A method and system for automatically deriving a cross-layer calling relationship of Xincheng assets

Abstract

The application discloses a kind of Xinxin assets cross-layer calling relationship automatic derivation method and system.The method includes collecting transaction tracking data and host network connection data, the calling relationship between the assets entity in multi-level asset account is generated, the incremental running state of calling record, hierarchical attribution mapping and topological edge is maintained;In response to account change, incrementally update topological edge, form positive change record and negative change record, and determine the calling relationship of shared endpoint;Based on change record, identify non-Xinxin asset entity reduction and Xinxin asset entity increase, determine substitution relationship through calling boundary and bidirectional cross constraint, and update the lower source of topological edge accordingly.The application can realize the continuous derivation of cross-layer calling relationship in Xinxin substitution process.

Description

Technical Field

[0001] This invention relates to the fields of information technology and information technology innovation asset management technology, specifically to a method and system for automatically deriving cross-layer calling relationships of information technology innovation assets. Background Technology

[0002] In the operation and maintenance management of IT assets under the domestic IT innovation environment, determining the calling relationship between assets is an important foundation for building the operation topology and supporting the decision-making of domestic IT innovation replacement.

[0003] Currently, there are several main methods for determining asset call relationships: distributed link tracing, which involves deploying probes on the application side to collect trace and span data, reconstructing the call chain of requests between various services, and then aggregating it to form a service-level call topology; host network connection data, which involves collecting TCP connection information at the operating system layer to obtain the communication relationships between processes and hosts; and configuration management database, which involves obtaining asset configuration attributes and relationships through automatic discovery methods such as agent scanning and network scanning.

[0004] In the management of IT innovation assets, the asset ledger records the brand, version, and whether it is domestically produced of various assets, which is the basic basis for determining whether an asset belongs to IT innovation assets. IT innovation replacement usually adopts a gray-scale switching strategy, in which non-IT innovation assets and IT innovation assets operate in parallel for a period of time and are gradually switched.

[0005] Applying the above technologies to multi-level asset operation topology management in a domestic IT innovation environment presents the following challenges: Existing call relationship determination technologies are typically geared towards a single level, with link tracing corresponding to the service layer and network connection data corresponding to the host or process layer. When a domestic IT innovation replacement occurs in the underlying asset entity, causing changes in the topology edges, the asset entities at both ends of the upper-level topology edges may not actually have changed after the hierarchical attribution mapping. However, the underlying changes will still propagate layer by layer, generating unnecessary upper-level topology changes and increasing the complexity for operations and maintenance personnel in interpreting topology changes. In scenarios where multiple applications share resources such as databases and middleware, the same underlying communication endpoint may be mapped to multiple upper-level asset entities. It is difficult to determine the actual attribution of the call based on IP and port information. In a domestic IT innovation environment, the replacement of shared resources involves multiple... Adjustments to the relationships between upper-layer application systems and uncertain ownership of shared endpoints can lead to undetermined ownership of related call relationships, making it impossible to determine some domestic IT innovation replacement relationships. Furthermore, during the domestic IT innovation replacement process, non-domestic IT assets are gradually withdrawn while domestic IT assets are gradually integrated. Changes in call relationships include not only topological changes caused by domestic IT innovation replacement but also routine changes due to fluctuations in business load. These two types of changes are difficult to distinguish in terms of the increase or decrease in the number of call records. Especially in scenarios where replacement is not one-to-one, the replacement progress at each call boundary may be asynchronous; some may be completed while others are still in parallel or have already been rolled back. Without a targeted mechanism, it is difficult to identify the call relationship transfer process and its completion status caused by domestic IT innovation replacement from continuously changing operational data, based on call boundaries. Summary of the Invention

[0006] To this end, the present invention discloses an automatic derivation method for cross-layer access relationships of information technology innovation assets, characterized by comprising: S1. Collect transaction tracking data and host network connection data, and generate call records in a standardized manner; based on the asset ledger, map the call record endpoints to asset entities at each preset level; maintain the call records, level attribution mapping, and topology edge set within the time window as incremental running status; S2. Respond to changes in the ledger, retrieve affected call records, and refresh the hierarchical attribution mapping corresponding to the changed fields. Update the topology edges incrementally accordingly. When a lower-level topology edge is added or removed, map the corresponding change to the upper-level topology edge through the upper-level attribution. When mapped to the same upper-level topology edge and the net change amount is 0, retain the corresponding positive change record and negative change record. Retain shared endpoints for judgment and determine the topology edges based on the attribution of upstream and downstream endpoints in the trace. S3. Based on the positive and negative change records, distinguish between the decrease of non-IT innovation asset entities and the increase of IT innovation asset entities; determine the call boundary using endpoints whose affiliation remains unchanged before and after the change; based on multiple call boundaries, apply bidirectional cross constraints to candidate IT innovation asset entities and candidate non-IT innovation asset entities, and determine the substitution relationship when both bidirectional results are unique; update the lower-level source of the topology edge according to the substitution relationship, and keep the upper-level topology edge unchanged when mapping to the same upper-level topology edge before and after the substitution.

[0007] This invention also discloses an automatic derivation system for cross-layer access relationships of information technology innovation assets, characterized by comprising: The data processing module is used to collect transaction tracking data and host network connection data, generate call records, and map the endpoints of the call records to asset entities layer by layer based on the asset ledger to form an incremental running status. The incremental update module is used to respond to changes in the ledger. It updates the topology edge set according to the hierarchical affiliation before and after the change, forming positive change records and negative change records, and determines the affiliation of call records involving shared endpoints. The substitution relationship derivation module is used to determine the substitution relationship between non-IT innovation asset entities and IT innovation asset entities based on the positive change records and negative change records, and update the lower-level source of the topology edge based on the substitution relationship.

[0008] The advantages of this invention are: 1. This invention can address call relationships involving shared databases, middleware, message queues, and gateways. It retains call records that may be filtered by the same entity due to multiple attributions, pending judgment. It treats continuous shared call intervals as a whole and restores the actual callers based on the unique non-shared endpoints of the upstream and downstream of the interval. This avoids filtering or merging call relationships between different upper-layer asset entities in shared asset scenarios, reducing omissions, incorrect associations, and incorrect merging of call edges between different upper-layer asset entities. It improves the completeness and accuracy of the actual business dependencies in cross-layer call topologies and provides more complete call boundary inputs for subsequent replacement relationship derivation when domestic IT innovation involves shared assets.

[0009] 2. This invention can stop the corresponding changes from propagating upwards when lower-level topology edges are added or removed, but the source asset entity, target asset entity, and existence status of the upper-level topology edge remain unchanged after hierarchical mapping. At the same time, it retains the positive and negative change records of the lower-level edges that form the upper-level topology edge. This ensures that when the upper-level topology call relationship remains stable, it can still record the addition or removal of lower-level asset entities associated with the call relationship. This avoids the loss of underlying implementation changes such as the exit of non-IT innovation assets and the access of IT innovation assets during the layer-by-layer aggregation process, reduces the need for repeated updates to the upper-level topology that has not actually changed, and achieves stable business call relationships and continuous recording of changes in underlying IT innovation assets.

[0010] 3. Based on the retained positive and negative change records, this invention can form multiple call boundaries by using call endpoints that remain unchanged before and after the change. By combining the co-occurrence relationship between domestic and non-domestic innovation assets in the multiple call boundaries, as well as the bidirectional uniqueness result, the actual replacement relationship can be determined. There is no need to pre-configure the migration correspondence between non-domestic innovation assets and domestic innovation assets. When IP, instance identifier, deployment location, or number of calls changes, or when multiple domestic innovation instances are accessed in parallel, the possibility of incorrect pairing of new and old assets can be reduced. This enables automatic determination from production operation data of which domestic innovation asset actually replaces the non-domestic innovation asset.

[0011] 4. When the replacement relationship is determined for the first time, the present invention can take a snapshot of each call boundary that actually has a valid call record before the replacement by non-IT innovation assets. Subsequent operational changes can be verified according to the snapshot records to see whether each original call boundary has been replaced by the determined IT innovation assets. It can distinguish between the original call that has completed the IT innovation replacement and the original call that has directly stopped existing, identify the call relationship that has not yet completed the replacement, and determine the completion of IT innovation switching only after all snapshot call boundaries have completed the transfer, thereby improving the accuracy of the determination of the completion status of IT innovation switching. Attached Figure Description

[0012] Figure 1 Flowchart of the method for automatically deriving cross-layer access relationships of information technology innovation assets; Figure 2 System architecture diagram for automatic derivation of cross-layer access relationships for IT innovation assets Detailed Implementation Example 1, such as Figure 1 As shown in the figure, this embodiment of the invention provides a method for automatically deriving cross-layer call relationships of information technology innovation assets, the method comprising: S1 collects transaction tracking data and host network connection data, and standardizes the generation of call records from the source end to the target end; based on the asset ledger, it maps the endpoints of each call record to multiple preset asset entities level by level; and maintains the call record set, level attribution mapping and topology edge set within the time window as an incremental running state.

[0013] An endpoint refers to the two running objects involved in a call in a call record, including the source end that initiates the call and the target end that receives the call. Endpoints can be identified by one or more of the following: IP address, port, process identifier, host identifier, component instance identifier, or application instance identifier.

[0014] Specifically, transaction trace data and host network connection data are collected from the runtime environment. Transaction trace data is used to obtain the order of calls between spans and the corresponding endpoint information during a transaction process, including at least trace identifier, span identifier, source end information, target end information and collection time. Host network connection data is used to obtain the actual network connections established between running processes, including at least source IP, source port, target IP, target port, protocol, process identifier and collection time.

[0015] Data from different sources is processed in a unified manner, and the data is organized into call records with a unified data structure. Each call record includes at least a call record identifier, source identifier, target identifier, call direction, source, and collection time. The source identifier and target identifier may include one or more of the following, depending on the data that can be obtained: IP address, port, process identifier, host identifier, component instance identifier, or application instance identifier.

[0016] For temporary source ports dynamically allocated by the operating system in host network connections, the source port is unified as a combination of the source IP and process identifier before generating call records. For example, if the same business process accesses the same database successively through 10.10.1.21:49172 and 10.10.1.21:49308, and the process identifier corresponding to both connections is P1, then the source port of both connections is unified as 10.10.1.21+P1, which can prevent the same process from being identified as multiple different call endpoints due to changes in temporary ports. If the target port is a fixed service port, the target IP, target port, and obtainable asset identifiers can continue to be retained as target port identification information.

[0017] After standardization, the asset ledger is queried for both the source and target ends of each call record. The asset entity to which the record belongs is determined layer by layer according to a preset hierarchy. In this embodiment, the preset hierarchy includes, in sequence, the process layer, component layer, host layer, user layer, and deployment environment layer. The affiliation between each layer is determined based on the existing instance relationships in the asset ledger. For example, if process P1 belongs to middleware component M1, component M1 belongs to application A, process P1 runs on host H1, and host H1 is located in deployment environment E1, then endpoint P1 is associated with the process layer P1, the component layer M1, the host layer H1, the user layer A, and the deployment environment layer E1. Furthermore, for database and middleware assets, entity matching can be completed by combining fields such as IP address, port, instance name, and system affiliation in the ledger.

[0018] Preferably, for each call record, the asset entity identifiers obtained by the source and target ends at each level are saved to form a hierarchical attribution mapping. For example, a call record of process P1 accessing database process P2 can be mapped to business component M1 to database component D1 at the component level, to host H1 to host H2 at the host level, and further mapped to application A to application B at the user level. This allows the same underlying running call record to serve as the common source of multiple hierarchical call relationships.

[0019] For any level, when both the source and target ends of a call record are uniquely mapped to a single asset entity, the asset entity identifiers at both ends are compared. If both ends belong to the same asset entity, no topological edge is generated for that level. If both ends belong to different asset entities, a directed topological edge is determined using the source and target asset entities. For multiple call records mapped to the same source and target asset entities, the number of associated call records for that topological edge is increased. Each topological edge records at least the source asset entity, the target asset entity, and the number of currently valid associated call records.

[0020] When any endpoint at any level can be mapped to multiple asset entities simultaneously based on the current asset ledger, the call record is retained as a pending record at the corresponding level. The multiple asset entities obtained by the endpoint are retained. The pending record is still stored in the call record set for subsequent steps to further determine the call relationship based on the added trace or other attribution criteria.

[0021] A time window is set to maintain the current valid operational relationships. This time window can be determined based on the operational data collection cycle and business call frequency, such as the most recent 30 minutes, 1 hour, or other durations that can cover several consecutive collection cycles. When a new call record enters the time window, it is written into the call record set, and a corresponding hierarchical attribution mapping is established. For call records with unique and different attributions at both ends, a topological edge is created synchronously, or the number of associated call records of an existing topological edge is increased.

[0022] When the collection time of a call record exceeds the time window, the call record is removed from the current call record set, and the number of associated call records for the associated topology edge is deducted simultaneously. If, after deduction, the topology edge still has at least one valid associated call record, the topology edge is retained; if the number of associated call records drops to 0, the topology edge is removed from the current topology edge set. In addition, for call records pending judgment, they are also removed from the pending judgment records after the time window has expired.

[0023] S2, in response to the ledger change, retrieve the call records whose affiliation is affected from the incremental running status, refresh the hierarchical affiliation mapping of the endpoints according to the changed fields, update the topology edge set according to the affiliation before and after the change, and retain the positive change records and negative change records formed by the incremental update of the topology edges; mark the endpoints mapped to multiple asset entities as shared endpoints of this level, and retain the call records in which the shared endpoints participate for judgment; when a new trace is received, determine the continuous shared call interval along the span sequence, and determine the topology edge according to the non-shared endpoints with unique affiliation upstream and downstream of the interval.

[0024] Specifically, one or more changes to asset attributes that occur during a single update of the asset ledger are considered as a ledger change batch. Each ledger change batch must at least obtain the identifier of the asset entity that has been changed, the hierarchical level of the asset entity, the fields changed in this update, and the field values ​​before and after the change. Based on the asset entity identifier, the call records belonging to the asset entity from the source or target end are retrieved from the incremental running status maintained by S1, and only the hierarchical affiliation of the call records is updated.

[0025] Preferably, a reverse index for call records is established according to the asset entity identifiers at each level. This allows for the direct retrieval of call records with the asset entity as the source or target when a change occurs in the ledger, without the need for re-traversal. The reverse index is only used for call record retrieval and can be established using conventional methods such as key-value indexes.

[0026] For the retrieved call records, it is determined whether the changed fields include the ledger fields on which the original hierarchical attribution mapping was based. If the ledger fields on which the mapping was based have not changed, and the endpoint can be directly identified by the entity identifier carried by the collected record itself, then the original hierarchical attribution mapping and related topological edges of this endpoint remain unchanged. For example, if a process call record already carries a process identifier that can uniquely identify process P1, and this ledger change only modifies the operation and maintenance attributes of host H1, then it is not necessary to re-determine the attribution of P1 at the process layer.

[0027] If the changed field is involved in the corresponding level's attribution mapping, the attribution asset entity of the endpoint at that level is re-determined based on the changed ledger value, and the attribution before and after the change is saved separately. If the entity identifier carried by the collected record itself is inconsistent with the attribution obtained from the ledger mapping, both the asset entities determined by the record and the ledger are added to the attribution entity set of the endpoint at the current level. When the attribution entity set contains two or more asset entities, the endpoint is marked as a shared endpoint at the current level.

[0028] After updating endpoint attribution, the corresponding topology edges are corrected based on the hierarchical attribution before and after the change. Specifically, for call records where both the source and target endpoints were uniquely attributed before the change and belonged to different asset entities, the associated number of the call record is deducted from the original topology edge. Preferably, one valid call record corresponds to a deduction of 1. When the number of associated call records for the original topology edge decreases to 0, the topology edge is removed from the current topology edge set, forming a negative change record.

[0029] For call records where both the source and target ends are uniquely mapped to a single asset entity after the change, and belong to different asset entities, the corresponding number of associated call records is increased in the new topology edge. If the topology edge does not yet exist, it is created to form a positive change record. If the topology edge already exists, the corresponding number is increased based on the original number of associated call records. When any endpoint after the change belongs to a shared endpoint of the current level, the corresponding call record is retained as pending judgment.

[0030] The positive change record is used to indicate the addition of a topological edge or the increase of associated sources, and the negative change record is used to indicate the removal of a topological edge or the decrease of associated sources. Each change record retains at least the level at which the change occurred, the source asset entity, the target asset entity, the change type, the associated call record, and the change time, so that it can be distinguished later from which lower-level call relationships caused the same topological edge change.

[0031] Furthermore, the changes in the topology edges at this level are mapped to the next higher level according to the upper-level affiliation of the source and target ends. For multiple changes in the lower-level topology edges that are mapped to the same upper-level source and target asset entities, the increases and decreases are merged. Positive changes are recorded as increases, and negative changes are recorded as decreases. When the net change after merging the increases and decreases of the same upper-level topology edge is zero, the changes are no longer propagated to higher levels.

[0032] When the number of lower-level topological edges has increased or decreased, but after upper-level attribution mapping, the source asset entity, target asset entity, and the state of the topological edge remain unchanged before and after the change, positive change records and negative change records that formed the lower-level change are still retained, and the number of positive change records and negative change records are not required to be equal.

[0033] For example, if application component M1 originally called non-domestic database component D1, after changes in the ledger and operational calls, the associated call records from M1 to D1 decreased by 8, while 11 new call records from M1 to domestic database component D2 appeared. In this case, the negative change records from M1 to D1 and the positive change records from M1 to D2 are retained at the component layer. If D1 and D2 both belong to application B at the user layer, and M1 belongs to application A, then both types of component layer changes are mapped to user layer A to B, and the user layer topology edges still exist before and after the changes.

[0034] For shared endpoints formed due to conflicts in ledger ownership or one asset serving multiple upper-level assets simultaneously, processing continues based on the ownership status of the peer endpoint in the call record. If the peer endpoint uniquely belongs to one asset entity, and that asset entity is one of multiple owned asset entities at the current level of the shared endpoint, then the call record is marked as pending judgment; if both endpoints in the call record are shared endpoints at the current level, the call record is also retained pending judgment.

[0035] The pending call record is retained in the call record set maintained by S1, and retains the call direction, the set of entities to which the endpoint belongs, the trace identifier, and the time information, which are used to determine its upstream and downstream non-shared endpoints along the transaction call chain.

[0036] Upon receiving new transaction tracing data, a span sequence is formed based on the parent-child relationship of each span within the same trace and the order of calls. The call records obtained from the span conversion are checked sequentially. When the intermediate endpoints of several consecutive call records all belong to shared endpoints at the target level, the consecutive call records are determined as a consecutive shared call interval. The nearest non-shared endpoint is searched upstream and downstream along the call chain along the interval. When both the upstream and downstream non-shared endpoints can be uniquely mapped to an asset entity at the current level, the topological edge corresponding to the interval is determined based on the two uniquely belonging asset entities.

[0037] For example, in the component layer, business component M1 calls shared MQ component Q1, and Q1 then passes the message to business component M2. If Q1 belongs to multiple applications, and the final cross-application call relationship cannot be determined solely based on M1 to Q1 or Q1 to M2, then both call records are kept pending. When the span sequence of the same trace indicates that M1 to Q1 to M2 occurs consecutively, and M1 uniquely belongs to application A and M2 uniquely belongs to application B, A and B are used as the non-shared belonging entities on both sides of the shared call interval to determine the topological edge from application layer A to B. If the belonging cannot be uniquely determined on either side of the interval, it is kept pending.

[0038] After determining the topological edge through continuous shared call intervals, each call record within the interval is registered as the associated source of that topological edge, and a trace identifier used to determine the relationship is recorded. The same topological edge can be associated with multiple valid trace identifiers simultaneously. Preferably, the validity period of the trace identifier adopts a time range consistent with the time window of the S1 call record.

[0039] When all trace identifiers associated with a topological edge formed by a certain shared call interval have expired, check whether there are other valid call records or other valid attribution criteria for the topological edge. If there are other valid sources of association, the topological edge is retained. If all trace identifiers have expired and there are no other valid sources of association, the topological edge is removed from the current set of topological edges.

[0040] For pending call records where the hierarchical affiliation cannot be determined solely by relying on trace, a second type of affiliation basis related to the corresponding running asset can also be read. This second type of affiliation basis includes the association between the user identifier and the endpoint in the database session, the authorization relationship between the producer and consumer in the message queue, and the registration relationship between the service instance and the registration subject in the name service.

[0041] Specifically, when the database endpoint is a shared database, the actual business affiliation of the calling end can be determined based on the user identifier of the current database session; when the MQ endpoint serves multiple applications simultaneously, the asset entities corresponding to both ends of the call record can be determined based on the authorization relationship between the producer and consumer; when a service is registered through a name service and accessed by multiple callers, the asset entity to which the endpoint belongs can be determined with the assistance of the service instance registration relationship. When the second type of affiliation criteria enables both ends of the call record to obtain a unique and different affiliated asset entity at the current level, a corresponding topological edge is generated accordingly; if multiple possible affiliations still exist, the process remains pending.

[0042] S3, receive the positive change record and negative change record, distinguish between the decrease of non-IT innovation asset entities and the increase of IT innovation asset entities; determine the call boundary using endpoints with unchanged affiliation; for the same non-IT innovation asset entity, obtain candidate IT innovation asset entities at multiple call boundaries and perform cross constraints, and obtain candidate non-IT innovation asset entities from the candidate IT innovation asset entities and perform cross constraints, and when both bidirectional results are unique, determine the substitution relationship; according to the substitution relationship, update the lower-level source of the topology edge, and when non-IT innovation asset entities and IT innovation asset entities are mapped to the same upper-level topology edge, keep the upper-level topology edge unchanged.

[0043] Specifically, the system receives and retains positive and negative change records generated and retained by S2 during the incremental update of the topology edges. Based on the asset entity identifier in the change record, it queries the asset ledger to obtain the corresponding asset entity's asset model type, whether it is domestically produced, brand, and version. For asset entities reduced in negative change records, if the "whether it is domestically produced" field indicates a non-domestic innovation asset entity, the change record is identified as a non-domestic innovation exit record. For asset entities added in positive change records, if the asset entity is a domestic innovation asset entity, the change record is identified as a domestic innovation access record. Other change records that do not meet the above conditions continue to be retained as ordinary topology change records.

[0044] For non-IT innovation exit records and IT innovation access records, the two ends of the changed topology edge are checked respectively. If the asset entity corresponding to one endpoint changes, while the other endpoint maintains the same hierarchical affiliation before and after the change, the endpoint with unchanged affiliation is used as the comparison benchmark. The call boundary is constructed according to the upper-level affiliation asset entity of the changed asset entity, the peer entity with unchanged affiliation, the call direction, and the asset model type. The call direction is determined according to the position of the changed asset entity relative to the endpoint with unchanged affiliation, including two directions: as the caller and as the callee.

[0045] For example, component D1 is a non-IT innovation database component, and its upper layer belongs to user-layer asset B. Component M1 calls D1. During the migration, M1 to D1 decreases, while M1 to D2 appears, where D2 is an IT innovation database component and also belongs to user-layer asset B. Since M1 remains unchanged before and after the change, a call boundary can be determined according to "upper-layer asset entity B, peer entity M1, call direction is called, and asset model type is database component". The negative change from M1 to D1 and the positive change from M1 to D2 are written into the same call boundary.

[0046] When the hierarchical affiliation of both ends of a change record changes simultaneously, or when it is impossible to determine the counterpart asset entity that remains unchanged before and after the change, the corresponding change record will be retained for judgment. The change record to be judged will continue to be retained within the time window set in S1. When a positive or negative change record that can determine the call boundary appears in a subsequent batch, it will then participate in the determination of the replacement relationship.

[0047] Furthermore, cross-constraints are applied to the changes of the same non-IT innovation asset entity across multiple call boundaries. For a non-IT innovation asset entity X, in each call boundary where both a non-IT innovation exit record and an IT innovation access record of X appear simultaneously, IT innovation asset entities with the same call boundary as X are obtained, forming a set of IT innovation asset entities under the call boundary. The sets of IT innovation asset entities obtained by X across multiple call boundaries are cross-compared, and only IT innovation asset entities that appear repeatedly in multiple call boundaries are retained as candidate IT innovation asset entities of X.

[0048] Preferably, the cross constraint is executed when there are two or more call boundaries. When only a single call boundary is obtained, the substitution relationship can be temporarily determined, and the relevant change records can be retained until subsequent changes arrive within the time window. This aims to avoid directly determining the substitution relationship based solely on the reduction or increase of a single occasional call.

[0049] For example, before replacement, non-IT innovation database component D1 has three call boundaries: M1 to D1, M2 to D1, and D1 to M3. During the migration, IT innovation database components D2 and D3 appear in the first call boundary, only D2 appears in the second call boundary, and D2 also appears in the third call boundary. After cross-comparison of multiple call boundaries, only D2 remains in multiple boundaries. Therefore, D2 is considered a candidate IT innovation asset entity for D1. The above processing does not require the number of call records reduced and increased in each call boundary to be exactly equal. The number of calls is only used to determine the number of transfers later.

[0050] After obtaining candidate IT innovation asset entities, reverse cross-constraints are executed starting from these candidate IT innovation asset entities. Specifically, for a candidate IT innovation asset entity Y, non-IT innovation asset entities that have non-IT innovation exit records at the same call boundary as Y are obtained, and cross-comparisons are performed across multiple call boundaries to form candidate non-IT innovation asset entities for Y. If, after forward cross-constraint of non-IT innovation asset entity X, only IT innovation asset entity Y is obtained, and after reverse cross-constraint of IT innovation asset entity Y, only non-IT innovation asset entity X is obtained, then a substitution relationship is determined between X and Y.

[0051] For example, D1 may only obtain D2 after positive cross constraint. However, if D2 corresponds to the exit changes of D1 and another non-IT innovation database component D4 in the same multiple call boundaries, the reverse result is not unique. Only when the non-IT innovation asset entity obtained by D2 in reverse is also uniquely D1, it is determined that D1 is replaced by D2.

[0052] When multiple IT innovation asset entities still exist after cross constraints, it is further determined whether the multiple IT innovation asset entities repeatedly appear in multiple call boundaries, and the model dictionary is read to determine the asset model type. If the asset model types of multiple IT innovation asset entities are the same, or satisfy the substitutable model relationship pre-registered in the model dictionary, then the non-IT innovation asset entities and the multiple IT innovation asset entities are determined to be in a group substitution relationship; the same processing is used when multiple non-IT innovation asset entities exist.

[0053] The alternative model relationships are preferably pre-registered based on the asset's purpose and model type. For example, "relational database component with relational database component" and "message queue component with message queue component" can be registered in the model dictionary as model relationships that allow for substitution.

[0054] After the substitution relationship is determined, for each call boundary, the absolute value N1 of the number of negative change call records in the non-IT innovation exit records and the number N2 of the number of positive change call records in the IT innovation access records are counted. Based on this, the number of transfers in this change for the call boundary is determined. Preferably, the number of transfers T in this batch can be determined as the smaller value between N1 and N2, i.e., T=min(N1, N2). For example, if the non-IT innovation asset entity in the same call boundary reduces 8 valid call records, while the IT innovation asset entity increases 11 valid call records, then the number of transfers that can be determined in this change is 8, and the remaining 3 newly added call records are retained as the call sources added by the IT innovation asset entity itself.

[0055] Furthermore, when a substitution relationship is first determined for a certain non-IT innovation asset entity, all call boundaries with valid call records of the non-IT innovation asset entity before the substitution occurs are read from the incremental operation status maintained by S1, forming a call boundary snapshot of the non-IT innovation asset entity. The call boundary snapshot records at least the upper-level asset entity, peer entity, call direction, asset model type and valid call relationship before substitution corresponding to each call boundary. When the snapshot is established, each call boundary is initially marked as non-IT innovation operation status.

[0056] When subsequent positive and negative change records arrive, the status in the call boundary snapshot is updated respectively. When both non-IT innovation asset entities and identified IT innovation asset entities have valid call records at the same call boundary, the call boundary is marked as being transferred. When a non-IT innovation asset entity no longer has a valid call record at the call boundary, but the corresponding IT innovation asset entity or set of IT innovation asset entities has a valid call record, the call boundary is marked as being replaced by IT innovation. When the call record of a non-IT innovation asset entity has been removed, but no valid call record of an identified IT innovation asset entity is found at the call boundary, the call boundary is marked as not being replaced.

[0057] For example, before the replacement of the non-IT-based database component D1, there are three call boundaries: M1 to D1, M2 to D1, and M3 to D1. If D2 has completely replaced M1 to D1, and M2 to D1 and M2 to D2 exist simultaneously, then the former call boundary is marked as having been replaced by IT-based technology, and the latter call boundary is marked as being transferred. If M3 to D1 disappears and M3 to D2 has not yet been found, then the corresponding call boundary is marked as not having been replaced.

[0058] If a domestic IT asset entity with a confirmed replacement relationship experiences a negative change at a certain call boundary, and a non-domestic IT asset entity experiences a positive change again at the same call boundary, then it is determined that the call boundary has rolled back. Based on the number of domestic IT asset entities reduced and the number of non-domestic IT asset entities restored, the number of transfers already recorded is deducted. Preferably, the number of rollbacks is the smaller of the two types of changes.

[0059] When the IT innovation asset entity still retains some valid call records, the call boundary is marked as a rollback state; when the IT innovation asset entity no longer has valid call records on its call boundary, and a non-IT innovation asset entity becomes a valid call source again, the call boundary is restored to the non-IT innovation running state.

[0060] The system continuously checks each call boundary in the call boundary snapshot. When all call boundaries in the snapshot change from non-IT innovation running, transfer in progress, or unreplaced state to IT innovation replaced state, the state of the non-IT innovation asset entity in the corresponding cross-layer call relationship is determined as the switchover is complete. If there is still any call boundary in the transfer in progress, unreplaced, rollback, or non-IT innovation running state, the corresponding state is maintained and the switchover is not determined as complete.

[0061] The lower-level sources recorded in the corresponding topology edges are updated according to the status of each call boundary. When the call boundary is in a transition state, both non-IT innovation asset entities and IT innovation asset entities are retained as the lower-level sources of the corresponding topology edges. When the call boundary is in a state where IT innovation has been replaced, the lower-level sources of the topology edges are updated from non-IT innovation asset entities to IT innovation asset entities or a set of IT innovation asset entities. When the call boundary is in a state where it has not been replaced, the original non-IT innovation asset entities are retained as the lower-level sources. For related call records that are still in a pending state, the original lower-level sources and the pending state are retained.

[0062] For non-IT-compliant and IT-compliant asset entities, even if they form different asset entities and different topological edges at the lower level, they still belong to the same upper-level asset entity after hierarchical mapping. In this case, only the lower-level source recorded in the upper-level topological edge is updated; it is not necessary to delete and re-establish the upper-level topological edge. For example, if non-IT-compliant database component D1 and IT-compliant database component D2 both belong to user-level asset B, and the calling terminal component M1 belongs to user-level asset A, then the component-level topology can gradually change from M1 to D1 to M1 to D2. After mapping to the upper level, both form A to B. Therefore, during the process of confirming the substitution relationship from D1 to D2 and updating the lower-level source, the user-level A to B remains the same topological edge, thus preserving the continuity of cross-level call relationships during the lower-level operation implementation switchover.

[0063] Example 1, such as Figure 2 As shown, this embodiment of the invention provides an automatic derivation system for cross-layer call relationships of domestically developed assets. The method includes: a data processing module, used to collect transaction tracking data and host network connection data, generate call records, and map the endpoints of the call records to asset entities layer by layer based on the asset ledger, forming an incremental running state; an incremental update module, used to respond to ledger changes, update the topology edge set according to the hierarchical affiliation before and after the change, form positive change records and negative change records, and determine the affiliation of call records involving shared endpoints; and a substitution relationship derivation module, used to determine the substitution relationship between non-domestic developed asset entities and domestically developed asset entities based on the positive change records and negative change records, and update the lower-layer source of the topology edges according to the substitution relationship.

Claims

1. A method for automatically deriving cross-layer access relationships of information technology innovation assets, characterized in that, include: S1. Collect transaction tracking data and host network connection data, and generate standardized call records; Based on the asset ledger, call record endpoints are mapped to asset entities at each preset level; call records, level attribution mappings, and topology edge sets within the maintenance time window are used as incremental running states; S2. Respond to changes in the ledger, retrieve the affected call records, and refresh the hierarchical attribution mapping corresponding to the changed fields, thereby incrementally updating the topology edges. When a lower-level topology edge is added or removed, the corresponding change is mapped to the upper-level topology edge through the upper-level attribution. When mapped to the same upper-level topology edge and the net change amount is 0, the corresponding positive change record and negative change record are retained. Shared endpoints are retained pending judgment, and the topology edge is determined based on the upstream and downstream endpoints of the trace. S3. Based on the positive and negative change records, distinguish between the decrease of non-IT innovation asset entities and the increase of IT innovation asset entities; determine the call boundary using endpoints whose affiliation remains unchanged before and after the change; Based on multiple call boundaries, bidirectional cross constraints are applied to candidate IT innovation asset entities and candidate non-IT innovation asset entities. When both bidirectional results are unique, the substitution relationship is determined. The lower-level source of the topology edge is updated according to the substitution relationship. When mapping to the same upper-level topology edge before and after substitution, the upper-level topology edge remains unchanged.

2. The method for automatically deriving cross-layer access relationships of information technology innovation assets according to claim 1, characterized in that, In S1: Before standardization, the source temporary port is normalized to a combination of IP and process identifier; in the hierarchical mapping, for call records that are unique at both ends, only call records belonging to different asset entities are included in the topology edge set; when any endpoint belongs to multiple asset entities, the corresponding call records are retained for judgment; each topology edge records the source entity, target entity, and the number of associated call records; when a call record exceeds the time window, it is removed and the topology edge is updated synchronously.

3. The method for automatically deriving cross-layer access relationships of information technology innovation assets according to claim 1, characterized in that, In S2: When the entity identifier in the collection record is inconsistent with the attribution determined by the asset ledger mapping, it is added to the attribution asset entity set of the endpoint at that level. When updating the topology edge set, the number of original topology edge associated call records is reduced by 1, and removed when it drops to 0; the number of new topology edge associated call records is increased by 1 when both ends are uniquely determined, and created if they do not exist; if any endpoint is a shared endpoint of the level, the call record is retained for further judgment. The changes in the topological edges generated by the update are mapped to the next higher level along the upper-level affiliation of each endpoint. Changes mapped to the same upper-level topological edge are merged. When the net change is 0, the upward propagation stops. The corresponding positive change records and negative change records are retained. The number of positive change records and negative change records may be different.

4. The method for automatically deriving cross-layer access relationships of information technology innovation assets according to claim 3, characterized in that, In S2: Endpoints mapped to multiple asset entities are marked as shared endpoints; if the peer belongs to any of the asset entities belonging to the shared endpoint, or if the peer is a shared endpoint, then same-entity filtering is not performed and the endpoint is retained for judgment. The pending call record is retained for upstream and downstream lookup. Based on the trace, the upstream and downstream asset entities and trace identifiers corresponding to the shared call interval are determined, and each call record within the interval serves as the source of association for the topology edge; Topological edges defined by the shared call interval are retained within the validity period of the corresponding trace identifier and removed when all trace identifiers expire and there are no other valid sources of association. The hierarchical affiliation of the call record to be judged is determined by the second type of affiliation criteria. The second type of affiliation criteria includes at least the association between the user identifier of the database session and the endpoint, the production and consumption authorization relationship of the message queue, and the registration relationship of the name service. When the second type of affiliation criteria determines that the two ends have different hierarchical affiliations, a corresponding topological edge is generated.

5. The method for automatically deriving cross-layer access relationships of information technology innovation assets according to claim 1, characterized in that, In S3: The call boundary includes at least the upper-level asset entity, the peer entity, the call direction, and the asset model type; when the entity reduced in the negative change record is a non-IT innovation asset entity, it is determined as a non-IT innovation exit record; when the entity added in the positive change record is an IT innovation asset entity, it is determined as an IT innovation access record. If the hierarchical affiliation of the peer entity remains unchanged before and after the change, the corresponding change record will be written to the call boundary; If the asset entities at both ends of the hierarchy or the endpoints change simultaneously, the corresponding change records will be retained for further judgment. Change records for which no substitution relationship has been determined in the current batch are retained within the stated time window; The cross constraint includes obtaining the candidate sets of information technology innovation assets corresponding to each call boundary for the same non-information technology innovation asset entity, and finding their intersection; the reverse cross constraint includes obtaining the candidate sets of non-information technology innovation asset entities for the candidate information technology innovation asset entities in the obtained intersection according to the same call boundary, and finding their intersection. When there are still multiple candidate IT innovation asset entities after cross-constraint, if the multiple IT innovation asset entities repeatedly appear together in multiple call boundaries, and the asset model type is the same or satisfies the preset substitutable model relationship in the model dictionary, the non-IT innovation asset entities and the set of IT innovation asset entities are determined as a group substitution relationship; the same processing is applied when there are multiple candidate non-IT innovation asset entities. Once the substitution relationship is determined, within the same call boundary, the absolute value of the number of call records with negative changes in non-IT innovation asset entities and the number of call records with positive changes in IT innovation asset entities are counted respectively; based on the number of negative changes in non-IT innovation and the number of positive changes in IT innovation within the same call boundary, the number of transferred call records of the call boundary is determined.

6. The method for automatically deriving cross-layer access relationships of information technology innovation assets according to claim 5, characterized in that, In S3: When it is first determined that there is a substitution relationship between non-IT innovation asset entities, according to the incremental operation status, each call boundary of the non-IT innovation asset entity with valid call records before substitution is read to form a call boundary snapshot. Each call boundary in the snapshot is initially marked as non-IT innovation operation status. For each call boundary in the call boundary snapshot, maintain the status of non-IT innovation running, in transfer, IT innovation replaced, not replaced, and rolled back respectively; when non-IT innovation asset entities and IT innovation asset entities have call records in the same call boundary at the same time, they are marked as in transfer. If a non-IT innovation asset entity no longer has a call record at the call boundary, and a confirmed IT innovation asset entity or set of IT innovation asset entities has a call record at the call boundary, it is marked as having been replaced by IT innovation; if the call record of a non-IT innovation asset entity is removed and no IT innovation asset entity is found to replace it, it is marked as not replaced. If a domestic IT innovation asset entity with a confirmed substitution relationship experiences a negative change at the call boundary, and a non-domestic IT innovation asset entity experiences a positive change at the same call boundary, then the call boundary is updated to a rollback state by deducting the number of transferred call records at the call boundary based on the corresponding number of changes. When all call boundaries in the call boundary snapshot are marked as having been replaced by IT innovation, the status of non-IT innovation asset entities in the corresponding cross-layer call relationship is determined as the switchover is complete; Update the lower-level source of the corresponding topology edge according to the status of each call boundary; during the transfer, mark both non-IT innovation asset entities and IT innovation asset entities as the lower-level source of the topology edge; when IT innovation has taken over, update the lower-level source from non-IT innovation asset entities to IT innovation asset entities or a set of IT innovation asset entities; when not taken over, retain the original lower-level source; if the call record is still pending judgment, retain the original lower-level source and pending judgment status.

7. An automatic derivation system for cross-layer access relationships of information technology innovation assets, characterized in that, include: The data processing module is used to collect transaction tracking data and host network connection data, generate call records, and map the endpoints of the call records to asset entities layer by layer based on the asset ledger to form an incremental running status. The incremental update module is used to respond to changes in the ledger. It updates the topology edge set according to the hierarchical affiliation before and after the change, forming positive change records and negative change records, and determines the affiliation of call records involving shared endpoints. The substitution relationship derivation module is used to determine the substitution relationship between non-IT innovation asset entities and IT innovation asset entities based on the positive change records and negative change records, and update the lower-level source of the topology edge based on the substitution relationship.

Images