target MAC address with a virtual MAC address of a next hop virtual switch instance; and after the virtual switch instance of the target physical host is decapsulated, delivering the audio packet to the target cloud host based on the virtual port forwarding table. According to the method, efficient virtualization transmission of audio data is realized through a distributed virtual switching layer and a consistent Hash mapping mechanism, QoS guarantee and low-delay transmission are supported, and the method is suitable for audio network resource management in a large-scale cloud environment.

Description

Technical Field

[0001] The invention belongs to the technical field of audio network resource management, and in particular relates to a virtualization method for audio integrated network resources. Background Art

[0002] With the rapid development of cloud computing and virtualization technologies, the demand for real-time transmission of audio streams (such as voice calls, video conferencing, etc.) in cloud environments is increasing.

[0003] Traditional virtualized networks typically use centralized virtual switches or virtual network devices to manage network traffic in cloud environments. This architecture has the following problems when processing audio streams: Traditional virtualized network resource allocation strategies are usually static and cannot dynamically adjust resource parameters such as the bandwidth and buffer size of virtual ports based on real-time audio stream requirements, resulting in low resource utilization or performance bottlenecks; traditional virtualized networks lack dedicated identification and optimization mechanisms for audio streams. Audio streams are mixed with other types of data streams (such as video streams and ordinary data packets) for transmission, which can easily lead to audio stream delays, jitter and packet loss problems, affecting user experience; traditional virtualized network architectures typically rely on complex physical network devices (such as physical switches, routers, etc.) and virtual network devices (such as virtual switches, virtual firewalls, etc.). This multi-layer architecture will lead to increased network latency and management complexity in large-scale cloud environments.

[0004] Therefore, it is necessary to propose a virtualization method for audio integrated network resources to solve the above problems. Summary of the Invention

[0005] In view of this, the purpose of the present invention is to provide a virtualization method for audio integrated network resources, which realizes efficient transmission of audio streams and dynamic optimization of resources through the collaborative work of audio stream identification, virtual port manager and policy execution engine.

[0006] In order to achieve the above object, the present invention provides the following technical solutions: The present invention provides a method for virtualizing audio-integrated network resources, comprising the following steps: S1: constructing a distributed virtual switching layer covering a subnet, embedding a virtual switch instance into the kernel-mode network stack of each physical host in a cluster, wherein the virtual switch instance includes an audio stream identification module, a virtual port manager, and a policy execution engine; S2: Dynamically map the cloud hosts in the subnet to the physical host group using a consistent hashing algorithm. The mapping relationship satisfies: Where, is the SHA-256 hash function, which is used to map the input to a fixed-length hash value. is the unique identifier of the virtual machine, is the unique identifier of the physical host, is a bitwise exclusive OR operation, is the preset threshold, ; S3: When the cloud host sends an audio packet, the source physical host virtual switch instance performs the following operations: extracting the DSCP field and marking it as EF, encapsulating it through the GRE-over-UDP tunnel, and replacing the destination MAC address with the virtual MAC address of the next-hop virtual switch instance; S4: After decapsulation, the virtual switch instance of the target physical host delivers the audio packet to the target cloud host based on the virtual port forwarding table.

[0007] Furthermore, the audio stream identification module adopts a dual-threshold trigger mechanism, wherein: The first threshold is: 5 consecutive packet payload sizes Bytes and Packet Interval ; The second threshold is: RTP sequence number increment Deviation rate ; Where, is the actual serial number increment, is the theoretical serial number increment, ; When the first threshold and the second threshold are met at the same time, the audio dedicated forwarding channel is activated. , it is determined to be a discontinuous audio stream or a network attack stream, and the security policy is triggered: the data packet is discarded and an alarm event is generated.

[0008] Furthermore, the audio dedicated forwarding channel execution steps include: A1: Bypass the traditional Linux TC flow table and directly call the eBPF program for forwarding; A2: Allocate an independent memory pool; A3: Enable forward error correction coding: , in =4, is the forward error correction block, To perform continuous XOR operations from the 1st to the kth data packets, For the i Raw audio data packets.

[0009] Furthermore, in step A3, the forward error correction coding adopts a non-uniform protection mechanism, including: When the silence detection flag is False, it is determined to be a key audio frame, and the key audio frame is divided into two parts according to the redundancy. To distribute; where, is the current path packet loss rate; When the silence detection flag is True, it is determined to be a non-critical audio frame, and the non-critical audio frame is divided into two parts according to the redundancy. Make allocation.

[0010] Furthermore, the optimization strategies implemented by the strategy execution engine include: real-time monitoring of the peak bandwidth of the audio stream in the cluster , when bandwidth hour, Enable zero-copy buffer expansion: Trigger vCPU affinity rebinding to bind the interrupt thread processing the audio stream to the physical core P-core; Dynamically adjust the GRE tunnel MTU value: Where, is the base bandwidth, is the size of the buffer after expansion, is the buffer size before expansion, is the total bandwidth capacity, For the adjusted value.

[0011] Furthermore, the steps of GRE-over-UDP tunnel encapsulation include: B1: Create a compression dictionary , including source IP, destination IP, GRE Key, and UDP port; B2: Perform differential encoding and Huffman encoding on the data packet header using the following compression algorithm, Where, is the compressed header information, is the header information of the current data packet, For the dictionary database, is the dictionary base index, is the Huffman encoding operation, is the difference extraction operator, is the data connector; B3: When the audio stream is detected to last for a period of time T At 2s, update the compression dictionary: Where, Update the smoothing factor for the dictionary, take , is the updated dictionary, This is the dictionary before updating.

[0012] Furthermore, in step S4, the virtual port forwarding table is updated using the voiceprint-based QoS scheduling algorithm. The updating steps are: C1: Extract MFCC feature vector of audio stream ; C2: Calculate the cosine similarity between the current audio stream and the high-priority audio stream : Where, , is the first MFCC feature vector of the current audio stream A quantity, is the first MFCC feature vector of the high priority audio stream Quantity C3: When When , Qos weight is assigned according to the similarity interval : and schedule priority .

[0013] Further, in step S3, the target MAC is replaced by a path selection based on the network situation and the path quality is selected. And hop count Path; where Where, 、 、 are weight factors, which are 0.5, 0.3, and 0.2 respectively. is the normalized path delay, is the path packet loss rate, Path jitter.

[0014] The beneficial effects of the present invention are: This invention deeply couples the distributed switching architecture with voice service features, breaking through the latency bottleneck of traditional centralized architecture by embedding vSwitch + audio stream feature recognition + dynamic hash mapping in kernel state; end-to-end latency is reduced to below 50ms, voice packet loss rate is <0.3%, and elastic expansion to >1000 nodes is supported.

[0015] Other advantages, objectives and features of the present invention will be described in the following description and will be apparent to those skilled in the art to some extent, or those skilled in the art can be taught from the practice of the present invention. The objectives and other advantages of the present invention can be realized and obtained through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to make the purpose, technical solutions and beneficial effects of the present invention more clear, the present invention provides the following drawings for illustration: Figure 1 Flowchart of an embodiment of the present invention. DETAILED DESCRIPTION

[0017] like Figure 1 As shown, the present invention provides a method for virtualizing audio integrated network resources, comprising the following steps: S1: Build a distributed virtual switching layer covering the subnets and embed a virtual switch instance (vSwitch Instance) into the kernel network stack of each physical host in the cluster. The vSwitch instance includes a voice stream identification module, a virtual port manager, and a policy execution engine. The audio stream identification module identifies audio stream packets based on the Payload Type field in the Real-time Transport Protocol (RTP) header and the feature that the packet length is ≤ 250 bytes. The virtual port manager supports 802.1Q VLAN tagging and is responsible for dynamically adjusting the resource allocation and performance parameters of the virtual port; The policy execution engine stores ACL rules and QoS queues and is responsible for executing predefined optimization policies based on real-time system status; S2: Dynamically map the cloud hosts in the subnet to the physical host group using a consistent hashing algorithm. The mapping relationship satisfies: Where, is the SHA-256 hash function, which is used to map the input to a fixed-length hash value. is the unique identifier of the virtual machine, is the unique identifier of the physical host, is a bitwise exclusive OR operation, is the preset threshold, ; S3: When the cloud host sends an audio packet, the source physical host virtual switch instance performs the following operations: extracting the DSCP field and marking it as EF, encapsulating it through the GRE-over-UDP tunnel, and replacing the destination MAC address with the virtual MAC address of the next-hop virtual switch instance; Among them, the target MAC replacement adopts the path selection based on the network situation and selects the path quality And hop count Path; where Where, 、 、 are weight factors, which are 0.5, 0.3, and 0.2 respectively. is the normalized path delay, is the path packet loss rate, is path jitter; S4: After decapsulation, the virtual switch instance of the target physical host delivers the audio packet to the target cloud host based on the virtual port forwarding table.

[0018] In this solution, the distributed switching architecture is deeply coupled with voice service characteristics. By embedding vSwitch in the kernel, audio stream feature recognition, and dynamic hash mapping, the latency bottleneck of the traditional centralized architecture is broken through. The end-to-end latency is reduced to below 50ms, the voice packet loss rate is <0.3%, and elastic expansion to >1000 nodes is supported.

[0019] In one embodiment of the present invention, the audio stream identification module adopts a dual-threshold trigger mechanism: First threshold: 5 consecutive packet payload sizes Bytes and Packet Interval ; Second threshold: RTP sequence number increment Deviation rate satisfy: Where, is the actual serial number increment, is the theoretical serial number increment, ; When both thresholds are met, the audio-only forwarding channel is activated, which performs: A1: Bypass the traditional Linux TC flow table and directly call the eBPF program for forwarding; A2: Allocate an independent memory pool; A3: Enable forward error correction coding: , in =4, is the forward error correction block, To perform continuous XOR operations from the 1st to the kth data packets, For the i Original audio data packets; If the deviation rate , it is determined to be a discontinuous audio stream or a network attack stream, and the security policy is triggered: the data packet is discarded and an alarm event is generated.

[0020] This solution builds a dual verification model based on the spatiotemporal characteristics of audio packets and combines it with the eBPF kernel bypass technology to solve the jitter sensitivity problem.

[0021] In one embodiment of the present application, the forward error correction coding adopts a non-uniform protection mechanism to improve the anti-packet loss capability of the key audio segment, and the non-uniform protection mechanism comprises: When the silence detection flag is False, it is determined that the key audio frame, and the key audio frame is allocated according to the redundancy ; in the formula, is the current path packet loss rate; When the silence detection flag is True, it is determined that the non-key audio frame, and the non-key audio frame is allocated according to the redundancy .

[0022] In one embodiment of the present application, the optimization strategy executed by the policy execution engine comprises: monitoring the audio stream peak bandwidth in the cluster in real time , when the bandwidth , , the zero-copy buffer expansion is started: Triggering vCPU affinity rebinding, binding the interrupt thread processing the audio stream to the physical core P-core; Dynamically adjusting the GRE tunnel MTU value: In the formula, is the reference bandwidth, is the buffer size after expansion, is the buffer size before expansion, is the total bandwidth capacity, is the adjusted value.

[0023] The scheme directly associates network resource allocation with audio service load, realizes sub-second level elastic response, reduces voice interruption time under burst traffic, and improves resource utilization.

[0024] In one embodiment of the present application, the steps of GRE-over-UDP tunnel encapsulation comprise: B1: Establishing a compression dictionary , including source IP, destination IP, GRE Key, and UDP port; B2: Differentiating and Huffman coding the packet header through the following compression algorithm, In the formula, is the compressed header information, is the header information of the current packet, is the dictionary database, is the dictionary reference index, is the Huffman coding operation, is the difference extraction operator, is the data connector; B3: When the audio stream is detected to last for a period of time T At 2s, update the compression dictionary: Where, Update the smoothing factor for the dictionary, take , is the updated dictionary, This is the dictionary before updating.

[0025] This solution designs an incremental dictionary update mechanism based on the long connection characteristics of audio streams, breaking through the fixed dictionary limitations of the traditional ROHC protocol, compressing the tunnel header overhead from 40 bytes to an average of 8 bytes, and improving bandwidth utilization by 25%.

[0026] In one embodiment of the present invention, in step S4, the virtual port forwarding table is updated using a voiceprint-based QoS scheduling algorithm, and the updating steps are: C1: Extract MFCC feature vector of audio stream ; C2: Calculate the cosine similarity between the current audio stream and the high-priority audio stream : Where, , is the first MFCC feature vector of the current audio stream A quantity, is the first MFCC feature vector of the high priority audio stream Quantity C3: When When , Qos weight is assigned according to the similarity interval : and schedule priority .

[0027] This solution gives high priority to critical voice traffic and improves the MOS score.

[0028] Finally, it should be noted that the above preferred embodiments are only used to illustrate the technical solutions of the present invention and are not limiting. Although the present invention has been described in detail through the above preferred embodiments, those skilled in the art should understand that various changes can be made in form and details without departing from the scope defined by the claims of the present invention.

Claims

1. A method for virtualizing audio integrated network resources, characterized in that: The following steps are involved: S1: Builds a distributed virtual switching layer covering the subnets, embedding a virtual switch instance into the kernel-mode network stack of each physical host in the cluster. The virtual switch instance includes a voice stream identification module, a virtual port manager, and a policy execution engine. S2: Dynamically map the cloud hosts in the subnet to the physical host group using a consistent hashing algorithm. The mapping relationship satisfies: Where, is the SHA-256 hash function, which is used to map the input to a fixed-length hash value. is the unique identifier of the virtual machine, is the unique identifier of the physical host, is a bitwise exclusive OR operation, is the preset threshold, ; S3: When the cloud host sends an audio packet, the source physical host virtual switch instance performs the following operations: extracting the DSCP field and marking it as EF, encapsulating it through the GRE-over-UDP tunnel, and replacing the destination MAC address with the virtual MAC address of the next-hop virtual switch instance; S4: After decapsulation, the virtual switch instance of the target physical host delivers the audio packet to the target cloud host based on the virtual port forwarding table.

2. The method for virtualizing audio integrated network resources according to claim 1, wherein: The audio stream identification module adopts a dual threshold trigger mechanism, in which: The first threshold is: 5 consecutive packet payload sizes Bytes and Packet Interval ; The second threshold is: RTP sequence number increment Deviation rate <0.1 ; Where, is the actual serial number increment, is the theoretical serial number increment, ; When the first threshold and the second threshold are met at the same time, the audio dedicated forwarding channel is activated. , it is determined to be a discontinuous audio stream or a network attack stream, and the security policy is triggered: the data packet is discarded and an alarm event is generated.

3. The method for virtualizing audio integrated network resources according to claim 2, wherein: The steps of executing the dedicated audio forwarding channel include: A1: Bypass the traditional Linux TC flow table and directly call the eBPF program for forwarding; A2: Allocate an independent memory pool; A3: Enable forward error correction coding: , in =4, is the forward error correction block, To perform continuous XOR operations from the 1st to the kth data packets, For the i Raw audio data packets.

4. The method for virtualizing audio integrated network resources according to claim 3, wherein: In step A3, the forward error correction coding adopts a non-uniform protection mechanism, including: When the silence detection flag is False, it is determined to be a key audio frame, and the key audio frame is divided into two parts according to the redundancy. To distribute; where, is the current path packet loss rate; When the silence detection flag is True, it is determined to be a non-critical audio frame, and the non-critical audio frame is divided into two parts according to the redundancy. Make allocation.

5. The method for virtualizing audio integrated network resources according to claim 1, wherein: The optimization strategies implemented by the policy execution engine include: real-time monitoring of the peak bandwidth of the audio stream in the cluster , when bandwidth hour, Enable zero-copy buffer expansion: Trigger vCPU affinity rebinding to bind the interrupt thread processing the audio stream to the physical core P-core; Dynamically adjust the GRE tunnel MTU value: Where, is the base bandwidth, is the size of the buffer after expansion, is the buffer size before expansion, is the total bandwidth capacity, For the adjusted value.

6. The method for virtualizing audio integrated network resources according to claim 1, wherein: The steps of GRE-over-UDP tunnel encapsulation include: B1: Create a compression dictionary , including source IP, destination IP, GRE Key, and UDP port; B2: Perform differential encoding and Huffman encoding on the data packet header using the following compression algorithm, Where, is the compressed header information, is the header information of the current data packet, For the dictionary database, is the dictionary base index, is the Huffman encoding operation, is the difference extraction operator, is the data connector; B3: When the audio stream is detected to last for a period of time T At 2s, update the compression dictionary: Where, Update the smoothing factor for the dictionary, take , is the updated dictionary, This is the dictionary before updating.

7. The method for virtualizing audio integrated network resources according to claim 1, wherein: In step S4, the virtual port forwarding table is updated using the voiceprint-based QoS scheduling algorithm. The updating steps are: C1: Extract MFCC feature vector of audio stream ; C2: Calculate the cosine similarity between the current audio stream and the high-priority audio stream : Where, , is the first MFCC feature vector of the current audio stream A quantity, is the first MFCC feature vector of the high priority audio stream Quantity C3: When When , Qos weight is assigned according to the similarity interval : and schedule priority .

8. The method for virtualizing audio integrated network resources according to claim 1, wherein: In step S3, the target MAC is replaced by a path selection based on the network situation and the path quality is selected. And hop count Path; where Where, 、 、 are weight factors, which are 0.5, 0.3, and 0.2 respectively. is the normalized path delay, is the path packet loss rate, Path jitter.

Images