A method and system for reconstructing and restoring six-dimensional coupled data in an overload saturation pathway
By establishing a six-dimensional coupled index and circular storage pages in the overload saturation path, and dividing and marking the mutual constraint relationships of the six-dimensional fields, the contradictory problems of data recovery in the overload saturation path are resolved, and the reliable and consistent recovery of data is achieved.
Patent Information
- Application Number
- CN202611593832.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-10-15
- Publication Date
- 2026-11-10
AI Technical Summary
In overload saturation pathways, existing technologies cannot effectively utilize the mutual constraints between the six-dimensional fields for data recovery, resulting in contradictory recovery results. Furthermore, the recoverable and unreliable positions of missing fields cannot be distinguished, affecting data reliability.
By establishing a six-dimensional coupled index, writing the six-dimensional fields into a circular storage page, generating a six-dimensional coupled data block, dividing the saturation interval, marking the recoverability of the mutual constraints of the six-dimensional fields, and performing cross-completion and backfilling, the coupling relationship and time alignment between fields are maintained during the recovery process.
It achieves the maintenance of coupling relationships and time alignment between six-dimensional fields in overload saturation paths, avoids internal state contradictions and timing misalignments in recovery data, and ensures that recovery data can directly participate in port queue scheduling and path state updates.
Smart Images

Figure CN123074980A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of communication data recovery technology, specifically to a method and system for reconstructing and recovering six-dimensional coupled data from an overload saturated path. Background Technology
[0002] When a transmission path becomes overloaded, data such as path occupancy timing, queue backlog, node buffer status, link error flags, service priority identifiers, and clock offsets will change simultaneously, and these data are interdependent. Path occupancy timing reflects port occupancy status, queue backlog reflects the degree of data accumulation, node buffer status reflects buffer page availability, link error flags reflect field reliability, and service priority identifiers and clock offsets affect the data slice scheduling order. Existing technologies typically store collected data in strips according to path number and timestamp. When a field is missing, recovery is achieved using methods such as single-field interpolation, mean filling, or copying neighboring values, or by directly reading backup fields from the buffer queue for backfilling.
[0003] The above methods are prone to producing contradictory field recovery results in overloaded and saturated paths. Single-field filling only considers the field's own historical values, failing to utilize the constraints between different fields. This leads to a mismatch between the recovered queue backlog and the path's occupancy timing, or inconsistencies between node cache status and link error markings. Cache backup backfilling lacks detailed segmentation of the saturation interval and field alignment processing. Backfilled fields deviate from the original context in terms of time order, business priority, and clock offset, making it difficult for recovered data to participate in scheduling decisions after being written to the path. Fragment boundaries formed during the overload saturation process are not marked, making it impossible to distinguish between recoverable and untrusted field positions. The recovery process may treat untrusted data as valid data for backfilling, affecting the reliability of the reconstructed data. Multi-source overload records have duplicate timestamps and field version changes during the acquisition phase. Existing storage methods do not generate version identifiers for fields with duplicate timestamps under the same path number, making it difficult to determine the latest state of fields during recovery.
[0004] To address the aforementioned shortcomings, it is necessary to clarify how to determine recoverable locations and perform cross-completion backfilling when multiple fields are simultaneously missing or unreliable in an overloaded saturated path, and how to organize overloaded data through six-dimensional coupled indexes and circular storage pages to maintain temporal and state coupling between fields during the recovery process. Summary of the Invention
[0005] This paper provides a method and system for reconstructing and recovering six-dimensional coupled data in overloaded saturated transmission paths. Addressing the recovery needs when multi-source data fields are missing or unreliable in overloaded saturated transmission paths, the method establishes a six-dimensional coupling index and writes the six-dimensional fields into a circular storage page to generate a six-dimensional coupled data block. It then segments the saturated interval and marks the recoverability of the mutual constraints between the six-dimensional fields. Based on the set of reconstruction and recovery marks, it performs cross-completion and backfilling, writes the reconstructed and recovered data back to the target transmission path according to the path number, and updates the path status, thus maintaining the coupling and time alignment relationships between the six-dimensional fields during the recovery process.
[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for reconstructing and recovering six-dimensional coupled data in an overload saturation pathway, the method comprising:
[0007] Acquire multi-source overload records of the target transmission path to form an overload saturated path data set. The overload saturated path data set includes path number, timestamp, and six-dimensional fields. The six-dimensional fields are path occupancy time sequence, queue backlog, node buffer status, link error mark, service priority identifier, and clock offset.
[0008] A six-dimensional coupled index is established based on the path number and the timestamp, and the six-dimensional field is written into a circular storage page to generate a six-dimensional coupled data block;
[0009] The six-dimensional coupled data block is segmented into saturation intervals to generate a set of overloaded saturated path segments;
[0010] The set of overload saturated pathway segments is marked for recoverability according to the mutual constraint relationship of the six-dimensional fields, and a set of reconstruction recovery marks is generated.
[0011] Based on the reconstructed recovery marker set, cross-filling and backfilling are performed on the missing fields of the six-dimensional coupled data block to generate reconstructed recovery data;
[0012] The reconstructed and restored data is written back to the target transmission path according to the path number, and the path status is updated.
[0013] Preferably, the step of establishing a six-dimensional coupled index based on the path number and the timestamp, writing the six-dimensional fields into a circular storage page, and generating a six-dimensional coupled data block includes:
[0014] Extract the path number and timestamp from the overload saturation path data set to generate a path event identifier;
[0015] The path event identifier is bound to the path occupancy time sequence, queue backlog, node cache status, link error mark, service priority identifier and clock offset in the overload saturated path data set, respectively, to obtain six-dimensional field slots;
[0016] The six-dimensional coupled data block is generated by setting chained pointers in the circular storage page according to the field slots of the six dimensions.
[0017] Preferably, the step of segmenting the six-dimensional coupled data block into saturation intervals to generate an overloaded saturated path fragment set includes:
[0018] Read the path occupancy timing and queue backlog in the six-dimensional coupled data block;
[0019] When the path occupancy time sequence is continuously full and the queue backlog remains non-decreasing during the corresponding time period, mark the overload start point and overload end point.
[0020] The six-dimensional fields in the six-dimensional coupled data block are extracted according to the overload start point and the overload end point to generate the overload saturation path segment set.
[0021] Preferably, the step of marking the overload saturated pathway fragment set with recoverability according to the mutual constraint relationship of the six-dimensional fields to generate a reconstructed recovery mark set includes:
[0022] Extract node buffer status, link error flags, and clock offsets from the overload saturated path segment set;
[0023] Based on the node cache status, determine the cache pages that can be refilled; based on the link error flag, determine the untrusted field bits; based on the clock offset, determine the field alignment offset.
[0024] Write the backfillable cache page, the untrusted field bit, and the field alignment offset into the reconstructed recovery tag set.
[0025] Preferably, the step of performing cross-filling and backfilling on the missing fields of the six-dimensional coupled data block based on the reconstructed recovery marker set to generate reconstructed recovery data includes:
[0026] Read the backup fields from the refillable cache pages in the reconstructed recovery mark set;
[0027] The corresponding fields in the six-dimensional coupled data block are masked according to the untrusted field bits in the reconstructed recovery mark set;
[0028] Based on the field alignment offset in the reconstructed recovery mark set, the backup field is mapped to the six-dimensional field to generate a cross-completion result;
[0029] The cross-completion results are backfilled into the empty positions of the circular storage page according to the path number to generate the reconstructed recovery data.
[0030] Preferably, the step of writing the reconstructed and restored data back to the target transmission path according to the path number and updating the path status includes:
[0031] The port queue of the target transmission path is located based on the path number in the reconstructed and restored data;
[0032] The reconstructed and restored data is divided into data slices corresponding to the business priority identifier;
[0033] The data slices are sorted by transmission time according to the clock offset;
[0034] The sorted data slices are written to the port queue, and a recovery completion flag is written to the scheduling table of the target transmission path to update the path status.
[0035] Preferably, the step of acquiring multi-source overload records of the target transmission path to form an overload saturated path data set includes:
[0036] Collect the path occupancy timing, queue backlog, node buffer status, link error markers, service priority identifiers, and clock offset of the target transmission path;
[0037] Write the above six data items into the six-dimensional field according to the channel number and timestamp;
[0038] For the six-dimensional fields with repeated timestamps under the same path number, a field version number is generated, and the six-dimensional fields with the field version number are aggregated into the overload saturation path data set.
[0039] Preferably, the step of sorting the data slices by transmission time according to the clock offset includes:
[0040] The offset direction between the local clock and the upstream clock of the target transmission path is read from the clock offset.
[0041] A transmission sequence number is assigned to each data segment according to the offset direction;
[0042] The data slices are sorted by transmission time according to the transmission sequence number and the service priority identifier.
[0043] Preferably, writing a recovery completion flag to the scheduling table of the target transmission path and updating the path status includes:
[0044] Read the occupancy threshold index of the port queue from the scheduling table of the target transmission path;
[0045] Associate the recovery completion marker with the occupancy threshold index;
[0046] After the port queue receives the recovery completion flag, it updates the credit count and link permission bit of the port queue.
[0047] Preferably, the present invention also includes an overload saturation pathway six-dimensional coupled data reconstruction and recovery system, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of the overload saturation pathway six-dimensional coupled data reconstruction and recovery method described above.
[0048] The technical effects and advantages provided by the present invention in the above technical solution are as follows:
[0049] By establishing a six-dimensional coupled index based on the path number and timestamp, path occupancy time sequence, queue backlog, node cache status, link error flags, service priority identifiers, and clock offsets are bound to path event identifiers, forming six-dimensional field slots. A chained pointer is then used in a circular storage page to generate a six-dimensional coupled data block. The six-dimensional fields remain associated within the same data block, and the field slots establish access paths based on path event identifiers via chained pointers. When overload saturation occurs, multiple dimensions of the same path event can be read synchronously. Compared to the conventional method of storing data sequentially by a single dimension and recovering data independently by field, the six-dimensional coupled index and circular storage page maintain path-level and time-level coupling for overload records. The correspondence between multi-dimensional states is not lost due to scattered field storage. Recoverability marking and cross-completion can be performed based on the complete six-dimensional context, and the recovery result no longer depends on extrapolation of historical values of a single field or neighbor copying.
[0050] By extracting node cache status, link error markers, and clock offsets from the overload saturated path segment set, refillable cache pages, untrusted field bits, and field alignment offsets are determined, generating a reconstruction recovery marker set. During cross-completion and backfilling, backup fields are read based on refillable cache pages, corresponding fields in the six-dimensional coupled data block are masked based on untrusted field bits, backup fields are mapped to six-dimensional fields based on field alignment offsets, and backfilled into vacant positions in the circular storage page. This method utilizes the mutual constraints between six-dimensional fields to distinguish refillable positions from untrusted fields. The backfill source is constrained by node cache status, the backfill object is filtered by untrusted field bits, and the timing correspondence between the backfilled field and the original field is determined by the clock offset. Unlike conventional methods that simply fill in a single field value or directly read cache backups for backfilling, cross-filling and backfilling synchronously maintain constraints between node cache status, link error markers, and clock offsets during the recovery process. This ensures that the queue backlog after backfilling is consistent with the path occupancy timing, the node cache status is consistent with the link error markers, and the service priority identifier is consistent with the sorted data slice transmission time. This avoids state contradictions or timing misalignments within the recovered data. After the recovered data is written back to the target transmission path, it can directly participate in port queue scheduling and path status updates. Attached Figure Description
[0051] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.
[0052] Figure 1 This is a flowchart of the six-dimensional coupled data reconstruction and recovery method for overload saturation pathways;
[0053] Figure 2 This is a flowchart of the overload saturation pathway segment extraction process;
[0054] Figure 3 This is a flowchart of the data generation method for reconstructing and restoring data. Detailed Implementation
[0055] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0056] See Figure 1This invention provides a method for reconstructing and recovering overload saturated transmission paths using six-dimensional coupled data. The method involves acquiring multi-source overload records of the target transmission path to form an overload saturated transmission path data set. This data set includes a path number, a timestamp, and six-dimensional fields: path occupancy time sequence, queue backlog, node buffer status, link error flag, service priority identifier, and clock offset. A six-dimensional coupling index is established based on the path number and timestamp. The six-dimensional fields are written to a circular storage page to generate a six-dimensional coupled data block. The six-dimensional coupled data block is then segmented into saturated intervals to generate an overload saturated transmission path fragment set. The overload saturated transmission path fragment set is marked for recoverability based on the mutual constraints of the six-dimensional fields, generating a reconstruction and recovery flag set. Based on the reconstruction and recovery flag set, missing fields in the six-dimensional coupled data block are cross-filled and backfilled to generate reconstructed and recovered data. Finally, the reconstructed and recovered data is written back to the target transmission path according to the path number to update the path status.
[0057] In practical implementation, the following parameters are collected for the target transmission path: path occupancy timing, queue backlog, node buffer status, link error flags, service priority flags, and clock offset. Path occupancy timing is generated by the port occupancy monitoring unit of the target transmission path, indicating the occupancy status of the target transmission path at different sampling times. Queue backlog is generated by the queue management unit of the target transmission path, indicating the amount of data waiting to be sent in the queue of the corresponding port of the target transmission path. Node buffer status is generated by the node buffer management unit of the target transmission path, indicating the occupancy status of the node buffer areas passed through by the target transmission path. Link error flags are generated by the link monitoring unit of the target transmission path, indicating the error indication of the corresponding link of the target transmission path. Service priority flags are generated by the service scheduling unit of the target transmission path, indicating the priority of the services carried by the target transmission path. Clock offset is generated by the clock synchronization unit of the target transmission path, indicating the offset value between the local clock of the target transmission path and the upstream clock.
[0058] The path occupancy timeline, queue backlog, node buffer status, link error flag, service priority identifier, and clock offset are written to a six-dimensional field according to the path number and timestamp. During the writing process, the writing order is established with the path number as the primary key and the timestamp as the secondary key. The path occupancy timeline is written to the first dimension of the six-dimensional field, the queue backlog is written to the second dimension, the node buffer status is written to the third dimension, the link error flag is written to the fourth dimension, the service priority identifier is written to the fifth dimension, and the clock offset is written to the sixth dimension.
[0059] Generate a version number for the six-dimensional field with duplicate timestamps under the same channel number. When a six-dimensional field with the same timestamp is detected under the same channel number, a version number is assigned to the six-dimensional field corresponding to the same timestamp. The initial value of the version number is 1, and the version number is incremented by 1 each time a six-dimensional field with the same channel number and the same timestamp is detected.
[0060] The six-dimensional fields with version numbers are aggregated into an overload saturation path data set. Each record in the overload saturation path data set includes the path number, timestamp, field version number, path occupancy time sequence, queue backlog, node cache status, link error flag, service priority identifier, and clock offset.
[0061] Extract the path ID and timestamp from the overload saturation path dataset to generate a path event identifier. The path event identifier is a string concatenated from the path ID and timestamp, with the path ID first and the timestamp second, using a nanosecond-level integer value.
[0062] By binding the path event identifier with the path occupancy timeline, queue backlog, node cache status, link error flag, service priority flag, and clock offset in the overload / saturation path dataset, six-dimensional field slots are obtained. Each of the six-dimensional field slots corresponds to a separate storage area, with each storage area storing the field value for one dimension. The binding method uses the path event identifier as the common index for the six-dimensional field slots, which are arranged in the order of path occupancy timeline, queue backlog, node cache status, link error flag, service priority flag, and clock offset.
[0063] A chained pointer is set in the circular storage page according to the six-dimensional field slots to generate a six-dimensional coupled data block. The circular storage page consists of multiple fixed-length storage pages and is managed using a circular write method. The chained pointers include forward pointers and backward pointers. The forward pointer points to the starting address of the next six-dimensional coupled data block with the same path number in the circular storage page, and the backward pointer points to the starting address of the previous six-dimensional coupled data block with the same path number in the circular storage page. Slot pointers are set for the six-dimensional field slots, and the slot pointers point to the offset address of the corresponding dimension field value in the circular storage page. A six-dimensional coupled data block is generated, which contains the path event identifier, the six-dimensional field slots, the chained pointers, and the slot pointers.
[0064] In specific implementation, please refer to Figure 2The process reads the path occupancy timing and queue backlog from the six-dimensional coupled data block. The read location is the offset address of the circular storage page pointed to by the field slots of the six dimensions in the six-dimensional coupled data block. When reading the path occupancy timing, the offset address of the path occupancy timing field value in the circular storage page is obtained based on the corresponding slot pointer in the six-dimensional coupled data block, and the path occupancy timing field value is read from the circular storage page. When reading the queue backlog, the offset address of the queue backlog value in the circular storage page is obtained based on the corresponding slot pointer in the six-dimensional coupled data block, and the queue backlog value is read from the circular storage page.
[0065] After reading the path occupancy timing and queue backlog from the six-dimensional coupled data block, it is determined whether the path occupancy timing is in a state of continuous full occupancy. Continuous full occupancy means that at multiple consecutive sampling times, the path occupancy timing field value is a full flag. The full flag is written by the target transmission path port occupancy monitoring unit when the port is in a state of full load occupancy. The path occupancy timing field value is scanned in timestamp order. When it is detected that the path occupancy timing field value corresponding to multiple consecutive sampling times is a full flag, it is determined that the path occupancy timing is in a state of continuous full occupancy.
[0066] When the channel occupancy time sequence is in a continuous full-occupancy state, it is determined whether the queue backlog remains non-decreasing during the corresponding period of continuous full-occupancy. The determination method is to compare the queue backlog field values at each sampling time within the corresponding period of continuous full-occupancy in chronological order. If the queue backlog field value at any later sampling time is less than the queue backlog field value at the previous sampling time, it is determined that the queue backlog has not remained non-decreasing; if the queue backlog field values at all later sampling times are greater than or equal to the queue backlog field value at the previous sampling time, it is determined that the queue backlog has remained non-decreasing during the corresponding period.
[0067] When the path occupancy time sequence is continuously at full capacity and the queue backlog remains non-decreasing during the corresponding time period, mark the overload start point and overload end point. The overload start point is the timestamp of the first sampling moment in the corresponding time period of continuous full capacity occupancy, and the overload end point is the timestamp of the last sampling moment in the corresponding time period of continuous full capacity occupancy. Write the overload start point to the overload start point marker field and the overload end point to the overload end point marker field.
[0068] After marking the overload start and end points, the six-dimensional fields in the six-dimensional coupled data block are extracted according to the overload start and end points. During extraction, the overload start point is used as the starting position and the overload end point is used as the ending position. The access timing, queue backlog, node buffer status, link error mark, service priority identifier, and clock offset between the overload start and overload end points are extracted from the six-dimensional coupled data block to obtain the overload saturated path segment.
[0069] The extracted overload saturated path segments are added to the overload saturated path segment set. Each overload saturated path segment in the set includes a path number, overload start point, overload end point, path occupancy timing, queue backlog, node buffer status, link error flag, service priority identifier, and clock offset. The overload saturated path segment set is arranged in order of the timestamp of the overload start point.
[0070] In specific implementation, please refer to Figure 3 The system extracts node cache status, link error flags, and clock offsets from the overload saturated path segment set. Each overload saturated path segment in the set includes a path number, overload start point, overload end point, path occupancy timing, queue backlog, node cache status, link error flags, service priority identifier, and clock offset. When extracting the node cache status, the system reads the node cache status field value from the overload saturated path segment. The node cache status field value includes the cache page number and cache page status identifier, which includes valid backup identifiers and invalid backup identifiers. When extracting the link error flags, the system reads the link error flag field value from the overload saturated path segment. The link error flag field value includes an error bitmap, the length of which is the same as the total number of fields in the six-dimensional field. When extracting the clock offset, the system reads the clock offset field value from the overload saturated path segment. The clock offset field value includes the offset direction and offset value. The offset direction includes positive offset and negative offset. Positive offset indicates that the local clock of the target transmission path is ahead of the upstream clock, and negative offset indicates that the local clock of the target transmission path is behind the upstream clock.
[0071] The ability to refill cache pages is determined based on the node cache status. In the node cache status field, when the cache page status identifier is a valid backup identifier, the corresponding cache page number is identified as a refillable cache page. A refillable cache page is a cache page in the node cache management unit that contains a usable backup field.
[0072] Untrusted field bits are determined based on link error flags. In the error bitmap of the link error flag field values, each bit corresponds to one dimension field in the six-dimensional field. The bit order from least significant bit to most significant bit corresponds to the path occupancy timing, queue backlog, node buffer status, link error flag, service priority identifier, and clock offset, respectively. When a bit in the error bitmap is 1, the corresponding dimension field is determined as an untrusted field bit. An untrusted field bit indicates that the current value of the corresponding dimension field in the six-dimensional coupled data block cannot be used for reconstruction and recovery.
[0073] The field alignment offset is determined based on the clock offset. The field alignment offset is the time alignment offset value that needs to be adjusted when writing the backup field to the six-dimensional field. The field alignment offset is determined by the offset direction and offset value in the clock offset field. The field alignment offset is calculated as follows:
[0074]
[0075] Where 's' represents the field alignment offset, and 'Δt' represents the offset value in the clock offset field, in nanoseconds. This indicates the sampling period of the target transmission path, in nanoseconds. This indicates the floor function. The sampling clock configuration of the target transmission path is determined and read from the sampling clock configuration register during system initialization. The value of is an integer greater than 0. When the offset direction is negative, the field alignment offset is the opposite of s; when the offset direction is positive, the field alignment offset is s.
[0076] Write the backfillable cache page, untrusted field bit, and field alignment offset to the refactoring recovery tag set. Each refactoring recovery tag in the refactoring recovery tag set contains the path number, overload start point, overload end point, backfillable cache page, untrusted field bit, and field alignment offset.
[0077] The backup fields are read from the refillable cache pages in the reconstruction and recovery mark set. The cache page address corresponding to the refillable cache page is located in the node cache management unit, and the backup fields are read from the refillable cache page. The backup fields are backup values for missing fields in the six-dimensional fields, including path occupancy timing backup value, queue backlog backup value, node cache status backup value, link error mark backup value, service priority identifier backup value, and clock offset backup value.
[0078] Based on the untrusted field bits in the reconstruction recovery mark set, the corresponding fields in the six-dimensional coupled data block are masked. In the six-dimensional coupled data block, the dimension field values corresponding to the untrusted field bits are set to invalid flags. The invalid flag is a preset invalid value, which is used to indicate that the corresponding field cannot participate in subsequent cross-padding.
[0079] Based on the field alignment offset in the reconstruction and recovery mark set, the backup fields are mapped to the six-dimensional fields to generate cross-completion results. For each dimension field, the backup values of the corresponding dimension field in the backup field are time-aligned according to the field alignment offset. The time-aligned backup fields overwrite the corresponding dimension fields in the six-dimensional coupled data block that have undergone masking, generating cross-completion results. The cross-completion results include the path number, timestamp, path occupancy time sequence, queue backlog, node cache status, link error mark, service priority identifier, and clock offset.
[0080] The cross-completion results are backfilled into the empty positions of the circular storage page according to the path number, generating reconstruction recovery data. Based on the path number in the cross-completion results, the storage area corresponding to the path number in the circular storage page is located. Empty positions are found within the storage area of the circular storage page; these empty positions are locations in the circular storage page where no valid data or invalid markers have been written. The cross-completion results are written to the empty positions. After writing, the chain pointer and slot pointer of the six-dimensional coupled data block in the circular storage page are updated, generating reconstruction recovery data.
[0081] In practice, the port queue of the target transmission path is located based on the path number in the reconstruction and recovery data. The reconstruction and recovery data includes the path number, timestamp, path occupancy timeline, queue backlog, node buffer status, link error flags, service priority identifiers, and clock offset. The port queue of the target transmission path is maintained by the port queue management unit, which stores multiple port queues, each corresponding to a path number. The path number is extracted from the reconstruction and recovery data, and the port queue in the port queue management unit that matches the path number in the reconstruction and recovery data is searched. This searched port queue is then used as the port queue of the target transmission path.
[0082] The reconstructed and recovered data is segmented into data slices corresponding to business priority identifiers. The business priority identifiers are extracted from the reconstructed and recovered data; each identifier contains multiple priority level values. The reconstructed and recovered data is then segmented according to the priority level values in the business priority identifiers, with data slices having the same priority level value grouped together. Each data slice contains a priority level value, slice content, and slice length. The slice content is a set of fields in the reconstructed and recovered data that belong to the same priority level value.
[0083] The clock offset is used to read the offset direction between the local clock of the target transmission path and the upstream clock. The clock offset is stored in the reconstruction and recovery data. It includes the offset direction and the offset value. The offset direction can be positive or negative. A positive offset indicates that the local clock of the target transmission path is ahead of the upstream clock, while a negative offset indicates that the local clock of the target transmission path is behind the upstream clock. The offset direction in the clock offset is read to determine whether it is a positive or negative offset.
[0084] A transmission sequence number is assigned to each data segment based on the offset direction. When the offset direction is positive, the rule for assigning a transmission sequence number to each data segment is as follows: Data segments are arranged in descending order of priority level, and an incremental transmission sequence number is assigned to each data segment, starting from 1 with a step size of 1. When the offset direction is negative, the rule for assigning a transmission sequence number to each data segment is as follows: Data segments are arranged in ascending order of priority level, and an incremental transmission sequence number is assigned to each data segment, starting from 1 with a step size of 1.
[0085] Data slices are sorted by transmission sequence number and service priority identifier according to their transmission time. The transmission time of each data slice is calculated using the following formula:
[0086]
[0087] in, Indicates the transmission time of the data segment, in nanoseconds; The base time for sorting transmission times is indicated by nanoseconds. Retrieve the current scheduling time of the port queue of the target transmission path and read it from the scheduling table of the target transmission path; Indicates the sequence number of the data segment being transmitted. The sequence number to be sent for each data slice based on the offset direction. The value of is an integer greater than or equal to 1; This indicates the length of the transmission time slot, in nanoseconds. Configured by the port queue management unit of the target transmission path, and read from the time slot configuration register of the port queue management unit during system initialization. The value of is an integer greater than 0; This indicates the priority level value in the service priority identifier corresponding to the data slice. The value range is the priority level value range defined by the service scheduling unit of the target transmission path; Indicates priority weight. The priority ranking weights set for the service scheduling units of the target transmission path The value of is an integer greater than or equal to 0. The priority weight configuration register of the service scheduling unit is read during system initialization; The priority time weight is indicated in nanoseconds. This is based on the calculated... Sort the data slices from smallest to largest.
[0088] The sorted data slices are written to the port queue. Following the sorted order, each data slice is written sequentially to the port queue of the target transmission path. During writing to the port queue, the data slice content is written to the queue buffer unit of the port queue, the priority level value of the data slice is written to the priority flag storage unit of the port queue, and the transmission time of the data slice is written to the transmission time storage unit of the port queue.
[0089] Read the occupancy threshold index of the port queue from the scheduling table of the target transmission path. The scheduling table of the target transmission path stores scheduling information for multiple port queues, including occupancy threshold index, credit count, and link permission bit. The occupancy threshold index is the index number of the port queue's occupancy threshold level in the scheduling table. Based on the position of the port queue in the scheduling table of the target transmission path, read the occupancy threshold index of the port queue.
[0090] Associate the recovery completion flag with the occupancy threshold index. Generate a recovery completion flag, which includes a recovery completion identifier and a recovery completion timestamp. Write the recovery completion flag into the scheduling table of the target transmission path at the scheduling information storage location corresponding to the read occupancy threshold index, so that the recovery completion flag and the occupancy threshold index are associated in the same scheduling information storage location.
[0091] After receiving the recovery completion flag, the port queue updates its credit count and link permission bit. Upon receiving the recovery completion flag, the port queue reads the corresponding credit count and link permission bit from the target transmission path's scheduling table. Based on the recovery completion flag, the credit count is incremented by the recovery credit increment value. This recovery credit increment value is configured by the target transmission path's scheduling table and is read from the recovery credit configuration register during system initialization. The value of the recovery credit increment value is an integer greater than 0. Based on the recovery completion flag, the link permission bit is set to the permission valid state, indicating that the target transmission path's port queue is allowed to transmit via the link. After updating the port queue's credit count and link permission bit, the path status is updated to the recovery completion state.
[0092] This invention provides a six-dimensional coupled data reconstruction and recovery system for overload saturation pathways, comprising a memory, a processor, and a computer program stored in the memory and running on the processor. The memory stores an overload saturation pathway data set, a six-dimensional coupled data block, an overload saturation pathway segment set, a reconstruction and recovery marker set, and reconstruction and recovery data. The processor executes instructions from the computer program.
[0093] When the processor executes the computer program, it acquires multi-source overload records of the target transmission path, forming an overload saturated path data set. The processor reads the path occupancy timing from the port occupancy monitoring unit of the target transmission path, the queue backlog from the queue management unit, the node cache status from the node cache management unit, the link error flag from the link monitoring unit, the service priority identifier from the service scheduling unit, and the clock offset from the clock synchronization unit. The processor writes the read path occupancy timing, queue backlog, node cache status, link error flag, service priority identifier, and clock offset into the overload saturated path data set according to the path number and timestamp. The overload saturated path data set is stored in memory.
[0094] The processor establishes a six-dimensional coupled index based on the path number and timestamp, writes the six-dimensional fields into a circular storage page, and generates a six-dimensional coupled data block. The processor extracts the path number and timestamp from the overload-saturated path data set to generate a path event identifier. The processor binds the path event identifier to the path occupancy timing, queue backlog, node cache status, link error flag, service priority identifier, and clock offset from the overload-saturated path data set, obtaining six-dimensional field slots. The processor sets a chained pointer in the circular storage page according to the six-dimensional field slots, generating a six-dimensional coupled data block. The circular storage page is located in memory and consists of multiple fixed-length storage pages, managed using a circular write method.
[0095] In practice, when the processor writes a six-dimensional field to a circular memory page, it uses the circular memory page write address calculation formula to determine the write address. The processor then writes the six-dimensional field to the corresponding offset address within the circular memory page based on this write address. The circular memory page write address calculation formula is as follows:
[0096]
[0097] in, This indicates the write address of the six-dimensional field within the circular memory page. The unit is bytes; This indicates the starting address of the circular memory page. It is determined by the physical start address of the circular memory page in the memory. The unit is bytes; The numerical hash value representing the path event identifier. It is obtained by hashing the string that identifies the pathway event. It is an integer. The value of is an integer greater than or equal to 0; This indicates the number of fixed-length memory pages contained within a circular memory page. Set to 256, The setting is based on the fact that the maximum number of concurrent paths in the target transmission path does not exceed 256; This indicates the number of bytes occupied by each six-dimensional field slot in a circular storage page. The storage length is determined by the fixed length of each dimension field in the six-dimensional field. The unit is bytes; mod represents the modulo operation.
[0098] The processor segments the six-dimensional coupled data block into saturation intervals, generating an overload saturated path segment set. The processor reads the path occupancy timing and queue backlog from the six-dimensional coupled data block, scanning the path occupancy timing in timestamp order. When the path occupancy timing is continuously at full capacity and the queue backlog remains non-decreasing during the corresponding time period, the processor marks the overload start and end points. Based on the overload start and end points, the processor extracts the path occupancy timing, queue backlog, node buffer status, link error flags, service priority identifiers, and clock offsets from the six-dimensional coupled data block to generate the overload saturated path segment set, which is then stored in memory.
[0099] The processor performs recoverability marking on the overload-saturated path segment set according to the mutual constraints of the six-dimensional fields, generating a reconstruction recovery mark set. The processor extracts node cache status, link error flags, and clock offsets from the overload-saturated path segment set. Based on the node cache status, it determines the refillable cache pages; based on the link error flags, it determines the untrusted field bits; and based on the clock offset, it determines the field alignment offset. The processor writes the refillable cache pages, untrusted field bits, and field alignment offsets into the reconstruction recovery mark set, which is stored in memory.
[0100] The processor performs cross-padding and backfilling on missing fields of the six-dimensional coupled data block based on the reconstruction and recovery mark set, generating reconstruction and recovery data. The processor reads backup fields from the node cache management unit based on the backfillable cache pages in the reconstruction and recovery mark set, masks the corresponding fields in the six-dimensional coupled data block based on the untrusted field bits in the reconstruction and recovery mark set, and maps the backup fields to the six-dimensional fields based on the field alignment offsets in the reconstruction and recovery mark set, generating cross-padding results. The processor backfills the cross-padding results to the empty positions of the circular storage pages according to the path numbers, generating reconstruction and recovery data, which is stored in memory.
[0101] The processor writes the reconstructed recovery data back to the target transmission path according to the path number, updating the path status. Based on the path number in the reconstructed recovery data, the processor locates the port queue of the target transmission path, segments the reconstructed recovery data into data fragments corresponding to service priority identifiers, sorts the data fragments by transmission time according to clock offset, and writes the sorted data fragments into the port queue. The processor writes a recovery completion flag to the scheduling table of the target transmission path, updates the credit count and link permission bit of the port queue, and updates the path status to recovery complete.
[0102] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.
Claims
1. A method for reconstructing and recovering six-dimensional coupled data from an overloaded saturated pathway, characterized in that, The method includes: Acquire multi-source overload records of the target transmission path to form an overload saturated path data set. The overload saturated path data set includes path number, timestamp, and six-dimensional fields. The six-dimensional fields are path occupancy time sequence, queue backlog, node buffer status, link error mark, service priority identifier, and clock offset. A six-dimensional coupled index is established based on the path number and the timestamp, and the six-dimensional field is written into a circular storage page to generate a six-dimensional coupled data block; The six-dimensional coupled data block is segmented into saturation intervals to generate a set of overloaded saturated path segments; The set of overload saturated pathway segments is marked for recoverability according to the mutual constraint relationship of the six-dimensional fields, and a set of reconstruction recovery marks is generated. Based on the reconstructed recovery marker set, cross-filling and backfilling are performed on the missing fields of the six-dimensional coupled data block to generate reconstructed recovery data; The reconstructed and restored data is written back to the target transmission path according to the path number, and the path status is updated.
2. The method for reconstructing and recovering overloaded saturated pathway six-dimensional coupled data according to claim 1, characterized in that, The process of establishing a six-dimensional coupled index based on the path number and the timestamp, writing the six-dimensional fields into a circular storage page, and generating a six-dimensional coupled data block includes: Extract the path number and timestamp from the overload saturation path data set to generate a path event identifier; The path event identifier is bound to the path occupancy time sequence, queue backlog, node cache status, link error mark, service priority identifier and clock offset in the overload saturated path data set, respectively, to obtain six-dimensional field slots; The six-dimensional coupled data block is generated by setting chained pointers in the circular storage page according to the field slots of the six dimensions.
3. The method for reconstructing and recovering overloaded saturated pathway six-dimensional coupled data according to claim 1, characterized in that, The step of segmenting the six-dimensional coupled data block into saturated intervals to generate an overloaded saturated path fragment set includes: Read the path occupancy timing and queue backlog in the six-dimensional coupled data block; When the path occupancy time sequence is continuously full and the queue backlog remains non-decreasing during the corresponding time period, mark the overload start point and overload end point. The six-dimensional fields in the six-dimensional coupled data block are extracted according to the overload start point and the overload end point to generate the overload saturation path segment set.
4. The method for reconstructing and recovering overloaded saturated pathway six-dimensional coupled data according to claim 1, characterized in that, The process of marking the overloaded saturated pathway fragment set for recoverability according to the mutual constraint relationship of the six-dimensional fields, and generating a reconstructed recovery mark set, includes: Extract node buffer status, link error flags, and clock offsets from the overload saturated path segment set; Based on the node cache status, determine the cache pages that can be refilled; based on the link error flag, determine the untrusted field bits; based on the clock offset, determine the field alignment offset. Write the backfillable cache page, the untrusted field bit, and the field alignment offset into the reconstructed recovery tag set.
5. The method for reconstructing and recovering overloaded saturated pathway six-dimensional coupled data according to claim 4, characterized in that, The process of performing cross-filling and backfilling on the missing fields of the six-dimensional coupled data block based on the reconstructed recovery marker set to generate reconstructed recovery data includes: Read the backup fields from the refillable cache pages in the reconstructed recovery mark set; The corresponding fields in the six-dimensional coupled data block are masked according to the untrusted field bits in the reconstructed recovery mark set; Based on the field alignment offset in the reconstructed recovery mark set, the backup field is mapped to the six-dimensional field to generate a cross-completion result; The cross-completion results are backfilled into the empty positions of the circular storage page according to the path number to generate the reconstructed recovery data.
6. The method for reconstructing and recovering overloaded saturated pathway six-dimensional coupled data according to claim 1, characterized in that, The step of writing the reconstructed and restored data back to the target transmission path according to the path number and updating the path status includes: The port queue of the target transmission path is located based on the path number in the reconstructed and restored data; The reconstructed and restored data is divided into data slices corresponding to the business priority identifier; The data slices are sorted by transmission time according to the clock offset; The sorted data slices are written to the port queue, and a recovery completion flag is written to the scheduling table of the target transmission path to update the path status.
7. The method for reconstructing and recovering overloaded saturated pathway six-dimensional coupled data according to claim 1, characterized in that, The acquisition of multi-source overload records of the target transmission path to form an overload saturated path data set includes: Collect the path occupancy timing, queue backlog, node buffer status, link error markers, service priority identifiers, and clock offset of the target transmission path; Write the above six data items into the six-dimensional field according to the channel number and timestamp; For the six-dimensional fields with repeated timestamps under the same path number, a field version number is generated, and the six-dimensional fields with the field version number are aggregated into the overload saturation path data set.
8. The method for reconstructing and recovering overloaded saturated pathway six-dimensional coupled data according to claim 6, characterized in that, The step of sorting the data slices by transmission time according to the clock offset includes: The offset direction between the local clock and the upstream clock of the target transmission path is read from the clock offset. A transmission sequence number is assigned to each data segment according to the offset direction; The data slices are sorted by transmission time according to the transmission sequence number and the service priority identifier.
9. The method for reconstructing and recovering overloaded saturated pathway six-dimensional coupled data according to claim 6, characterized in that, The step of writing a recovery completion flag to the scheduling table of the target transmission path and updating the path status includes: Read the occupancy threshold index of the port queue from the scheduling table of the target transmission path; Associate the recovery completion marker with the occupancy threshold index; After the port queue receives the recovery completion flag, it updates the credit count and link permission bit of the port queue.
10. A six-dimensional coupled data reconstruction and recovery system for overload saturation pathways, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the overload saturation path six-dimensional coupled data reconstruction and recovery method according to any one of claims 1 to 9.
