A network security device N+1 cluster hot backup system and method based on chain physical cascade

Abstract

The application discloses a network security device N+1 cluster hot backup system and method based on a chain type physical cascade, which comprises N main devices and at least one public backup machine, and each device is connected into a daisy chain topology through a special cascade interface. Each device is internally integrated with an encryption module and a physical switching matrix composed of multiple mechanical latching type optical switches. The system defines three modes of service processing, fault source bypass and transparent relay: when a device suddenly fails or is powered off, the optical switch automatically maintains the redirected optical path by relying on the mechanical latching characteristics, and physically bypasses the local port traffic to the cascade channel; after detecting the cascade signal, the downstream normal node directly establishes a straight-through transparent optical path by using the internal optical switch matrix, so that the fault traffic is relayed at the physical layer. The application realizes "physical layer complete unloading" of fault traffic, realizes zero computing power consumption for intermediate nodes, has a forwarding delay of less than 1 mu s, effectively eliminates the performance avalanche problem of a high computing power cluster, and has high availability of power failure self-healing.

Description

Technical Field

[0001] This invention relates to the field of network communication security architecture, and in particular to an N+1 cluster high-availability hot standby system and method based on physical layer optical path switching technology and mechanical latching characteristics. Background Technology

[0002] In enterprise-level network egress, security devices such as VPN gateways and encryption machines are critical nodes. Because encryption modules require a huge amount of CPU computing power to perform encryption and decryption operations (such as the Chinese national cryptographic algorithms SM1 / SM4), the performance bottleneck of the device often lies in computing resources. Existing high availability solutions mainly include dual-machine hot standby (1+1) and N+1 clusters based on routing protocols (such as OSPF). This approach has significant drawbacks: 1. Dual-machine hot standby is too costly, as each main machine needs to be equipped with a standby machine that is idle for a long time, resulting in low resource utilization. 2. Traditional N+1 clusters heavily rely on software-layer routing and forwarding. When the primary device fails, the failed traffic switching to the common backup device often requires the CPU of intermediate healthy nodes to perform protocol parsing and software forwarding. For high-computing-power-consuming encryption devices, forcing intermediate healthy nodes to process and forward additional failed traffic can easily lead to CPU overload of intermediate nodes, sluggishness of their own encryption services, or even direct system crashes, thus triggering a chain reaction of "avalanche effects" in the network. Therefore, there is an urgent need for a low-level hardware-level switching technology that can achieve N+1 hot standby and enable intermediate transit nodes to achieve "zero computing power consumption" for faulty traffic. Summary of the Invention

[0003] The purpose of this invention is to provide an N+1 cluster hot standby system and method for network security devices, which utilizes a mechanically latched optical switch to construct a switching matrix and implements it through an independent physical cascading interface. This invention innovatively introduces a "transparent relay mode" in logic by defining a physical switching matrix in hardware, thereby achieving complete physical layer offloading of faulty traffic. The beneficial effects of this invention are as follows: 1. Zero computing power consumption (completely eliminates the avalanche effect): Innovatively utilizes an optical switch matrix to establish a "physical layer offloading" mechanism. When the transit traffic of a faulty device passes through an intermediate normal node, it is directly transmitted at the optical physical layer, without entering or occupying the CPU and encryption module resources of the intermediate node, thus absolutely guaranteeing the stability of the intermediate node's own services. 2. Physical-level high reliability (power-off self-healing): The switching matrix uses optical switches with mechanical latching characteristics. Even if the device node experiences a sudden and complete power outage, its internal optical physical channel will still remain in the conducting state by mechanical force, without electrical reset, thus achieving true physical self-healing after power failure. 3. Extremely low latency and scalable networking: Hardware-level optical path switching response time is less than 10ms; through a transparent offloading mechanism at the pure physical layer, the forwarding latency of transit relays is less than 1μs. Simultaneously, the system scientifically limits the number of daisy-chained units to no more than 16 through rigorous physical layer link budgeting, perfectly balancing the relationship between optical link insertion loss and backup receiver sensitivity. Attached Figure Description Figure 1 The diagram shows the N+1 chain-type physical cascade system architecture and fault flow provided in the embodiments of the present invention (the main device A is in a power failure / abnormal state, and the main device B is in a transparent relay state). Figure 2 This diagram illustrates the physical connection of the optical switch under different logic control modes according to an embodiment of the present invention. The left diagram shows the normal service connection state (P1-P3 are on), and the right diagram shows the redirection / relay state (P1-P2 are on). Detailed Implementation The present invention will now be described in further detail with reference to the accompanying drawings. Example 1: Device Hardware Interface and Daisy-Chain Topology like Figure 1 As shown, this system includes multiple master devices (such as master device A and master device B) and at least one common backup device C. All devices are equipped with external WAN ports, LAN ports, and dedicated cascading input interfaces (including cascading input WAN and cascading input LAN) and cascading output interfaces (including cascading output WAN and cascading output LAN) for cluster communication. During network setup, each device is connected via fiber optic patch cords, linking the "cascade output interface" of the previous level device to the "cascade input interface" of the next level device, thus forming a typical daisy chain topology at the physical layer. A common backup machine C is deployed at the end of the data link, serving as a highly available backup computing power pool for unified allocation across the entire cluster. Example 2: Internal switching matrix and optical switch logic of the device are as follows Figure 1 and Figure 2 As shown, in addition to the core encryption module, main control unit and optical power monitoring module integrated inside the device, its core innovation lies in the design of a physical switching matrix consisting of four optical switches. In this embodiment, all optical switches are mechanically latched optical switches (LatchType) with power-off retention characteristics. For example... Figure 2As shown, this type of optical switch includes a common terminal P1 and two throwing terminals P2 and P3. The normal operating state of the optical switch is defined as P1-P3 being turned on, and the fault / bypass state is defined as P1-P2 being turned on. To achieve independent business logic and cascading send / receive logic, the switching matrix is ​​finely divided into: WAN side input optical switch: P1 connects to the WAN port, P3 connects to the encryption module, and P2 connects to the internal bypass. WAN-side cascaded optical switches: P1 is connected to the cascaded output WAN, P3 is connected to the cascaded input WAN, and P2 is connected to the aforementioned internal bypass. LAN-side output optical switch: P1 connects to the LAN port, P3 connects to the encryption module, and P2 connects to the internal backhaul bypass. LAN-side cascaded optical switches: P1 connects to the cascaded output LAN, P3 connects to the cascaded input LAN, and P2 connects to the aforementioned internal backhaul bypass (Note: Since the LAN-side backhaul traffic flows in the reverse direction from bottom to top on the physical link, this traffic actually enters through the cascaded output LAN port of the local machine. Therefore, the port definitions of P1 / P3 maintain a physically symmetrical architecture with the WAN side, which can perfectly realize reverse flow pass-through). Example 3: Analysis of Three Core Working Modes and Data Flow Based on the aforementioned optical switch matrix, the main control unit intelligently or passively switches between the following three modes according to the device status: 1. Business processing mode (main equipment in normal operating state, refer to...) Figure 1 (Left side state) (1) Triggering condition: The device self-test is completely normal. (2) Optical switch status: All four optical switches remain in the normal state, i.e., P1-P3 are turned on. (3) Traffic flow: External service traffic enters from the WAN port, is introduced into the encryption module through optical switches P1-P3, and after processing, is normally output from the LAN port through optical switches P3-P1 on the LAN side. At this time, a direct backup physical channel is formed between the cascade input and cascade output ports through the P3-P1 path, and it is in standby mode. 2. Fault source bypass mode (main device A power failure / fault state, refer to...) Figure 1 Right-side main device A): (1) Triggering conditions: The main control unit detects a fault or experiences a physical power failure. (2) Optical switch status: The four optical switches automatically switch to the fault state by mechanical force, that is, P1-P2 are turned on. (3) Traffic flow: External WAN traffic no longer enters the invalid encryption module, but is directly guided to the "cascaded output WAN port" through the P1-P2 bypass of the two sets of optical switches and sent to the next level node; at the same time, the return traffic processed by the standby machine and transmitted back to the "cascaded output LAN port" of the local machine in reverse along the daisy chain is accurately guided to the output of the local machine's LAN port through the P1-P2 bypass of the two sets of optical switches on the LAN side. 3. Transparent Relay Mode (Core Creative Highlight, see reference) Figure 1 Right-side main device B): (1) Triggering conditions: The machine itself is operating normally, but an optical signal sent from upstream is detected at its cascade interface. (2) Optical switch state: Maintain the normal physical state, that is, all four optical switches are kept P1-P3 conducting. (3) Traffic Flow: The stunning "complete physical layer offloading" occurs here. Since the cascaded optical switches keep P1-P3 conducting, the upstream fault traffic enters from the "cascaded input WAN port" and flows directly out from the "cascaded output WAN port" along the P3-P1 optical path. Similarly, the LAN-side return traffic enters from the "cascaded output LAN port" and is transmitted upwards unimpeded from the "cascaded input LAN port" along the P1-P3 optical path. (4) Core advantages: such as Figure 1 As clearly shown, in this mode, relay traffic is transmitted purely through the independent physical channel of the optical switch, completely isolated at the physical layer from normal services being processed by the local machine through the WAN / LAN port. Real-world testing demonstrates that the forwarding latency in this mode is less than 1μs, completely bypassing the CPU and encryption module, achieving true "zero computing power consumption," and physically immune to the performance avalanche effect during cluster concurrency failures. Example 4: Calculation of System Limiting Performance and Physical Cascading Scale In this physical cascade system, each time an optical signal passes through a mechanically latched optical switch and its flange connector, an insertion loss of approximately 0.5 dB to 0.8 dB is generated. To ensure that the optical power received by the last public standby device C in the daisy chain remains within its high-sensitivity reception threshold (ensuring that the optical signal is not distorted and there is no packet loss), this embodiment limits the maximum number of cascaded cluster master devices N to 16 based on a strict optical physical layer link budget. Calculation Basis: Assume the system uses optical modules with a transmit power of -3dBm and a receive sensitivity threshold of -24dBm. If the system cascades N=16 master devices, and the fault occurs at the head of the chain, the cumulative optical switch insertion loss transmitted to the tail of the chain is approximately 16 × 0.6dB = 9.6dB. Adding fiber optic patch cord loss and appropriate redundancy, the total link loss is strictly controlled within 13dB. At this point, the actual received optical power of the backup unit C at the end is approximately -16dBm, still possessing an absolute safety power margin of approximately 8dB from the -24dBm failure threshold. Through this rigorous physical scale hardware constraint, this solution meets the needs of large enterprise network ports for ultra-large-scale clusters while ensuring the absolute reliability of optical link physical layer transmission.

Claims

1. A network security device N+1 cluster hot standby system based on chained physical cascading, characterized in that, It includes N main devices and at least one common backup device, where N is a positive integer not greater than 16; the main devices and the common backup device are cascaded sequentially through fiber optic links to form a daisy-chain topology; each device is equipped with: a WAN port, a LAN port, a cascaded input WAN port, a cascaded output WAN port, a cascaded input LAN port, and a cascaded output LAN port; each device integrates an encryption module, an optical power monitoring module, a main control unit, and a switching matrix composed of four mechanically latched optical switches; each optical switch has a common terminal P1, a first throwing terminal P3, and a second throwing terminal P2; The switching matrix includes a WAN-side input optical switch, a WAN-side cascaded optical switch, a LAN-side output optical switch, and a LAN-side cascaded optical switch, and its physical connection topology is configured as follows: The P1 terminal of the WAN-side input optical switch is connected to the WAN port, and the P3 terminal is connected to the input terminal of the encryption module. The P1 terminal of the WAN-side cascaded optical switch is connected to the cascaded output WAN port, the P3 terminal is connected to the cascaded input WAN port, and the P2 terminal is internally connected to the P2 terminal of the WAN-side input optical switch. The P1 terminal of the LAN-side output optical switch is connected to the LAN port, and the P3 terminal is connected to the output terminal of the encryption module. The P1 terminal of the LAN-side cascaded optical switch is connected to the cascaded output LAN port, the P3 terminal is connected to the cascaded input LAN port, and the P2 terminal is internally connected to the P2 terminal of the LAN-side output optical switch. The main control unit controls the switching matrix to switch between the following modes based on the equipment operating status: Normal Service / Transparent Relay Mode: When the device itself is functioning normally, all four optical switches are connected at P1 and P3. At this time, WAN and LAN network traffic enters the local encryption module for processing; simultaneously, the cascade input and cascade output channels form a physical pass-through. If upstream fault traffic enters the cascade interface, it will be directly transmitted through the P1-P3 path of each cascade optical switch. Fault Source Bypass Mode: When a device fault or power failure is detected, all four optical switches switch to the P1 and P2 terminals in the on state using mechanical latching characteristics or master control commands. This establishes a physical bypass from the WAN port to the cascaded output WAN port, and a physical backhaul bypass from the cascaded output LAN port (backhaul receiver) to the LAN port.

2. The system according to claim 1, characterized in that, When the optical switch performs the switching action to bypass mode to the fault source, the physical switching delay is less than 10ms; and in the event of a sudden power failure, the optical switch relies on the internal mechanical locking mechanism to maintain the current conduction state of P1 and P2 terminals or P1 and P3 terminals.

3. The system according to claim 1, characterized in that, The triggering condition for the transparent relay mode is as follows: the optical power monitoring module detects the optical power of the cascaded input port in real time and compares it with a preset threshold. If the power exceeds the preset threshold, it directly determines that the upstream device has failed based on the physical layer signal strength and sends bypass traffic. No protocol layer handshake is required, and the pure physical layer forwarding delay in the transparent relay mode is less than 1μs.

4. A cluster hot standby switching method based on the system described in any one of claims 1 to 3, characterized in that, Includes the following steps: 1) Connect N master devices to the cascade ports of the common backup unit sequentially via fiber optic links to form a daisy-chain physical topology; 2) The main control unit of each device monitors its own operating status in real time, and the optical power monitoring module monitors the physical optical power signal of the cascaded input interface in real time; 3) When a main device loses power or malfunctions, the fault source bypass mode is triggered. The optical switch matrix inside the device instantly reconstructs the physical optical path by relying on the mechanical locking mechanism, and physically redirects the WAN-side traffic that should have entered the local machine to the cascaded output WAN port. 4) When the downstream intermediate main equipment detects an optical signal entering the cascaded input WAN port, it triggers or maintains the transparent relay mode. The fault traffic directly passes through the P1-P2 transparent transmission channel of the cascaded side optical switch inside the intermediate main equipment, bypassing the CPU and encryption module of the intermediate node, realizing physical layer offloading with zero computing power consumption and transparent transmission downstream. 5) The common backup machine is located at the end of the link. It receives and uses its own encryption module to process the fault traffic transmitted through the transparent transmission. The processed legitimate traffic is transmitted back through the cascaded LAN channels of each intermediate node in the same transparent transmission method, and finally outputs to the internal network from the LAN port of the fault source device.

Images