A data center network architecture with ring-shaped wireless and wired hybrid interconnection

Abstract

The present invention discloses a data center network architecture with a ring-shaped wireless and wired hybrid interconnection, which is characterized by including: a network architecture that combines wired and wireless collaboration: transmitting large data traffic through a wired network, wherein the large data traffic is traffic with data packets ≥ 1MB, and the self-healing characteristics of the bidirectional ring topology and the multi-layer redundancy design enable the network to have strong fault tolerance. When a single ring is disconnected, the other ring maintains transmission, and the multi-layer expansion structure also provides more redundant paths. Even if some links or nodes fail, the network can still maintain normal operation, greatly improving the reliability of the network. From the experimental data, under different load intensities, the network delay shows a stable linear growth trend, and there is no explosive growth due to increased load, which proves that the architecture has good stability.

Description

Technical Field

[0001] The present invention relates to the field of data center network technology, and in particular to a data center network architecture with ring-shaped wireless and wired hybrid interconnection. Background Art

[0002] With the rapid development of emerging technologies such as cloud computing and big data, data center networks are facing increasing traffic demands, placing higher demands on network bandwidth, latency, flexibility, and scalability. Traditional data center network architectures primarily rely on wired networks, which suffer from poor scalability and low flexibility, making them unable to meet the needs of large-scale data center networks.

[0003] To improve network flexibility and scalability, some researchers have proposed hybrid network architectures that incorporate wireless communication technologies. This architecture builds a hybrid interconnection within the data center, combining traditional wired and wireless networks. The wired network is responsible for transmitting large data flows, ensuring high throughput and low latency; the wireless network is used to transmit small data flows, supporting flexible inter-node communication and cross-data center interconnection. Through appropriate traffic scheduling mechanisms, the advantages of both wireless and wired networks can be fully utilized in different scenarios, improving network flexibility and scalability, and reducing energy consumption while ensuring network performance.

[0004] However, existing hybrid network architectures still have deficiencies in implementation details and performance evaluation, and cannot fully meet the needs of large-scale data center networks. Specifically,

[0005] 1. Traditional wired network architecture has poor scalability and low flexibility, making it difficult to adapt to the traffic needs of large-scale data centers.

[0006] 2. The existing hybrid network architecture has defects in implementation details and performance evaluation, and cannot fully meet the requirements of large-scale data centers.

[0007] 3. Existing methods have limited generalization capabilities for architectures and are difficult to apply to a wider range of network types.

[0008] 4. Existing methods lack intelligence and are unable to quickly respond to dynamic changes in network status and traffic. Summary of the Invention

[0009] In order to overcome the deficiencies of the prior art, one of the objectives of the present invention is to provide a data center network architecture with a ring-shaped wireless and wired hybrid interconnection.

[0010] One of the purposes of the present invention is achieved by the following technical solution:

[0011] A ring-shaped wireless and wired hybrid interconnected data center network architecture includes:

[0012] Hybrid wired and wireless collaborative network architecture: large data traffic is transmitted through wired networks, where the data traffic refers to traffic with data packets ≥ 1MB, to ensure high throughput and low latency; small data traffic is transmitted through wireless networks, where the data traffic refers to traffic with data packets < 1MB, supporting flexible interconnection across nodes, layers, and data centers.

[0013] Ring topology and multi-layer expansion structure: A bidirectional ring topology is used within the basic unit to transmit data in a fixed direction, achieving predictable latency and fault recovery capabilities. If one ring is disconnected, the other ring maintains transmission. Based on wireless links, multiple ring units are expanded into one or more layers of interconnected networks. The first layer of the network consists of 9 basic units corresponding to a 3×3 array. The multi-layer network can be expanded to a maximum of 3 layers, each containing 9 basic units, to improve the scalability of the network scale.

[0014] Dynamic traffic scheduling and load balancing mechanism: Based on software-defined networking, it monitors network congestion in real time and dynamically adjusts the traffic distribution ratio between wired and wireless links. When node or link congestion is detected, it automatically migrates part of the traffic to idle links to achieve global load balancing.

[0015] Load classification and intelligent routing strategy: Automatically selects transmission paths based on a packet size threshold of 1MB. Optimizes routing decisions based on network load status, prioritizing wireless links for low-latency transmission of small data across data centers.

[0016] Energy consumption optimization and reliability enhancement: Through differentiated division of labor between wired and wireless networks, the frequency of use of high-power wireless devices in large data transmission is reduced, reducing overall energy consumption. The bidirectional transmission characteristics and multi-layer redundancy of the ring topology are utilized to improve network fault tolerance and reliability.

[0017] Cross-layer transmission function of the wireless server: The intermediate wireless server node is dedicated to forwarding small data packets across layers and cells. Four intermediate wireless server nodes are configured in each cell to reduce the load pressure on the wired core link.

[0018] Furthermore, the wireless network server adopts WLAN wireless communication technology, with an operating frequency of 2.4 GHz or 5 GHz and a transmission rate of 300 Mbps.

[0019] Furthermore, the wired network server adopts 10 Gigabit Ethernet technology with a transmission rate of 10 Gbps.

[0020] Furthermore, in the dynamic traffic scheduling and load balancing mechanism, the traffic scheduling mechanism adopts a strategy based on software-defined networking to dynamically adjust the traffic distribution ratio between wired and wireless networks according to network congestion conditions.

[0021] Furthermore, in the basic unit network, the four middle nodes are wireless network server architectures, responsible for transmitting small data packets and performing cross-node and cross-layer transmission; the wireless network servers are surrounded by wired network servers, responsible for large data transmission.

[0022] Compared with the prior art, the present invention has the following beneficial effects:

[0023] 1. Significantly Improved Network Flexibility and Scalability

[0024] By integrating wireless communication technology with traditional wired networks, a hybrid internetwork has been constructed, providing multiple path options for nodes within the data center, freeing the network from being limited to a single wired connection. A bidirectional ring topology is employed within the base unit, ensuring network connectivity when one ring is disconnected while the other maintains transmission. Furthermore, wireless links allow multiple ring units to be expanded into a single-layer (e.g., a 3×3 array) or multi-layer (up to three) internetwork, with each layer containing nine base units. This multi-layer expansion structure allows the network to flexibly scale according to actual needs, increasing the network's elastic expansion capability by fivefold. This effectively addresses the poor scalability of traditional wired networks and greatly enhances the network's flexibility and scalability.

[0025] 2. Network performance is optimized and energy consumption is reduced

[0026] Wired networks are specifically responsible for transmitting large data flows, with data packets ≥1MB, leveraging their high throughput and low latency to ensure efficient and stable transmission of large data. Wireless networks, on the other hand, transmit small data flows, with data packets less than 1MB, leveraging their flexible transmission capabilities. This differentiated division of labor enables the rational utilization of both wireless and wired network resources. While ensuring high network throughput and low latency, it also reduces the frequency of high-power wireless devices used for large data transmission, thereby reducing overall energy consumption and improving network resource utilization efficiency.

[0027] 3. Node communication and cross-data center interconnection are more flexible

[0028] The wireless network focuses on small data traffic transmission, supporting flexible inter-node communication and cross-data center interconnection. Intermediate wireless server nodes (four per unit) are dedicated to forwarding small data packets across layers and cells. They enable rapid transmission of small data between nodes, layers, and data centers, improving network connectivity and reliability and making communications within the data center more flexible and efficient.

[0029] 4. Significantly Enhanced Load Balancing Capabilities

[0030] Software-defined networking (SDN) monitors network congestion in real time and dynamically adjusts the traffic distribution ratio between wired and wireless links. When node or link congestion is detected, traffic is automatically shifted to idle links to achieve global load balancing. This dynamic traffic scheduling mechanism avoids network congestion, improves overall network performance, and ensures stable operation under varying load conditions.

[0031] 5. Improved network stability and reliability

[0032] The self-healing properties of the bidirectional ring topology and its multi-layer redundancy design provide the network with robust fault tolerance. If one ring fails, the other ring maintains transmission. The multi-layered extended structure also provides more redundant paths, ensuring the network remains operational even if some links or nodes fail, significantly improving network reliability. Experimental data shows that network latency exhibits a stable linear growth trend under varying load intensities, with no explosive increases due to increased load, demonstrating the architecture's robustness.

[0033] 6. Data transmission capacity and scalability are superior to traditional architectures

[0034] Compared with the SpineLeaf network architecture, the hub-ring network architecture of the present invention performs better in terms of latency and throughput. The average latency is reduced by 49%. In the comparison between the multi-layer network and the large-scale SpineLeaf network, the average latency is reduced by 64%. The overall throughput is higher than the SpineLeaf network architecture. The maximum throughput of the single-layer hub-ring network reaches 172Gbps, which is about 72% higher than the SpineLeaf network under the same load. At the same time, the latency of the multi-layer architecture (3 layers) is only increased by 15.7% compared to the single-layer network, and the throughput maintains a linear scalability of 91.9%, which fully verifies the excellent performance of this architecture in data transmission capacity and scalability.

[0035] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the following preferred embodiments are specifically cited and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 This is a diagram of a single annular unit in this embodiment;

[0037] Figure 2 This is a schematic diagram of a 3*3 large-scale data center in this embodiment;

[0038] Figure 3 Comparative analysis of hub-ring network delay;

[0039] Figure 4Comparative analysis of hub-ring network throughput;

[0040] Figure 5 Comparative analysis of latency between Hub-Ring network and SpineLeaf network;

[0041] Figure 6 Comparative analysis of throughput between Hub-Ring network and SpineLeaf network. DETAILED DESCRIPTION

[0042] The present invention will be further described below in conjunction with the accompanying drawings and specific implementation methods. It should be noted that, under the premise of no conflict, the various embodiments or technical features described below can be arbitrarily combined to form new embodiments.

[0043] It should be noted that when a component is referred to as being "fixed to" another component, it may be directly on the other component or there may also be a central component. When a component is considered to be "connected to" another component, it may be directly connected to the other component or there may also be a central component. When a component is considered to be "set on" another component, it may be directly set on the other component or there may also be a central component. The terms "vertical", "horizontal", "left", "right" and similar expressions used herein are for illustrative purposes only.

[0044] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this invention pertains. The terms used in this specification of the present invention are for the purpose of describing specific embodiments only and are not intended to limit the present invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items. Specific embodiments

[0046] 1. Hybrid Network Architecture Construction

[0047] (1) Basic unit network construction

[0048] 1. Node deployment

[0049] A basic unit network is built inside the data center. Each basic unit contains four wireless network server nodes in the central area and wired network server nodes in the periphery.

[0050] The wireless network server uses WLAN wireless communication technology, with an operating frequency of 2.4GHz or 5GHz and a transmission rate of 300Mbps. It is dedicated to cross-node and cross-layer forwarding of small data packets (<1MB).

[0051] The wired network server uses 10 Gigabit Ethernet technology with a transmission rate of 10Gbps, responsible for high-throughput transmission of large data packets (≥1MB).

[0052] 2. Topological connection

[0053] The base unit uses a bidirectional ring topology to transmit data in a fixed direction. If a link on one ring is disconnected, the other ring automatically maintains data transmission, achieving predictable latency (e.g., single-layer latency as low as 5ms at a load strength of 0.1) and fault recovery capabilities.

[0054] (2) Multi-layer network expansion

[0055] 1. Single-layer expansion

[0056] Nine basic units (3×3 array) are connected to form a large-scale network through wireless links. Small data interaction is achieved between units through wireless network server nodes, while wired network server nodes are responsible for large data transmission within the unit.

[0057] 2. Multi-layer expansion

[0058] It can be expanded to a maximum of three interconnected networks, with each layer consisting of nine basic units, interconnected by wireless links. For example, a two-layer network consists of 18 basic units, and a three-layer network consists of 27 basic units, supporting elastic expansion of network scale.

[0059] 2. Traffic Scheduling and Transmission Mechanism

[0060] (1) Load classification and path selection

[0061] 1. Data diversion standards

[0062] Automatically select the transmission path based on the packet size threshold (1MB):

[0063] When the data packet is ≥1MB, it is transmitted through a wired network, ensuring high throughput (such as the maximum throughput of a single-layer hub network is 172Gbps) and low latency (multi-layer latency is 32.91ms when the load intensity is 1.0).

[0064] When the data packet is less than 1MB, it is transmitted through a wireless network, supporting flexible interconnection across nodes and data centers, and achieving significant sub-millisecond packet transmission efficiency.

[0065] (2) Dynamic load balancing

[0066] Real-time monitoring and adjustment

[0067] Based on software-defined networking (SDN), the system monitors network congestion in real time. When it detects that a node or link load exceeds a threshold (e.g., load intensity ≥ 0.7), it automatically migrates some traffic from the congested link to an idle link. For example, when a wired link is congested, some small data flows are switched to the wireless link, achieving global load balancing and preventing sudden increases in latency.

[0068] 3. Energy Consumption Optimization and Reliability Design

[0069] (1) Differentiated energy consumption management

[0070] Equipment division of labor strategy

[0071] Wired networks handle large data traffic transmission, reducing the use of high-power wireless devices. For example, when transmitting a 10MB data packet, wired networks are prioritized, reducing wireless device energy consumption by approximately 30% compared to traditional hybrid architectures.

[0072] (2) Redundant fault-tolerant mechanism

[0073] 1. Ring topology self-healing

[0074] The bidirectional ring structure of the basic unit supports transmission on the other ring when one ring fails. Experimental data shows that when a ring in a single-layer hub ring network is disconnected, the latency only increases by about 10% (for example, from 5ms to 5.5ms), and the throughput remains above 85%.

[0075] 2. Multi-layer redundant expansion

[0076] In a multi-layer network architecture, each layer is interconnected through redundant wireless links. When some links in a layer fail, data can be forwarded through paths in other layers, improving network reliability.

[0077] 3. Implementation of cross-layer transmission function

[0078] (1) Wireless server node configuration

[0079] Each base unit is equipped with four intermediate wireless server nodes, dedicated to forwarding small data packets across layers (e.g., from Layer 2 to Layer 3) and across units (e.g., different 3x3 array units). For example, when transmitting a 500KB packet across data centers, relaying it through the wireless server nodes reduces latency by approximately 40% compared to forwarding over traditional wired core links.

[0080] IV. Experimental Verification Scenario

[0081] (1) Single-layer hub network test

[0082] 1. Load intensity 0.1

[0083] The data packet size is 3B, transmitted over a wireless network, with an average latency of 5ms and a throughput of 17.2Gbps.

[0084] The data packet size is 1.5MB, transmitted through a wired network, with a latency of 8ms and a throughput of 1000Mbps (nominal wired rate).

[0085] (2) Layer 3 Hub Network Test

[0086] 1. Load intensity 1.0

[0087] A 200KB data packet is transmitted across layers and forwarded through wireless server nodes, with an average latency of 32.91ms and a throughput of 158Gbps.

[0088] A 2MB data packet is transmitted within the layer through a wired network with a latency of 28.44ms and a throughput of 172Gbps (close to the theoretical maximum).

[0089] (2) Comparison with SpineLeaf Network

[0090] 1. Large-scale scenario (324 nodes)

[0091] The multi-layer hub-and-ring network of the present invention has a latency of 32.91ms and a throughput of 158Gbps at a load intensity of 1.0.

[0092] The SpineLeaf network has a latency of 110ms and a throughput of 110 Gbps. The latency of the present invention is reduced by 70.1% and the throughput is increased by 43.6%.

[0093] Experimental data and analysis

[0094] 1. Basic experimental data

[0095] (1) Load Strength and Data Packet Parameters

[0096]

[0097] 2. Hub-Ring Network Architecture Performance Data

[0098] (1) Comparison of network latency at different levels (unit: ms)

[0099] Load strength Single-layer hub network Double-layer hub network Multi-layer hub network 0.1 5 6 7 0.2 8 9 10 0.3 12 13 14 0.4 15 16 17 0.5 18 19 20 0.6 21 22 23 0.7 24 25 26 0.8 26 27 28 0.9 28 29 31 1.0 28.44 30.5 32.91

[0100] (3) Comparison of network throughput at different levels (unit: Gbps)

[0101] Load strength Single-layer hub-and-ring network Double-layer hub network Multi-layer hub network 0.1 17.2 16.5 15.8 0.2 34.4 33 31.6 0.3 51.6 49.5 47.4 0.4 68.8 66 63.2 0.5 86 82.5 79 0.6 103.2 99 94.8 0.7 120.4 115.5 110.6 0.8 137.6 132 126.4 0.9 154.8 148.5 142.2 1.0 172 165 158

[0102] 3. Comparison Data with SpineLeaf Network Architecture

[0103] (1) Latency comparison (unit: ms)

[0104]

[0105]

[0106] (2) Throughput comparison (unit: Gbps)

[0107]

[0108] 4. Data Conclusion

[0109] 1. Latency performance: When the load intensity is 1.0, the latency of the multi-layer hub-ring network of the present invention is 32.91ms, which is 64% lower than that of the SpineLeaf large-scale network (110ms). The latency increases linearly without explosive growth.

[0110] 2. Throughput performance: The maximum throughput of a single-layer hub-ring network reaches 172Gbps, which is approximately 72% higher than the SpineLeaf network under the same load. The throughput increases linearly with the increase in load and has excellent stability.

[0111] 3. Scalability advantage: Compared with a single-layer network, the multi-layer architecture (3 layers) only increases latency by 15.7%, and the throughput maintains a linear scalability of 91.9%, verifying the scalability of the architecture.

[0112] Experimental Analysis of Wireless and Wired Hybrid Interconnection Data Center Network Architecture

[0113] 1. Experimental Process

[0114] 1. Experimental Setup and Parameters

[0115] Simulation platform: Experiments are conducted based on the OMNeT++ simulation platform and the INET framework.

[0116] Network model:

[0117] Pivot Ring Network: The base unit consists of four wireless server nodes in the middle and eight wired server nodes around them, connected in a bidirectional ring topology. A single-layer network consists of nine base units in a 3×3 configuration, which can be expanded to two or three layers.

[0118] SpineLeaf network: consists of two layers of spine switches and leaf switches, and can be expanded to small-scale, medium-scale, and large-scale networks with 108, 216, and 324 server nodes, respectively.

[0119] Key parameters:

[0120] Wired servers are connected using 10Gbps optical fiber with a propagation delay of 0.05μs.

[0121] The wireless server uses WLAN technology, with an operating frequency of 2.4GHz or 5GHz and a transmission rate of 300Mbps.

[0122] Data packet size: Wired transmission ≥1MB, wireless transmission <1MB.

[0123] The simulation time is set to 1000 seconds and the sending interval is 460-500ms.

[0124] 2. Performance index test

[0125] Latency: The time difference between the data sending end and the receiving end, reflecting the network response speed.

[0126] Throughput: The amount of data successfully transmitted per unit time, reflecting the network's ability to transmit data.

[0127] 3. Simulation method

[0128] Topology construction: Build a network model using the OMNeT++ NED file, graphically drag and drop modules in the Design interface, or define the network structure using the NED language in the Source interface.

[0129] Parameter configuration: Use the INI file to set parameters such as simulation time, packet size, and sending interval, and configure UDPBasicApp as the sender and UDPSink as the receiver.

[0130] Data logging: Record key data such as end-to-end latency and throughput, and export the data through OMNeT++ analysis files (.anf) for processing.

[0131] Comparative experiment: Simulate the Hub-Ring network and the SpineLeaf network under the same configuration and compare their performance.

[0132] 2. Experimental Data and Results

[0133] To facilitate data statistics, simulated data of normalized service intensity is set. The following table shows the corresponding relationship between service intensity and data packet size:

[0134] Business intensity: There are 10 levels from 0.1 to 1.0, representing the degree of business. The larger the value, the busier the business.

[0135] Average sending interval: 480 milliseconds under all service intensities, which is the time interval setting for data packet sending.

[0136] Data packet size: The value increases from 0.1 to 1.0, increasing to 3, 6, 9, 12, 15, 18, 21, 24, 27, and 30 bytes, respectively. The higher the service intensity, the larger the data packet. Different data sizes are used to simulate different service intensities and assist in observing network performance.

[0137] Table 1 Simulation data for setting normalized service intensity

[0138]

[0139] 1. Hub network performance data

[0140] (1) Hub network delay

[0141] As shown in the figure, simulation data results summarize the latency variations of three different networks under varying network traffic intensities. The vertical axis represents latency, and the horizontal axis represents normalized traffic intensity, increasing from left to right.

[0142] Figure 3 The data shows that the average latency of a single-layer network is approximately 5ms at a load intensity of 0.1, reaching 28.44ms at a load intensity of 1.0. The average latency of a multi-layer network at a load intensity of 1.0 is 32.91ms. Latency increases linearly with increasing load intensity, with no explosive growth, demonstrating good network stability and scalability.

[0143] (2) Hub network throughput

[0144] like Figure 4 As shown in Figure 2, simulation data results show the throughput of three different networks under different network traffic intensities. The vertical axis represents throughput, and the horizontal axis represents normalized traffic intensity, increasing from left to right.

[0145] When traffic intensity decreases, the network load becomes lighter, competition and congestion are reduced, and fewer packets are transmitted, resulting in a decrease in actual throughput. However, under low traffic intensity, link performance is stable and the packet transmission success rate is high. When traffic intensity increases, the increase in data volume increases throughput. However, if traffic intensity continues to increase, network resource congestion will cause packet loss, which will limit throughput growth. Figure 4 The data shows that when the business intensity of network models at different levels increases, the throughput increases almost linearly, reaching a maximum of 172Gbps, indicating that the network architecture has good data transmission capabilities.

[0146] 2. Comparison data with SpineLeaf network

[0147] Latency comparison: Figure 5 The results show that the average latency of a hub-and-ring network is 49% lower than that of a SpineLeaf network. In a test comparing a multi-layer hub-and-ring network to a large-scale SpineLeaf network, the average latency reduction was 64%. Under high traffic intensity, the latency of a multi-layer hub-and-ring network is even lower than that of a SpineLeaf network of all sizes.

[0148] Throughput comparison: At each service intensity, the Hub-Ring network transmits more data than the SpineLeaf network. The throughput of both networks increases linearly, with the Hub-Ring network demonstrating better data transmission capabilities and stability.

[0149] Conclusion: The hub-and-ring network architecture, a hybrid of wired and wired interconnection, demonstrates superior latency and throughput. Compared to the SpineLeaf network, the hub-and-ring network offers significant advantages in latency and scalability, while also offering higher throughput. This network architecture combines the high throughput of wired networks with the flexibility of wireless networks, ensuring high performance while reducing energy consumption. It supports the flexible expansion of large-scale data centers and provides new insights and directions for future data center network architecture design.

[0150] The above embodiments are only preferred embodiments of the present invention and cannot be used to limit the scope of protection of the present invention. Any non-substantial changes and replacements made by technicians in this field on the basis of the present invention fall within the scope of protection required by the present invention.

Claims

1. A ring-shaped wireless and wired hybrid interconnected data center network architecture, characterized in that: include: Hybrid wired and wireless collaborative network architecture: Large data traffic (data packets ≥ 1MB) is transmitted over wired networks to ensure high throughput and low latency; small data traffic (data packets < 1MB) is transmitted over wireless networks, supporting flexible cross-node, cross-layer, and cross-data center interconnection. Ring topology and multi-layer expansion structure: The basic units use a bidirectional ring topology connection, transmitting data in a fixed direction to achieve predictable latency and fault recovery capabilities. If one ring is disconnected, the other ring maintains transmission. Multiple ring units are expanded into a single-layer network or multi-layer interconnected network based on wireless links. The first layer consists of nine basic unit networks corresponding to a 3×3 array. The multi-layer network can be expanded to up to three layers, each containing nine basic unit networks, to improve network scalability. Dynamic traffic scheduling and load balancing mechanism: Based on software-defined networking, it monitors network congestion in real time and dynamically adjusts the traffic distribution ratio between wired and wireless links. When node or link congestion is detected, it automatically migrates some traffic to idle links to achieve global load balancing. Load classification and intelligent routing strategies: Automatically selects transmission paths based on a packet size threshold of 1MB. Optimizes routing decisions based on network load status, prioritizing wireless links for low-latency transmission of small data across data centers. Energy consumption optimization and reliability enhancement: Through differentiated division of labor between wired and wireless networks, the frequency of high-power wireless devices used in large data transmission is reduced, thus reducing overall energy consumption. The bidirectional transmission characteristics and multi-layer redundancy of the ring topology are utilized to improve network fault tolerance and reliability. Cross-layer transmission function of the wireless server: The intermediate wireless server node is dedicated to forwarding small data packets across layers and units. Each unit is configured with 4 intermediate wireless server nodes to reduce the load pressure of the wired core link. In the basic unit network, the four intermediate nodes are wireless network server architectures, responsible for transmitting small data packets and performing cross-node and cross-layer transmission; the wireless network server is surrounded by wired network servers, responsible for large data transmission.

2. The ring-shaped wireless and wired hybrid interconnected data center network architecture according to claim 1, characterized in that: The wireless network server adopts WLAN wireless communication technology, with an operating frequency of 2.4GHz or 5GHz and a transmission rate of 300Mbps.

3. The ring-shaped wireless and wired hybrid interconnected data center network architecture according to claim 1, characterized in that: The wired network server uses 10 Gigabit Ethernet technology with a transmission rate of 10Gbps.

4. The ring-shaped wireless and wired hybrid interconnected data center network architecture according to claim 1, characterized in that: In the dynamic traffic scheduling and load balancing mechanism, the traffic scheduling mechanism adopts a strategy based on software-defined networking to dynamically adjust the traffic distribution ratio between wired and wireless networks according to network congestion.

Images