A method, system, and related equipment for verifying optical modules based on host behavior models.

By adopting a verification method based on host behavior models, the interaction process between the CMIS host and the optical module is recorded and analyzed. Controlled perturbation variables are applied to generate a host behavior model, which solves the problems of static nature and insufficient process details in existing optical module verification methods, and achieves more efficient compatibility verification.

CN123122648APending Publication Date: 2026-11-17SHENZHEN HUANGUANG ERA TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611546502.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-10-08
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing optical module verification methods cannot delve into the detailed interaction process between the host and the module, leading to compatibility issues such as slow initialization and abnormal states under different host or software versions. Existing technologies lack the ability to explore dynamic process details and boundaries.

Method used

By using a verification method based on host behavior models, the interaction process between the CMIS host and the reference optical module is recorded, parsed into multiple CMIS semantic events, controlled perturbation variables are applied, host behavior is observed, a reusable host behavior model is generated, and compared with the optical module under test to output compatibility verification results.

Benefits of technology

Dynamic modeling and quantification of host tolerance boundaries improve the efficiency of optical module verification, enabling more accurate identification of compatibility issues and improvements to the verification process.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123122648A_ABST
    Figure CN123122648A_ABST
Patent Text Reader

Abstract

This invention provides a method, system, and related equipment for optical module verification based on a host behavior model. The method includes: first, collecting a baseline transaction flow through an interaction proxy during normal interaction between the CMIS host and a reference optical module, and parsing it into a semantic event sequence according to the CMIS protocol; then, implementing controlled perturbations of a single variable on the host in each round within the event window of the selected target CMIS semantic event; next, extracting the boundary parameters corresponding to the perturbation variables based on the host's behavior classification after perturbation; then, integrating the event information, perturbation information, boundary information, and host behavior branches from multiple rounds of testing to generate a reusable host behavior model; finally, measuring the actual response parameters of the optical module under test using the behavior model and comparing them with the corresponding boundaries to output the compatibility verification result. This invention transforms the static comparison verification process of existing technologies into dynamic behavior modeling and quantifies the host's tolerance boundaries, thereby improving the efficiency of optical module verification.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of optical communication device testing technology, and in particular to an optical module verification method, system and device based on host behavior model. Background Technology

[0002] As 400G, 800G and higher speed optical modules enter data centers and AI computing clusters, optical modules not only need to meet the optoelectronic performance requirements, but their management interfaces also need to work with the host software to complete identification, configuration, initialization and monitoring. The CMIS specification provides basic definitions for registers, states and configuration processes, but different host manufacturers do not implement them exactly the same in software. In actual debugging, we often encounter situations where the module fields seem to conform to the specification and can pass regular tests, but after switching to a certain host or software version, the module initialization is slow, the state polling is abnormal, or the data path state machine configuration process does not reach the expected state.

[0003] The current conventional approach is to establish a standard library of optical module parameters. During testing, the module's rate, status, and register parameters are read and compared with standard values ​​or ranges. This method is suitable for determining whether a certain parameter of the module falls within the range. Although it can reduce the workload of manually reading registers, it is a static and isolated parameter comparison. It does not record what the host reads first during actual access, how long it waits before reading again, how long it waits after page switching, or whether it retries after configuration writing. It also cannot explain at which handshake interaction stage a certain module's anomaly occurs on a certain host.

[0004] Some existing technologies involve creating compatibility lists, which involves testing certain host and module combinations with certain manufacturers in advance to determine compatibility or incompatibility. Subsequent searches can be made by model or manufacturer. This method is effective for combinations that have already been tested, but when new host software, new port configurations, or new module firmware are released, it is still necessary to re-test the combinations. More importantly, compatibility lists usually only save the final conclusions and do not save the specific behavior of the host during page access, application configuration, data path initialization, etc., that is, they lack detailed interaction processes and specific reasons.

[0005] During R&D and debugging, engineers sometimes capture I2C / CMIS communication logs between the host and the module for analysis. This method can show which read and write operations the host performed. However, a single packet capture is merely a fixed trajectory of the host under specific conditions. A fixed trajectory cannot indicate how long the host can tolerate latency, how many temporary NACKs it can endure, how long Clock Stretch will trigger a retry, or how long a state machine transition intermediate state will be considered abnormal. Moreover, it is limited to the equipment models available in the laboratory. If the log is simply replayed to the module under test, it is just repeating a fixed path and still cannot reveal where the critical point of CMIS host compatibility or incompatibility lies.

[0006] In summary, existing technologies are either too static, lack process details, or are unable to explore boundaries, making it difficult to fundamentally solve the deep-seated compatibility problems of CMIS. Therefore, there is an urgent need for a new optical module verification and testing method. Summary of the Invention

[0007] This invention provides a method, system, and device for verifying optical modules based on a host behavior model. The main objective of this invention is to solve the technical problems mentioned in the background section of the prior art.

[0008] The first aspect of this invention provides a method for verifying optical modules based on a host behavior model, comprising: When the CMIS host and the reference optical module interact normally, the baseline transaction process is recorded through the interaction agent device, and the baseline transaction process is parsed and divided into multiple CMIS semantic events according to the CMIS management protocol to obtain the baseline event sequence. Select a target CMIS semantic event. In the event window corresponding to the target CMIS semantic event, change only one perturbation variable in each round of testing and apply a controlled perturbation to the CMIS host through the interaction agent device. Observe and record the subsequent interactive behavior of the CMIS host after being disturbed, and extract the direct acceptance boundary, recoverable acceptance boundary and failure boundary corresponding to the disturbance variable according to the behavior classification. The event information, disturbance information, boundary information and host behavior branches obtained from multiple rounds of testing are associated and integrated according to the target CMIS semantic events and the corresponding disturbance variables to generate a reusable host behavior model. Load the host behavior model, measure the actual response parameters of the optical module under test corresponding to the perturbation variable under the target CMIS semantic event, compare the actual response parameters with the boundary corresponding to the target CMIS semantic event and the perturbation variable in the host behavior model, and output the compatibility verification result.

[0009] In an optional embodiment of the first aspect of the present invention, the step of recording the baseline transaction process through the interaction proxy device and parsing and dividing the baseline transaction process into multiple CMIS semantic events according to the CMIS management protocol during normal interaction between the CMIS host and the reference optical module includes: Based on the page field, register field, read / write direction, adjacent transaction time interval, and subsequent transaction type of the underlying I2C transactions recorded by the interactive proxy device, consecutive underlying I2C transactions are merged and identified as CMIS semantic events with protocol significance. The CMIS semantic events include: module information reading event, module status detection event, first reading event after page switching, application capability reading event, module status control event, application selection and configuration event, data path initialization event, data path status polling event, and status confirmation event; For each identified CMIS semantic event, generate event data including event ID, start transaction, end transaction, related field set, and event start and end time.

[0010] In an optional embodiment of the first aspect of the present invention, the selected target CMIS semantic event, within the event window corresponding to the target CMIS semantic event, involves changing only one perturbation variable in each round of testing, and applying a controlled perturbation to the CMIS host through the interactive proxy device, including: Determine the start and end transactions of the target CMIS semantic event to define the event window; Within the event window, the perturbation variables are applied to the CMIS host through the interaction agent device. The perturbation variables include one of the following: clock delay duration, number of temporary negative acknowledgments, page data preparation delay time, configuration status update time, module state machine intermediate state holding time, and data path state machine intermediate state holding time.

[0011] In an optional embodiment of the first aspect of the present invention, for the perturbation variable whose larger value is more unfavorable, multiple rounds of testing are performed using a stepwise expansion or binary search method, wherein each round of testing includes: Clear the cache and previous disturbance state of the interactive proxy device; Restore the same port configuration of the CMIS host, and restore the page selection state, application configuration state, and data path initial state of the reference optical module to the same starting point; Execute the baseline transaction process once without disturbance. If the baseline transaction process fails, the test data for this round is deemed invalid and will not be included in subsequent boundary calculations.

[0012] In an optional embodiment of the first aspect of the present invention, observing and recording the subsequent interactive behavior of the CMIS host after being disturbed, and extracting the direct acceptance boundary, recoverable acceptance boundary, and failure boundary corresponding to the disturbance variable according to the behavior classification includes: After each round of perturbation, if the CMIS host directly continues to execute the next event in the baseline event sequence, the current perturbation value is recorded as acceptance to determine the direct acceptance boundary; After each round of perturbation, if the CMIS host retryes the current transaction a limited number of times before continuing to execute the next event in the baseline event sequence, or re-executes the current configuration-related transaction before continuing to execute the next event in the baseline event sequence, then the current perturbation value is recorded as recoverable and acceptable, so as to determine the recoverable and acceptable boundary. After each round of perturbation, if the CMIS host continues to repeatedly read the current transaction or waits for more than a preset observation time, or stops subsequent CMIS access, the current perturbation value is recorded as unacceptable to determine the failure boundary.

[0013] In an optional embodiment of the first aspect of the present invention, the data structure information of the host behavior model includes basic information, event information, disturbance information, boundary information, and behavior branches: the basic information includes host model, software version, and CMIS version; the event information includes event ID, triggering condition, event window, and related field set; the disturbance information includes disturbance type, starting value, stepping method, and stopping condition; the boundary information includes the direct acceptance boundary, the recoverable acceptance boundary, and the failure boundary; and the behavior branches include host feedback features such as continue, retry, reconfigure, timeout, or stop access.

[0014] In an optional embodiment of the first aspect of the present invention, loading the host behavior model, measuring the actual response parameters of the optical module under test corresponding to the perturbation variable under the target CMIS semantic event, comparing the actual response parameters with the boundary corresponding to the target CMIS semantic event and the perturbation variable in the host behavior model, and outputting the compatibility verification result includes: The optical module under test is driven to execute the target CMIS semantic event specified in the host behavior model, the actual response parameters of the optical module under test are collected, and the response parameters are normalized according to the compatibility degradation direction. If the actual response parameter is less than or equal to the direct acceptance boundary, then the optical module under test is determined to be compatible with the target CMIS semantic event; If the direct acceptance boundary < the actual response parameter ≤ the recoverable acceptance boundary, then the optical module under test is determined to be recoverable compatible or low-risk in the target CMIS semantic event; If the recoverable acceptable boundary is less than the actual response parameter and less than the failure boundary, then it is determined to be a critical risk; If the actual response parameter is greater than or equal to the failure boundary, it is determined to be incompatible. When the incompatibility is determined, the failure event ID, the actual response parameter, the corresponding tolerance boundary, and the subsequent behavior of the host are output, and the retesting items are suggested.

[0015] A second aspect of the present invention provides an optical module verification system based on a host behavior model, the optical module verification system comprising: The transaction acquisition and identification module is used to record the baseline transaction process through the interaction agent device when the CMIS host and the reference optical module are interacting normally, and to parse the baseline transaction process according to the CMIS management protocol and divide it into multiple CMIS semantic events to obtain the baseline event sequence. The single-variable perturbation test module is used to select a target CMIS semantic event. In each round of testing, only one perturbation variable is changed within the event window corresponding to the target CMIS semantic event. The controlled perturbation is applied to the CMIS host through the interactive agent device. The host tolerance boundary search module is used to observe and record the subsequent interaction behavior of the CMIS host after being disturbed, and extract the direct acceptance boundary, recoverable acceptance boundary and failure boundary corresponding to the disturbance variable according to the behavior classification. The host behavior model generation module is used to integrate the event information, disturbance information, boundary information and host behavior branches obtained from multiple rounds of testing, according to the target CMIS semantic event and the corresponding disturbance variable, to generate a reusable host behavior model. The compatibility comparison and verification module is used to load the host behavior model, measure the actual response parameters of the optical module under test under the corresponding target CMIS semantic event and the perturbation variable, compare the actual response parameters with the boundary corresponding to the target CMIS semantic event and the perturbation variable in the host behavior model, and output the compatibility verification result.

[0016] A third aspect of the present invention provides an optical module verification device based on a host behavior model, the optical module verification device based on a host behavior model includes: a memory and at least one processor, the memory storing instructions, and the memory and the at least one processor being interconnected via a line; The at least one processor invokes the instructions in the memory to cause the host behavior model-based optical module verification device to perform the host behavior model-based optical module verification method as described in any one of the first aspects of the present invention.

[0017] A fourth aspect of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the optical module verification method based on a host behavior model as described in any one of the first aspects of the present invention.

[0018] Beneficial Effects: This invention provides a method, system, and related equipment for optical module verification based on a host behavior model. The method includes: first, collecting the baseline transaction flow through an interaction proxy during normal interaction between the CMIS host and the reference optical module, and parsing it into a semantic event sequence according to the CMIS protocol; then, implementing controlled perturbations of a single variable on the host in each round within the event window of the selected target CMIS semantic event; next, extracting the boundary parameters corresponding to the perturbation variables based on the host's behavior classification after perturbation; then, integrating the event information, perturbation information, boundary information, and host behavior branches from multiple rounds of testing to generate a reusable host behavior model; finally, measuring the actual response parameters of the optical module under test using the behavior model and comparing them with the corresponding boundaries to output the compatibility verification results. This invention transforms the static comparison verification process of existing technologies into dynamic behavior modeling and quantifies the host's tolerance boundaries, thereby improving the efficiency of optical module verification. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of an embodiment of the hardware architecture of an optical module verification system according to the present invention; Figure 2 This is a schematic diagram of an embodiment of the internal hardware of a CMIS interactive agent device, which is an example of the present invention. Figure 3 This is a schematic diagram of an embodiment of the system signal connection example based on the QSFP-DD package of the present invention; Figure 4 This is a schematic diagram of an embodiment of the system signal connection example based on OSFP packaging, which is an exemplary embodiment of the present invention; Figure 5 This is a schematic diagram of an embodiment of the optical module verification method based on a host behavior model, which is an exemplary embodiment of the present invention. Figure 6 This is a schematic diagram of an embodiment of the benchmark transaction acquisition and event identification process of the present invention; Figure 7 This is a schematic diagram of an embodiment of an event window and a single-variable perturbation mechanism of the present invention; Figure 8 This is a schematic diagram of an embodiment of the perturbation mechanism and physical implementation process of the present invention; Figure 9 This is a schematic diagram of an embodiment of the baseline recovery process before each round of boundary testing, which is an exemplary embodiment of the present invention; Figure 10 This is a schematic diagram of an embodiment of the host behavior classification and boundary search process of the present invention; Figure 11 This is a schematic diagram of an embodiment of the verification process of a module under test based on a host behavior model, which is an example of the present invention. Figure 12 This is a schematic diagram of an embodiment of a compatible anomaly localization process according to the present invention; Figure 13 This is a schematic diagram of an embodiment of the software architecture of an optical module verification system based on a host behavior model, which is an example of the present invention. Figure 14 This is a schematic diagram of an embodiment of an optical module verification device based on a host behavior model, which is an example of the present invention. Detailed Implementation

[0020] The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0021] The first aspect of this invention provides a method for verifying optical modules based on a host behavior model. The general approach involves first allowing a CMIS host to interact normally with a standard-compatible reference module once. An interaction proxy device records and transparently forwards the interaction to obtain a baseline transaction flow. Then, the baseline transaction is categorized according to CMIS meaning into semantic events such as page access, state detection, application configuration, and data path initialization. Next, an event is selected, and only one variable is changed in each round within the event window (e.g., I2C response delay during power-on, number of temporary NACKs, clock stretch time, or duration of various stable and transitional states). Then, based on the classification of subsequent interaction behaviors of the CMIS host, the acceptable and failure boundaries of the CMIS host are identified. Host behavior modeling is then performed based on the acquired event / disturbance / boundary / behavior information. Finally, the optical module's compatibility is verified using the established behavior model. This invention solves the problem that existing verification methods are too static and cannot detect the host's tolerance threshold.

[0022] See Figure 1The hardware architecture of the test system upon which the verification method of this invention relies mainly includes a host device 101 (i.e., a CMIS host), a CMIS interaction agent device 100 (i.e., an interaction agent device), a reference module or module under test 200 (i.e., a reference or optical module under test), a host computer analysis system 300, a host behavior model library 400, and a test configuration and result output module 500. The CMIS interaction agent device 100 is placed between the host device 101 and the reference or optical module under test 200, and can completely monitor the handshake process on the I2C management bus between the CMIS host and the optical module. The I2C management bus between the CMIS host and the optical module is not directly short-circuited, but the interaction agent device handles host-side transactions and module-side transactions respectively.

[0023] In this invention, the host computer analysis system 300 does not communicate with the interactive agent device through the optical module management bus, but connects to the interactive agent device through the USB HID independent control communication interface. The host computer analysis system is responsible for issuing test configurations, reading logs, identifying events, generating models, and outputting verification results. The host behavior model library 400 is used to store behavior model data corresponding to different hosts, different software versions, or different port configurations. The test configuration and result output module 500 is used to store the test items, boundary parameters, and risk judgment conditions required for subsequent verification of the module under test.

[0024] For example, the internal structure of a CMIS interactive agent device according to the present invention can be as follows: Figure 2 As shown, the CMIS interactive agent device includes a host-side CMIS interface 110, a module-side CMIS interface 120, a buffer and forwarding unit 130, a disturbance control unit 140, a unified time base unit 150, a low-speed signal acquisition and driving unit 160, and a control communication interface 170. In this invention, the host-side CMIS interface and the module-side CMIS interface are connected via the buffer and forwarding unit. This configuration allows the interactive agent device to buffer, delay, temporarily NACK, or maintain the status of accesses issued by the CMIS host without bus conflicts, thereby completing the CMIS host behavior analysis and compatibility testing of the optical module under test. In this invention, the unified time base unit is mainly used to provide a unified time base for host-side transactions, module-side transactions, variable disturbance actions, and low-speed control signal changes. The low-speed signal acquisition and driving unit is mainly used to acquire or forward low-speed management signals of the module as needed for testing. The control communication interface is mainly used to communicate with the host computer to implement configuration, log, and model data distribution.

[0025] In this invention, based on the different packaging forms of the optical module, there are two connection methods between the optical module and the host interface. When the system uses a QSFP-DD packaged optical module, the hardware connection diagram of the optical module, the interaction agent, and the host interface can be as follows: Figure 3As shown, in this connection method, the host side and the module side respectively bring out signals such as SCL, SDA, ModSelL, ResetL, LPMode, ModPrsL, IntL, and 3.3V / GND. SCL and SDA are connected to the host side target interface and the module side controller interface respectively, instead of simply passing through and shorting the two sides directly. In the QSFP-DD hardware example shown, ModSelL, ResetL, and LPMode are acquired by the low-speed signal acquisition and drive unit, and can be transparently forwarded or delayed according to the test configuration as needed. IntL and ModPrsL are mainly used as status inputs for monitoring and recording. ModPrsL is only used as the QSFP-DD module's in-situ status acquisition object in the system, and is not used as a changeable measurement quantity.

[0026] When the system uses an OSFP-packaged optical module, the hardware connection diagram of the optical module, interaction agent, and host interface can be as follows: Figure 4 As shown, in this connection method, the low-speed pin signals mainly include SCL, SDA, LPWn / PRSn, and INT / RSTn. SCL and SDA can be connected through two independent open-drain bus interfaces. For OSFP package pins LPWn / PRSn and INT / RSTn, they are multi-level dual-function signals. Then, the corresponding voltage comparator, 1.25V and 2.5V reference voltages, controllable resistor network and open-drain drive circuit are added inside the corresponding low-speed signal acquisition and drive unit according to the OSFP interface requirements. This enables the module to achieve in-situ and low-power functions, as well as voltage parsing of reset and alarm interrupt signals.

[0027] See Figure 5 The optical module verification method based on the host behavior model includes: S100. When the CMIS host and the reference optical module are interacting normally, the baseline transaction process is recorded through the interaction agent device, and the baseline transaction process is parsed and divided into multiple CMIS semantic events according to the CMIS management protocol to obtain the baseline event sequence.

[0028] In this invention, for each CMIS host, before model building, a round of baseline behavior collection of the CMIS host is performed (i.e., a round of baseline transaction process is executed). During the baseline collection phase, no disturbance is applied (i.e., normal interaction), only transparent forwarding is performed. The interaction proxy device (i.e., the CMIS interaction proxy device) records events such as address, page, register, read / write direction, data content, ACK / NACK, Clock Stretch, low-power / Reset pin actions, and timestamps during normal access between the CMIS host and the reference optical module. For example, an exemplary baseline transaction collection and event recognition process of this invention can be as follows: Figure 6 As shown.

[0029] In this invention, a complete benchmark acquisition is segmented into one port initialization and one configuration process. The acquisition window is set to a time of 2 to 20 seconds. Generally, the benchmark acquisition ends when all Datapaths of the module enter the DPActivated state. When it is necessary to confirm the host link status, the acquisition window can be further extended to the corresponding port of the host with normal linkup. The specific length is adjusted according to the target host process.

[0030] This step in the invention is not simply to save a log, but to determine the access path of the host under normal circumstances. For example, the host first reads basic information, monitors the changes in the Lowpower / Reset pin status, then polls ModuleState, then reads the Application Descriptor, then selects AppSelCode to write to the application selection and control field, and finally polls the Data Path status. This process is necessary because only with this normal path in place can we know in which event window to apply the disturbance and whether the host deviates from the normal process after the disturbance.

[0031] In an optional embodiment of the first aspect of the present invention, when the CMIS host and the reference optical module interact normally, recording the baseline transaction process through the interaction proxy device and parsing and dividing the baseline transaction process into multiple CMIS semantic events according to the CMIS management protocol includes: based on the page field, register field, read / write direction, adjacent transaction time interval, and subsequent transaction type of the underlying I2C transaction recorded by the interaction proxy device, merging and identifying the continuous underlying I2C transactions into CMIS semantic events with protocol significance; the CMIS semantic events include: module information reading event, module status detection event, first reading event after page switching, application capability reading event, module status control event, application selection and configuration event, data path initialization event, data path status polling event, and status confirmation event; and generating event data including event ID, start transaction, end transaction, related field set, and event start and end time for each identified CMIS semantic event.

[0032] Specifically, typical I2C events are merely low-level read / write records. However, in this invention, these events are further categorized into semantic events (i.e., CMIS semantic events) according to the CMIS management process. Semantic events include module information read events, module status detection events, first read events after page switching, application capability read events, module status control events (controlled by Lowpower / Reset pins or corresponding register bits), application selection and configuration events, data path initialization events, data path status polling events, and status confirmation events. In this invention, during event parsing, the system simultaneously examines page fields, register fields, and read / write records. The direction, the time interval between adjacent transactions, and the type of subsequent transactions are considered. For example, if the host continuously reads certain status fields until the status is satisfied before accessing the application configuration field, this segment can be identified as a module status detection event. Or, if the host writes the application selection field and then immediately reads the configuration status and data path status, this can be identified as an application configuration event and a data path initialization event. Each event is recorded with an event ID, start transaction, end transaction, related field set, event start time, event end time, and subsequent event ID. Thus, a complete CMIS access process is represented as an event sequence {E1, E2, ..., En}.

[0033] S200. Select a target CMIS semantic event. Within the event window corresponding to the target CMIS semantic event, change only one perturbation variable in each round of testing, and apply a controlled perturbation to the CMIS host through the interaction proxy device. In this invention, this step may include: determining the start and end transactions of the target CMIS semantic event to define the event window; within the event window, applying the perturbation variable to the CMIS host through the interaction proxy device. The perturbation variable includes one of the following: clock delay duration, number of temporary negative acknowledgments, page data preparation delay time, configuration state update time, module state machine intermediate state holding time, and data path state machine intermediate state holding time.

[0034] Specifically, in this invention, the relationship between the event window and the univariate perturbation can be as follows: Figure 7 As shown, the event window is the time range or transaction range (including the start and end of the transaction) for applying a disturbance and observing the host's subsequent behavior. For example: 1) The window for the first read event after page switching, starting from when the host writes the page selection field, and ending when the host completes the first read of the target page or enters another event; 2) The module status control event, starting from when the host triggers the Lowpower / Reset pin or the corresponding register bit control, until the module jumps to the correct state normally; 3) The window for the data path initialization event, starting from when the initialization trigger transaction, and ending when the host confirms the target state or stops polling.

[0035] In this invention, within an event window, each round of testing changes only one variable while keeping the other conditions unchanged. The perturbation variable includes one of the following: Clock Stretch duration, number or location of temporary NACK occurrences, page data preparation delay, configuration state update time, intermediate state holding time of module state machine, and intermediate state holding time of data path state machine. By using single-variable control, it is possible to avoid the inability to determine which factor caused the host behavior change after changing multiple conditions simultaneously.

[0036] In this invention, the physical implementation process of various disturbances can be as follows: Figure 8 As shown, in summary, the disturbances of this invention mainly include the following types of actions: The first type is transaction-level Clock Stretch, which means that the host-side target interface keeps SCL low during the protocol-allowed phase, thus delaying the transmission of the current byte; the second type is temporary NACK, which means that the host-side target interface does not pull SDA low in the address acknowledgment bit or data acknowledgment bit, and observes whether the host retryes; the third type is semantic state delay (page data delay feedback and legal intermediate state lengthening, etc.), which means that the host-side transaction is still completed normally, but the proxy device returns the original state value, the original page data, or a legal intermediate state conforming to the CMIS protocol within a certain period of time.

[0037] In this invention, for example, the Clock Stretch test can start from 10us and extend to a maximum of 25ms in fixed 10us steps; the temporary NACK test starts from one NACK and increases by 5 each time until the host stops continuing the current event. The purpose of this setting is to first find a coarse range and then perform refined confirmation near the boundary; as for the status field, the agent device can generate some status values ​​that exist in the CMIS specification, such as the transition states of the module state machine and data path state machine, as well as the feedback states of the application configuration. It mainly selects from the feedback status values ​​or legal transition intermediate states allowed by the target CMIS version and extends the holding time. This invention does this to ensure that the test actions are still within the scope of the CMIS management protocol.

[0038] Furthermore, in this invention, since the tolerance boundary needs to be obtained through multiple rounds of testing, in order to make each round of testing comparable, the baseline must be restored to the same starting point before each round of perturbation. An exemplary baseline restoration process before each round of boundary testing in this invention is as follows: Figure 9As shown, the recovery process includes clearing the proxy device cache and the previous round of disturbance state, restoring the same host port configuration, restoring the module page selection state, application configuration state, and data path initial state, confirming that the voltage, temperature, and interface rate are within the preset range, and executing a non-disturbance benchmark process. If the non-disturbance benchmark test itself fails, the test data of this round will not participate in the boundary calculation (i.e., in an optional embodiment of the first aspect of the present invention, before each round of testing, the following is included: clearing the cache of the interactive proxy device and the previous round of disturbance state; restoring the same port configuration of the CMIS host, and restoring the page selection state, application configuration state, and data path initial state of the reference optical module to the same starting point; executing a non-disturbance benchmark transaction process, and if the benchmark transaction process fails, the test data of this round is determined to be invalid and will not participate in the subsequent boundary calculation).

[0039] S300. Observe and record the subsequent interaction behavior of the CMIS host after being disturbed, and extract the direct acceptance boundary, recoverable acceptance boundary, and failure boundary corresponding to the disturbance variable according to the behavior classification. In this invention, this step may include: after each round of disturbance, if the CMIS host directly continues to execute the next event in the baseline event sequence, the current disturbance value is recorded as accepted to determine the direct acceptance boundary; after each round of disturbance, if the CMIS host retries the current transaction a limited number of times before continuing to execute the next event in the baseline event sequence, or re-executes the current configuration-related transaction before continuing to execute the next event in the baseline event sequence, the current disturbance value is recorded as recoverable acceptance to determine the recoverable acceptance boundary; after each round of disturbance, if the CMIS host continuously reads the current transaction repeatedly or waits for more than a preset observation time, or stops subsequent CMIS access, the current disturbance value is recorded as unacceptable to determine the failure boundary.

[0040] Specifically, in this invention, after each round of perturbation, the system continues to observe and record the host's subsequent behavior: 1) The host directly enters the next event; 2) The host continues after a limited number of retries; 3) The host rewrites the configuration; 4) The host keeps reading the same state until the timeout specified by the CMIS protocol; 5) The host stops all subsequent access or puts the module into Lowpower state.

[0041] Taking Table 1 below as an example, this invention categorizes subsequent host behaviors into five types based on the host's typical subsequent behavior: Type A is typically denoted as directly accepted; Types B and C are denoted as recoverable and acceptable according to test requirements; Types D and E are denoted as unacceptable. For example, the host behavior classification and tolerance boundary search algorithm process of this invention can be described as follows: Figure 10As shown, in this invention, before testing the disturbance variable P, the system first restores the baseline and confirms that the disturbance-free process has passed, and then sets a disturbance value P for testing. If the host behavior belongs to class A, the current P is recorded as accepted; if it belongs to the allowed class B or C, the current P is recorded as recoverable and acceptable; if the host behavior belongs to class D or E, the current P is recorded as unacceptable and fails.

[0042] Table 1. Exemplary Classification Table of Host Subsequent Behaviors in this Invention

[0043] In this invention, for variables such as response latency, number of NACKs, clock stretch time, and legal intermediate state holding time, which are "more unfavorable the larger the value", the system will use a step-by-step expansion plus binary search or step-by-step approximation method to conduct multiple rounds of testing (that is, in an optional embodiment of the first aspect of this invention, for the disturbance variables that are more unfavorable the larger the value, a step-by-step expansion or binary search method is used to conduct multiple rounds of testing). For variables with unclear monotonicity, an interval scan will be performed first. If multiple acceptable intervals are found, the system will save them separately instead of forcibly merging them into a single boundary.

[0044] Specifically, in this invention, the system will ultimately save the directly received boundary P. direct Recoverable Acceptable Boundary P recover and failure boundary P fail For variables whose disturbance increases with increasing value, P direct This indicates that, after repeated testing and confirmation, the host still maintains the maximum perturbation value for directly accepting behavior, P. recover P represents the maximum disturbance value that allows the host to continue the subsequent process through a limited number of retries or reconfigurations. fail This indicates that after repeated testing, unacceptable disturbance values ​​such as continuous waiting and access stoppage began to appear. In order to ensure accuracy, multiple rounds of testing will be carried out near the boundary, and different host ports will be used for retesting. The number of successful tests, the number of failed tests, behavior branches, and result stability will be recorded. If the baseline recovery fails, non-target variables change, or host behavior cannot be classified during the test, the data of that round will be marked as invalid, and the device and module need to be restarted for retesting.

[0045] In this invention, since application configuration is a process prone to compatibility differences, when performing benchmark data collection, the invention first identifies the transaction relationship between the host reading application description information, writing application selection or control parameters, and then reading configuration status fields. During the perturbation phase, the configuration status retention time, configuration status update time, or response delay of configuration-related fields are selected as optional variables. In each round of testing, other transaction contents are kept unchanged, only one variable is changed, and it is observed whether the host continues to read the status, whether it retryes the configuration, whether it enters the data path initialization phase, and whether the final status is successfully configured. After multiple rounds of perturbation, the host's waiting boundary and recoverable branch for the application configuration event are obtained.

[0046] S400. The event information, disturbance information, boundary information, and host behavior branches obtained from multiple rounds of testing are associated and integrated according to the target CMIS semantic events and the corresponding disturbance variables to generate a reusable host behavior model. In this invention, the data structure information of the host behavior model includes basic information, event information, disturbance information, boundary information, and behavior branches: the basic information includes the host model, software version, and CMIS version; the event information includes the event ID, triggering condition, event window, and related field set; the disturbance information includes the disturbance type, starting value, stepping method, and stopping condition; the boundary information includes the direct acceptance boundary, the recoverable acceptance boundary, and the failure boundary; the behavior branches include host feedback features such as continue, retry, reconfigure, timeout, or stop access.

[0047] Specifically, in this invention, the testing system will use the basic information (host / software / CMIS version), event information (event ID / window / field set), disturbance information (disturbance type / step / stopping condition), and boundary information (P) obtained from multiple rounds of testing. direct / P recover / P fail The host behavior model is comprehensively stored together with host behavior branches (continue / retry / timeout / stop), and the data structure table of an exemplary host behavior model of the present invention is shown in Table 2 below.

[0048] Table 2. Data structure table of an exemplary host behavior model of the present invention.

[0049] S500: Load the host behavior model, measure the actual response parameters of the optical module under test corresponding to the perturbation variable under the target CMIS semantic event, compare the actual response parameters with the boundary corresponding to the target CMIS semantic event and the perturbation variable in the host behavior model, and output the compatibility verification result. In this invention, the step includes: driving the optical module under test to execute the target CMIS semantic event specified in the host behavior model, collecting the actual response parameters of the optical module under test, and normalizing the response parameters according to the compatibility degradation direction; if the actual response parameters ≤ the direct acceptance boundary, the optical module under test is determined to be compatible with the target CMIS semantic event; if the direct acceptance boundary < the actual response parameters ≤ the recoverable acceptance boundary, the optical module under test is determined to be recoverably compatible or low-risk with respect to the target CMIS semantic event; if the recoverable acceptance boundary < the actual response parameters < the failure boundary, it is determined to be critically risky; if the actual response parameters ≥ the failure boundary, it is determined to be incompatible, and when it is determined to be incompatible, the failure event ID, the actual response parameters, the corresponding tolerance boundary, and the host's subsequent behavior are output, and a retest item is suggested.

[0050] Specifically, such as Figure 11 As shown, in this invention, during the verification testing phase of the optical module under test, the system loads the target host behavior model, generates event-level test items based on the event ID and boundary parameters in the model, and then drives the module under test to execute the same CMIS semantic events, collecting the actual response parameters R of the optical module under test, such as page data preparation time, NACK count, Clock Stretch time, status field update time, or legal intermediate state holding time. Before comparison, the system first normalizes the response parameters according to the direction of compatibility degradation. For example, an increase in the delay time value indicates a more severe disturbance. For parameters whose larger normalized values ​​are more unfavorable, if R is not greater than P... direct Determine if the event is compatible; if R is greater than P direct But not greater than P recover If R is greater than P, it is determined to be recoverable, compatible, or low-risk; recover But less than P fail If R is not less than P, it is considered a critical risk; fail It was determined to be incompatible; furthermore, such as Figure 12 As shown, when the optical module under test fails to verify, the system will also determine the failure event ID, and then classify the exception into the stage of page access, application configuration, data path initialization or status confirmation according to the event type. The system will output results including detailed parameters such as failure event ID, actual response parameters, corresponding acceptance boundary, failure boundary, and subsequent host behavior, and suggest retesting the project.

[0051] In this invention, for some special CMIS semantic events, such as the first read event after a page switch, taking the first read after a page switch as a typical CMIS event window as an example, the system will first collect a normal baseline to find the transaction sequence of the first read of the target page field after the host writes the Page Select field. Then, only within this transaction segment will the target page data preparation delay be increased, or the number of temporary NACKs be increased, while other access conditions remain unchanged. For example, when accessing Page 01h, the event window starts from the end of the Page Select write and ends at the end of the first read transaction of the target page. When testing the response delay, the host-side target interface can maintain the SCL low level during the protocol-allowed phase to generate a clock. Stretch, or NACK generated during the response phase when temporarily denying access during testing, can have an initial delay step of 100µs, which is subsequently increased based on the host's retry or cessation of access. If the host still reads the target page and enters subsequent events after the delay, the perturbation value is recorded as acceptance; if the host continues after a limited number of retries, it is recorded as recoverable acceptance; if the host continuously reads and stops access, or shows management failure in the optional port status interface, it is recorded as non-acceptance. When verifying the optical module under test, the test system measures the first readable time or NACK performance after page switching of the module under test and compares it with this boundary to determine whether the module meets the target host's requirements for page access response.

[0052] For example, the data path initialization event involves several actions: triggering, status polling, and status confirmation. The host is sensitive to the waiting time during this stage, making it an important test event in this invention. During benchmark acquisition, the system identifies the period, number of times, and conditions for entering the next event after the host triggers data path initialization. During the disturbance phase, the system changes the status field update time, the legal intermediate transition state holding time, or the status response delay time in rounds. If the host still enters the final status confirmation event according to the benchmark process, it is recorded as accepted. If the host continues to poll until the specified maximum observation time ends or stops subsequent access, it is recorded as unaccepted. During the verification of the module under test, the system measures its data path initialization state transition time and compares it with the boundary in the host behavior model. This allows the system to determine whether the event meets the management process requirements of the target host.

[0053] In summary, the main technical points of the optical module verification method based on host behavior model of this invention include the following: 1. Host behavior modeling based on CMIS semantic events: The underlying CMIS transactions are parsed into event units with protocol meaning, and a host behavior model is established based on the sequence, window, and subsequent branches between events.

[0054] 2. Single-variable controllable perturbation method within the event window: Within specific CMIS semantic event windows such as page switching, application configuration, and data path initialization, only one variable is changed in each round: response delay, number of temporary NACKs, clock stretch time, page data preparation delay, or state retention time.

[0055] 3. Host Behavior Transition Classification and Tolerance Boundary Search Method: Through multiple rounds of testing with different disturbance values, host behavior is classified into types such as continue, continue after limited retries, reconfiguration, continuous waiting or stopping access, and based on this, the direct acceptance boundary, recoverable acceptance boundary and failure boundary are obtained.

[0056] 4. Data structure of the executable host behavior model: The model stores event ID, triggering condition, window definition, disturbance type, boundary parameters, number of retries, failure branch, recovery branch and confidence information, so that the model can be used for subsequent verification of the module under test.

[0057] 5. Event-level verification method for optical modules under test based on host behavior model: Load the host behavior model, measure the actual response parameters of the module under test under the corresponding CMIS event, and output the event-level compatibility result based on the relationship between the parameter and the model boundary.

[0058] Based on the above technical points, compared with existing methods that mainly rely on parameter standard libraries or compatibility lists, the method of the present invention has at least the following technical advantages: (1) Breaking a normal CMIS access process into interpretable semantic events, when an anomaly occurs, it is possible to see which CMIS interaction process the problem occurred in, such as page access, application configuration, data path initialization or status confirmation stage, which is beneficial for compatibility analysis and module improvement, rather than just getting a simple result of overall incompatibility.

[0059] (2) In each round within the same event window, only one variable is changed to obtain the host's direct acceptance boundary, recoverable acceptance boundary, and failure boundary for a certain response condition. This allows us to obtain the host's tolerance range for the event.

[0060] (3) Establish a host behavior model that can be used for reuse testing in subsequent mass production and updated modules. Under the same host or the same software version, different modules under test can be tested using the generated model, reducing the workload of manually analyzing complete logs each time.

[0061] (4) A dual-bus active agent structure and a specific disturbance implementation method are provided. The host side and module side management buses are independent of each other. The agent device generates a controlled response through buffering, time base and disturbance control, avoiding electrical conflicts caused by direct forced driving on the same bus.

[0062] See Figure 13 A second aspect of the present invention provides an optical module verification system based on a host behavior model, the optical module verification system comprising: The transaction acquisition and identification module 10 is used to record the baseline transaction process through the interaction agent device when the CMIS host and the reference optical module are interacting normally, and to parse the baseline transaction process according to the CMIS management protocol and divide it into multiple CMIS semantic events to obtain a baseline event sequence. The single-variable perturbation test module 20 is used to select a target CMIS semantic event. In the event window corresponding to the target CMIS semantic event, only one perturbation variable is changed in each round of testing, and a controlled perturbation is applied to the CMIS host through the interactive agent device. The host tolerance boundary search module 30 is used to observe and record the subsequent interaction behavior of the CMIS host after being disturbed, and extract the direct acceptance boundary, recoverable acceptance boundary and failure boundary corresponding to the disturbance variable according to the behavior classification. The host behavior model generation module 40 is used to integrate the event information, disturbance information, boundary information and host behavior branches obtained from multiple rounds of testing according to the target CMIS semantic event and the corresponding disturbance variable to generate a reusable host behavior model. The compatibility comparison and verification module 50 is used to load the host behavior model, measure the actual response parameters of the optical module under test corresponding to the perturbation variable under the target CMIS semantic event, compare the actual response parameters with the boundary corresponding to the target CMIS semantic event and the perturbation variable in the host behavior model, and output the compatibility verification result.

[0063] In an optional embodiment of the second aspect of the present invention, the transaction acquisition and identification module includes: The event identification unit is used to identify consecutive underlying I2C transactions as CMIS semantic events with protocol significance based on the page field, register field, read / write direction, adjacent transaction time interval, and subsequent transaction type of the underlying I2C transactions recorded by the interactive agent device. The CMIS semantic events include: module information reading event, module status detection event, first reading event after page switching, application capability reading event, module status control event, application selection and configuration event, data path initialization event, data path status polling event, and status confirmation event; The event data generation unit is used to generate event data for each identified CMIS semantic event, including event ID, start transaction, end transaction, related field set, and event start and end time.

[0064] In an optional embodiment of the second aspect of the present invention, the univariate perturbation test module includes: An event window delimitation unit is used to determine the start and end transactions of the target CMIS semantic event to delimit the event window; The disturbance testing unit is used to apply the disturbance variables to the CMIS host through the interaction agent device within the event window. The disturbance variables include one of the following: clock delay duration, number of temporary negative acknowledgments, page data preparation delay time, configuration state update time, module state machine intermediate state holding time, and data path state machine intermediate state holding time.

[0065] In an optional embodiment of the second aspect of the present invention, for the perturbation variable whose larger value is more unfavorable, multiple rounds of testing are performed using a stepwise expansion or binary search method. Before each round of testing, the following steps are included: clearing the cache of the interactive proxy device and the perturbation state of the previous round; restoring the same port configuration of the CMIS host, and restoring the page selection state, application configuration state, and data path initial state of the reference optical module to the same starting point; executing a perturbation-free baseline transaction process. If the baseline transaction process fails, the test data of this round is determined to be invalid and will not participate in subsequent boundary calculations.

[0066] In an optional embodiment of the second aspect of the present invention, the host tolerance boundary search module includes: A direct acceptance boundary determination unit is used to determine the direct acceptance boundary by recording the current perturbation value as accepted after each round of perturbation if the CMIS host directly continues to execute the next event in the reference event sequence. A recoverable acceptance boundary determination unit is used to determine the recoverable acceptance boundary if, after each round of disturbance, the CMIS host retryes the current transaction a limited number of times before continuing to execute the next event in the baseline event sequence, or re-executes the current configuration-related transaction before continuing to execute the next event in the baseline event sequence. The failure boundary determination unit is used to determine the failure boundary by recording the current disturbance value as unacceptable after each round of disturbance if the CMIS host continues to repeatedly read the current transaction or waits for more than a preset observation time, or stops subsequent CMIS access.

[0067] In an optional embodiment of the second aspect of the present invention, the data structure information of the host behavior model includes basic information, event information, disturbance information, boundary information, and behavior branches: the basic information includes host model, software version, and CMIS version; the event information includes event ID, triggering condition, event window, and related field set; the disturbance information includes disturbance type, starting value, stepping method, and stopping condition; the boundary information includes the direct acceptance boundary, the recoverable acceptance boundary, and the failure boundary; and the behavior branches include host feedback features such as continue, retry, reconfigure, timeout, or stop access.

[0068] In an optional embodiment of the second aspect of the present invention, the compatibility comparison verification module includes: The response parameter acquisition unit is used to drive the optical module under test to execute the target CMIS semantic event specified in the host behavior model, collect the actual response parameters of the optical module under test, and normalize the response parameters according to the compatibility degradation direction. The compatibility determination unit is configured to determine that the optical module under test is compatible with the target CMIS semantic event if the actual response parameter is less than or equal to the direct acceptance boundary; if the direct acceptance boundary is less than or equal to the actual response parameter is less than the recoverable acceptance boundary, the optical module under test is recoverably compatible or low-risk with respect to the target CMIS semantic event; if the recoverable acceptance boundary is less than the actual response parameter is less than the failure boundary, the module is determined to be critically risky; if the actual response parameter is greater than or equal to the failure boundary, the module is determined to be incompatible. When the module is determined to be incompatible, the unit outputs the failure event ID, the actual response parameter, the corresponding tolerance boundary, and the host's subsequent behavior, and suggests retesting items.

[0069] Figure 14 This is a schematic diagram of the structure of an optical module verification device based on a host behavior model according to an embodiment of the present invention. This host behavior model-based optical module verification device can vary significantly due to different configurations or performance. It may include one or more processors 60 (central processing units, CPUs) (e.g., one or more processors) and memory 70, and one or more storage media 80 (e.g., one or more mass storage devices) for storing programs or data. The memory and storage media can be temporary or persistent storage. The program stored in the storage media may include one or more modules (not shown in the figure), each module including a series of instruction operations on the host behavior model-based optical module verification device. Furthermore, the processor may be configured to communicate with the storage media and execute the series of instruction operations in the storage media on the host behavior model-based optical module verification device.

[0070] The optical module verification device based on host behavior model in this embodiment of the invention may further include one or more power supplies 90, one or more wired or wireless network interfaces 91, one or more input / output interfaces 92, and / or one or more operating systems, such as Windows Server, Mac OS, Unix, Linux, FreeBSD, etc. Those skilled in the art will understand that... Figure 14 The illustrated optical module verification device structure based on host behavior model does not constitute a limitation on the optical module verification device based on host behavior model. It may include more or fewer components than illustrated, or combine certain components, or have different component arrangements.

[0071] This invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. The computer-readable storage medium stores instructions that, when executed on a computer, cause the computer to perform the steps of the optical module verification method based on the host behavior model.

[0072] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the system, module, or unit described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0073] If the integrated unit is implemented as a software functional unit or a combination of software and hardware and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiments of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a host behavior model-based optical module verification device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0074] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for optical module verification based on host behavior model, characterized in that, include: When the CMIS host and the reference optical module interact normally, the baseline transaction process is recorded through the interaction agent device, and the baseline transaction process is parsed and divided into multiple CMIS semantic events according to the CMIS management protocol to obtain the baseline event sequence. Select a target CMIS semantic event. In the event window corresponding to the target CMIS semantic event, change only one perturbation variable in each round of testing and apply a controlled perturbation to the CMIS host through the interaction agent device. Observe and record the subsequent interactive behavior of the CMIS host after being disturbed, and extract the direct acceptance boundary, recoverable acceptance boundary and failure boundary corresponding to the disturbance variable according to the behavior classification. The event information, disturbance information, boundary information and host behavior branches obtained from multiple rounds of testing are associated and integrated according to the target CMIS semantic events and the corresponding disturbance variables to generate a reusable host behavior model. Load the host behavior model, measure the actual response parameters of the optical module under test corresponding to the perturbation variable under the target CMIS semantic event, compare the actual response parameters with the boundary corresponding to the target CMIS semantic event and the perturbation variable in the host behavior model, and output the compatibility verification result.

2. The method of claim 1, wherein, During normal interaction between the CMIS host and the reference optical module, the baseline transaction process is recorded through the interaction agent device, and the baseline transaction process is parsed and divided into multiple CMIS semantic events according to the CMIS management protocol, including: Based on the page field, register field, read / write direction, adjacent transaction time interval, and subsequent transaction type of the underlying I2C transactions recorded by the interactive proxy device, consecutive underlying I2C transactions are merged and identified as CMIS semantic events with protocol significance. The CMIS semantic events include: module information reading event, module status detection event, first reading event after page switching, application capability reading event, module status control event, application selection and configuration event, data path initialization event, data path status polling event, and status confirmation event; For each identified CMIS semantic event, generate event data including event ID, start transaction, end transaction, related field set, and event start and end time.

3. The method of claim 1, wherein the host behavior model is based on a plurality of host behaviors. The selected target CMIS semantic event, within the event window corresponding to the target CMIS semantic event, involves changing only one perturbation variable in each round of testing. Applying controlled perturbation to the CMIS host through the interactive proxy device includes: Determine the start and end transactions of the target CMIS semantic event to define the event window; Within the event window, the perturbation variables are applied to the CMIS host through the interaction agent device. The perturbation variables include one of the following: clock delay duration, number of temporary negative acknowledgments, page data preparation delay time, configuration status update time, module state machine intermediate state holding time, and data path state machine intermediate state holding time.

4. The method of claim 1, wherein, For the perturbation variable whose larger value is more unfavorable, multiple rounds of testing are conducted using a stepwise expansion or binary search method. Before each round of testing, the following is included: Clear the cache and previous disturbance state of the interactive proxy device; Restore the same port configuration of the CMIS host, and restore the page selection state, application configuration state, and data path initial state of the reference optical module to the same starting point; Execute the baseline transaction process once without disturbance. If the baseline transaction process fails, the test data for this round is deemed invalid and will not be included in subsequent boundary calculations.

5. The method of claim 1, wherein, The observation and recording of the subsequent interactive behavior of the CMIS host after being disturbed, and the extraction of the direct acceptance boundary, recoverable acceptance boundary, and failure boundary corresponding to the disturbance variable based on behavior classification, include: After each round of perturbation, if the CMIS host directly continues to execute the next event in the baseline event sequence, the current perturbation value is recorded as acceptance to determine the direct acceptance boundary; After each round of perturbation, if the CMIS host retryes the current transaction a limited number of times before continuing to execute the next event in the baseline event sequence, or re-executes the current configuration-related transaction before continuing to execute the next event in the baseline event sequence, then the current perturbation value is recorded as recoverable and acceptable, so as to determine the recoverable and acceptable boundary. After each round of perturbation, if the CMIS host continues to repeatedly read the current transaction or waits for more than a preset observation time, or stops subsequent CMIS access, the current perturbation value is recorded as unacceptable to determine the failure boundary.

6. The method of claim 1, wherein, The data structure information of the host behavior model includes basic information, event information, disturbance information, boundary information, and behavior branches: the basic information includes host model, software version, and CMIS version; the event information includes event ID, triggering condition, event window, and related field set; the disturbance information includes disturbance type, starting value, stepping method, and stopping condition; the boundary information includes the direct acceptance boundary, the recoverable acceptance boundary, and the failure boundary; and the behavior branches include host feedback characteristics such as continue, retry, reconfigure, timeout, or stop access.

7. The method of claim 1, wherein, The process involves loading the host behavior model, measuring the actual response parameters of the optical module under test corresponding to the perturbation variable under the target CMIS semantic event, comparing the actual response parameters with the boundaries corresponding to the target CMIS semantic event and the perturbation variable in the host behavior model, and outputting compatibility verification results, including: The optical module under test is driven to execute the target CMIS semantic event specified in the host behavior model, the actual response parameters of the optical module under test are collected, and the response parameters are normalized according to the compatibility degradation direction. If the actual response parameter is less than or equal to the direct acceptance boundary, then the optical module under test is determined to be compatible with the target CMIS semantic event; If the direct acceptance boundary < the actual response parameter ≤ the recoverable acceptance boundary, then the optical module under test is determined to be recoverable compatible or low-risk in the target CMIS semantic event; If the recoverable acceptable boundary is less than the actual response parameter and less than the failure boundary, then it is determined to be a critical risk; If the actual response parameter is greater than or equal to the failure boundary, it is determined to be incompatible. When the incompatibility is determined, the failure event ID, the actual response parameter, the corresponding tolerance boundary, and the subsequent behavior of the host are output, and the retesting items are suggested.

8. A host behavior model based optical module verification system, comprising: The optical module verification system includes: The transaction acquisition and identification module is used to record the baseline transaction process through the interaction agent device when the CMIS host and the reference optical module are interacting normally, and to parse the baseline transaction process according to the CMIS management protocol and divide it into multiple CMIS semantic events to obtain the baseline event sequence. The single-variable perturbation test module is used to select a target CMIS semantic event. In each round of testing, only one perturbation variable is changed within the event window corresponding to the target CMIS semantic event. The controlled perturbation is applied to the CMIS host through the interactive agent device. The host tolerance boundary search module is used to observe and record the subsequent interaction behavior of the CMIS host after being disturbed, and extract the direct acceptance boundary, recoverable acceptance boundary and failure boundary corresponding to the disturbance variable according to the behavior classification. The host behavior model generation module is used to integrate event information, disturbance information, boundary information, and host behavior branches obtained from multiple rounds of testing, according to the target CMIS semantic event and the corresponding disturbance variable, to generate a reusable host behavior model. The compatibility comparison and verification module is used to load the host behavior model, measure the actual response parameters of the optical module under test under the corresponding target CMIS semantic event and the corresponding disturbance variable, compare the actual response parameters with the boundary corresponding to the target CMIS semantic event and the disturbance variable in the host behavior model, and output the compatibility verification result.

9. A host behavior model based optical module verification apparatus, comprising: The optical module verification device based on host behavior model includes: a memory and at least one processor, wherein the memory stores instructions, and the memory and the at least one processor are interconnected via a line; The at least one processor invokes the instructions in the memory to cause the host behavior model-based optical module verification device to execute the host behavior model-based optical module verification method as described in any one of claims 1-7.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that When the computer program is executed by the processor, it implements the optical module verification method based on the host behavior model as described in any one of claims 1-7.