A monitoring operating system for a data disaster backup control system

Abstract

The application relates to the technical field of data disaster backup control, and discloses a monitoring operating system for a data disaster backup control system, which comprises the following steps: recording the creation time and the final update time of a paragraph in a document, and classifying the paragraph based on a set word threshold and a time threshold; extracting the binary of the first and last characters of each sentence of the marked paragraph, calculating the XOR value, calculating the XOR value of the XOR value of each sentence and the creation time of the marked paragraph, generating a binary set of the marked paragraph, and generating a hash value of the marked paragraph based on the binary set; judging whether the hash value of the marked paragraph exists; when it is detected that the hash value of the marked paragraph does not exist, it is determined that the marked paragraph is deleted; slicing the marked paragraph into multiple slice sentences, calculating the slice identifier of the slice sentence according to the first and last characters of the slice sentence and the creation time of the marked paragraph, and matching the backup marked paragraph by using the slice identifier; and improving the intelligence, refinement and space utilization rate of the document backup.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data disaster recovery backup control, and particularly relates to a monitoring operating system for a data disaster recovery backup control system. BACKGROUND

[0002] In a data processing and document editing system, regular backup can significantly reduce the loss caused by sudden data loss. A reasonable backup strategy not only improves the disaster recovery capability of the system, but also quickly recovers the required information after data deletion or change, and is a key means to ensure data integrity and continuity.

[0003] The current mainstream document backup mechanism mostly adopts file-level saving, full-text difference comparison or periodic snapshot. Although these methods have certain effects in ensuring the integrity of the overall content, they have obvious deficiencies in processing paragraph-level deletion behavior. First, when a user permanently deletes a paragraph from a document, the traditional system usually cannot determine whether the paragraph should continue to be retained in the historical backup document, resulting in redundant retention of the deleted paragraph in multiple backup versions, causing waste of storage space. Secondly, the existing scheme lacks a structured tracking means for paragraph-level deletion behavior, and cannot establish the inheritance relationship of the paragraph between different versions, so it cannot effectively identify whether the paragraph has been replaced or evolved. Thirdly, the existing system mostly relies on content full-text comparison for change detection, which is difficult to quickly locate and identify the paragraph, increases the system computing overhead, and also causes repeated backup of repeated content, affecting the overall efficiency. In addition, there is a lack of a unified identification mechanism, which cannot match, track and deduplicate the paragraph between massive versions at low cost, thereby limiting the fine-grained ability of the system in space utilization and version management.

[0004] The present application provides a monitoring operating system for a data disaster recovery backup control system. SUMMARY

[0005] The present application provides a monitoring operating system for a data disaster recovery backup control system, which is used to promote the solution to the problems mentioned in the background.

[0006] The present application provides the following technical scheme: a monitoring operating system for a data disaster recovery backup control system, comprising:

[0007] A paragraph management module is configured to record the creation time and the final update time of each paragraph in a document, and classify the paragraphs based on a set word threshold and a time threshold, wherein the paragraphs optimized for backup are marked as marked paragraphs.

[0008] The paragraph identification generation module is configured to extract the binary of the first and last characters of each sentence of the marked paragraph, calculate the XOR value, calculate the XOR value between the XOR value of each sentence and the time when the marked paragraph is created, generate the binary set of the marked paragraph, and generate the hash value of the marked paragraph based on the binary set.

[0009] The paragraph deletion detection module is configured to determine whether the hash value of the marked paragraph exists, and when it is detected that the hash value of the marked paragraph does not exist, it is determined that the marked paragraph is deleted, and the marked paragraph is backed up and optimized.

[0010] The deleted paragraph tracking module is configured to slice the marked paragraph into multiple slice sentences, calculate the slice identifier of the slice sentence according to the first and last characters of the slice sentence and the time when the marked paragraph is created, and match the backup marked paragraph using the slice identifier.

[0011] The deleted paragraph optimal saving module is configured to determine the optimal way to save the marked paragraph according to the number of words of the backup marked paragraph and the deleted marked paragraph.

[0012] Optionally, the paragraph management module is configured to record the creation time and the final update time of each paragraph in the document, and classify the paragraphs based on a set word threshold and a time threshold, including:

[0013] The automatic saving interval is set, and the automatic saving interval is used to automatically save the document in units of the automatic saving interval when the user creates the document, and a backup document of the document is generated.

[0014] The word threshold is set.

[0015] The time threshold is set, and the time threshold is used to determine whether the paragraph is completed.

[0016] For any one paragraph in the document:

[0017] The number of words of the paragraph is counted, and the number of words is compared with the word threshold:

[0018] If the number of words is greater than or equal to the word threshold, the time interval between the final update time of the paragraph and the current time is calculated, and the time interval is compared with the time threshold:

[0019] If the time interval is greater than or equal to the time threshold, the paragraph is recorded as a marked paragraph for backup optimization.

[0020] Optionally, the paragraph identification generation module is configured to extract the binary of the first and last characters of each sentence of the marked paragraph, calculate the XOR value, calculate the XOR value between the XOR value of each sentence and the time when the marked paragraph is created, generate the binary set of the marked paragraph, and generate the hash value of the marked paragraph based on the binary set, including:

[0021] For any one paragraph:

[0022] Divide the paragraph into multiple sentences according to the period Wherein, is the i-th sentence, n is the number of sentences, :

[0023] Obtain the first character of the sentence ; Calculate the binary of the first character Wherein, is the binary of the first character, is used to calculate the binary of the character;

[0024] Obtain the last character of the sentence ; ;

[0025] Calculate the binary of the last character Wherein, is the binary of the last character;

[0026] Calculate the XOR operation of and : Wherein, is the XOR symbol, is the basic XOR value of the sentence ;

[0027] The creation time of the sentence is represented as a timestamp ;

[0028] Calculate the XOR operation of the timestamp and the basic XOR value: Wherein, is the sentence identifier;

[0029] For any marked paragraph:

[0030] Obtain the sentence identifier of all sentences of the marked paragraph, and obtain a binary set ;

[0031] Apply a hash function to the binary set to calculate the hash value of the marked paragraph , , is the hash function.

[0032] Optionally, the paragraph deletion detection module is configured to determine whether the hash value of the marked paragraph exists, and when it is detected that the hash value of the marked paragraph does not exist, it is determined that the marked paragraph is deleted, and the marked paragraph is backed up and optimized, including:

[0033] Arrange the backup documents according to the saving order;

[0034] counting hash values of all marked paragraphs in each backup document;

[0035] setting any one backup document as a target document;

[0036] setting a verification number;

[0037] obtaining a previous backup document of the target document as a forward reference document according to a saving order;

[0038] obtaining a set of backup documents after the target document as a backward reference document set according to the saving order, the set of backup documents including the target document and the verification number of backup documents after the target document;

[0039] comparing hash values of marked paragraphs in the forward reference document, the target document and the backward reference document set;

[0040] if a hash value of a marked paragraph existing in the target document exists in the forward reference document but does not exist in all the backward reference document set, it is determined that the marked paragraph is deleted in the target document;

[0041] setting the marked paragraph deleted in the target document as a target paragraph.

[0042] Optionally, the deletion paragraph tracking module is configured to slice the marked paragraph into a plurality of sliced sentences, including:

[0043] setting a sliced length, the sliced length being used to limit a number of characters of each sliced sentence;

[0044] setting a sliced overlap length, the sliced overlap length being used to limit a number of overlapping characters between adjacent two sliced sentences;

[0045] obtaining a first character of the target paragraph as a positioning first character;

[0046] S1, selecting a sliced length of characters from the positioning first character according to a character arrangement order of the target paragraph as a positioning last character;

[0047] S2, judging whether the positioning last character is empty or not;

[0048] if the positioning last character is not empty, setting the positioning first character and the positioning last character and a character sequence between the positioning first character and the positioning last character as a sliced sentence;

[0049] S3, calculating the sliced length minus the sliced overlap length plus 1, and a result is a neighboring length;

[0050] selecting a neighboring length of characters from the positioning first character according to the character arrangement order of the target paragraph as a new positioning first character, and repeating S1-S3;

[0051] S4, if the positioning last character is empty:

[0052] Obtain the last character of the target paragraph, and combine the first character of the target paragraph with the last character and the character sequence between the first character of the target paragraph and the last character to form a segmented sentence.

[0053] Optionally, the step of calculating the slice identifier of the sliced ​​sentence based on the first and last characters of the sliced ​​sentence and the creation time of the marked paragraph, and using the slice identifier to match the backed-up marked paragraph, includes:

[0054] Calculate the segment identifier for each segment clause, specifically including:

[0055] Get the binary representation of the first and last characters of the slice clause, calculate the XOR value of the binary representation of the first and last characters, and denot it as the slice XOR value.

[0056] Obtain the creation time of the target paragraph and represent the creation time as a timestamp;

[0057] Calculate the XOR value of the timestamp and the slice XOR value, and denote it as the slice identifier;

[0058] Use slice identifiers to match the marked segments in the backup, specifically including:

[0059] Get the save time of all backup documents and represent the save time as a timestamp;

[0060] Compare the timestamp of the target paragraph creation time with the timestamp of each backup document saving time, and obtain the backup file corresponding to the largest timestamp earlier than the target paragraph's timestamp as the backup document for the first time the target paragraph is saved;

[0061] Record the first backup document containing the target paragraph as the starting backup document;

[0062] Retrieve all backup documents between the starting backup document and the target document according to the order in which they were saved, and combine them with the starting backup document and the target document to form a paragraph-related document set.

[0063] Optionally, the step of calculating the slice identifier of the sliced ​​sentence based on the first and last characters of the sliced ​​sentence and the creation time of the marked paragraph, and using the slice identifier to match the backed-up marked paragraph, further includes:

[0064] For any backup document in the paragraph-related document set:

[0065] Set a time error threshold;

[0066] The time before the creation time of each paragraph in the backup document is recorded as the lower creation limit time, and the time after the creation time is recorded as the upper creation limit time.

[0067] The creation time interval is composed of the upper limit creation time and the lower limit creation time.

[0068] comparing the creation time of the target paragraph with the creation time interval of each paragraph in the backup document;

[0069] obtaining the creation time interval in which the creation time of the target paragraph is located, and recording the paragraph corresponding to the creation time interval as a candidate paragraph;

[0070] comparing the sentence identifiers of the candidate paragraph with all the slice identifiers of the target paragraph;

[0071] setting an equal judgment threshold;

[0072] when the number of the sentence identifiers equal to the slice identifiers in the candidate paragraph is greater than or equal to the equal judgment threshold, the candidate paragraph is determined as the evolution paragraph of the target paragraph.

[0073] Optionally, the deleting paragraph optimal saving module is used for judging the optimal way of saving the marked paragraph according to the word number of the backup marked paragraph and the deleted marked paragraph, and the optimal way of saving the marked paragraph comprises:

[0074] obtaining the word number of all the backup document evolution paragraphs in the paragraph associated document set, calculating the total sum and recording the total sum as evolution total word number;

[0075] obtaining the word number of the target paragraph, and comparing the word number with the evolution total word number;

[0076] if the word number is less than (the evolution total word number / 2), calculating the time length from the saving time of the starting backup document to the current time, and recording the time length as a first time length;

[0077] obtaining the number of times that the starting backup document is queried, and recording the number of times as a first number of times;

[0078] the time length from the saving time of the target document to the current time is a second time length;

[0079] obtaining the number of times that the target document is queried, and recording the number of times as a second number of times;

[0080] calculating the first number of times divided by the first time length, and recording the result as a first query frequency;

[0081] calculating the second number of times divided by the second time length, and recording the result as a second query frequency;

[0082] if the first query frequency is greater than the second query frequency, the evolution paragraph in the starting backup document is replaced by the target paragraph, and all the remaining evolution paragraphs in the paragraph associated document set are deleted;

[0083] if the first query frequency is less than or equal to the second query frequency, the target paragraph in the target document is retained, and all the remaining evolution paragraphs in the paragraph associated document set are deleted.

[0084] The application has the following beneficial effects:

[0085] 1. The monitoring operating system for the data disaster backup control system, by setting a time threshold, accurately identifying which paragraphs have editing completion characteristics, thereby dynamically managing the hierarchical management of document content, the system stores each automatic backup file, but for the marked paragraphs that have been edited and identified as no longer modified, if the marked paragraphs are determined to be permanently deleted, and the number of words in such paragraphs exceeds the word threshold, the memory cost of storing such marked paragraphs in the previous backup file is very large, so backup optimization should be performed for such whole paragraph deletion marked paragraphs to save memory resources.

[0086] 2. The monitoring operating system for the data disaster backup control system, by means of the binary XOR of the first and last characters of the sentence and the participation of the creation time, a unique and lightweight paragraph binary identifier is constructed. Since it is already identified that the marked paragraph will not change, the hash value of the marked paragraph will not change, avoiding the complexity of retrieving all the text in the whole paragraph deletion in the prior art, and the creation time is accurate to milliseconds, plus the calculation of the binary of the first and last of each sentence in each paragraph can uniquely identify the paragraph. This is equivalent to giving a fixed and unchanging paragraph a fixed and unchanging hash value as an index. In the case of large backup volume and frequent version, the retrieval and comparison efficiency is greatly improved, especially in the case of judging the whole paragraph deletion marked paragraph, it has good accuracy and security.

[0087] 3. The monitoring operating system for the data disaster backup control system, first, the definition of the marked paragraph determines that the marked paragraph is stable and will not be frequently modified, and all backup files where the marked paragraph is located are already completely determined, so it can be explained that the marked paragraph may appear completely in multiple backup documents. If the user deletes the marked paragraph, there must be a hash value of the marked paragraph existing in the target document in the forward reference document, but not existing in all backward reference document sets, that is, the user deletes the target paragraph after saving the target document. The benefit of optimizing only the whole paragraph deletion case is that optimization itself needs to consider time cost, cpu running resource consumption, etc. If every sentence or character modification is optimized in every backup file, the resource cost of optimization may be much greater than the benefit brought by optimization. In combination with multiple version comparison strategies, it avoids the accidental omission or editing reconstruction of the user from mistakenly deleting the paragraph, lays the foundation for subsequent paragraph optimization saving and version reorganization, and realizes the mechanism from discovering deletion to responding to optimization.

[0088] 4、The monitoring operating system for the data disaster backup control system, since multiple backup documents are continuously saved during the process of creating the target paragraph, the target paragraph stored in the backup document is incomplete, the target paragraph is cut according to the character number of the slice length limit, and multiple possible sentence breaking conditions are limited according to the slice overlap length, the continuity between the sentences is maximally maintained, and the multiple sentences that may appear in the evolution process of the target paragraph are provided, which facilitates the matching of the position of the target paragraph in the backup document and improves the accuracy of finding the paragraph in the evolution process of the target paragraph in the backup document.

[0089] 5、The monitoring operating system for the data disaster backup control system, the backup document in which the target paragraph is first saved is the first step of optimization backup, which can narrow the range of the backup document to be screened, and obtain the paragraph associated document set, the second step of screening is to find the predecessor of the target paragraph corresponding to each backup document in the paragraph associated document set, here, the creation time interval of the paragraph is first used for the first screening, according to the set time error threshold, the fuzzy screening caused by concurrent editing or system time precision is avoided, the slice identifier of the target paragraph and the sentence identifier of the sentence in the paragraph are compared, which is actually the comparison of the content in the paragraph, without using the traditional word-by-word comparison, but based on the identifier comparison of multiple character binary calculations, which can improve the comprehensiveness, accuracy and efficiency of the comparison.

[0090] 6、The monitoring operating system for the data disaster backup control system, when the number of words is less than (evolution total word number / 2), it means that the backup file is more efficient for optimization replacement, here, the cpu running consumption and time cost are simplified and equivalent to the cost of storing evolution total word number / 2, at this time, the memory is simplified to store only one complete version of the target paragraph, generally, the complete target paragraph is stored in the first backup file or the last backup file, but the user is more used to using which backup file, so the two efficiencies are calculated, the backup document corresponding to the highest query efficiency is used as the document for storing the complete target paragraph, which is convenient for the user to view, by retaining only the target paragraph or updating the evolution paragraph, and eliminating redundant content, the system realizes the triple optimization of storage space, access efficiency and data accuracy.

[0091] 7、The monitoring operating system for the data disaster recovery backup control system intelligently marks and classifies backup optimization paragraphs, accurately identifies paragraph deletion and evolution, and avoids redundant data repeated backup. By dynamically judging the backup paragraph word count and access frequency, the optimal saving strategy is determined, the storage space occupation is significantly reduced, the storage resource utilization is improved, the memory and storage space of the backup system are saved, more effective versions can be retained by the disaster recovery system, data can be quickly recovered in the case of hardware failure, system exception or human error operation, the data loss risk is minimized. At the same time, reducing redundant data also reduces storage maintenance cost and energy consumption, improves system economy and environmental protection. The safety, reliability and efficiency of data disaster recovery backup are effectively improved. BRIEF DESCRIPTION OF DRAWINGS

[0092] Figure 1 The system module schematic diagram of the present application. DETAILED DESCRIPTION

[0093] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0094] Embodiment one, refer to Figure 1 A monitoring operating system for a data disaster recovery backup control system, comprising:

[0095] A paragraph management module for recording the creation time and the final update time of each paragraph in a document, classifying the paragraphs based on a set word threshold and a time threshold, wherein the paragraphs optimized for backup are marked as marked paragraphs, comprising:

[0096] The automatic saving interval is set to 5 minutes, which is used to automatically save the document in units of automatic saving interval when the user creates the document, and a backup document of the document is generated;

[0097] The word threshold is set to 40 words;

[0098] The time threshold is set to 10 minutes, which is used to determine whether the paragraph is completed;

[0099] For a paragraph in a document:

[0100] The word count of the paragraph is 44 words, and the word count and the word threshold are compared:

[0101] 44 words are greater than 40 words, the time interval between the final update time of the paragraph and the current time is calculated as 15 minutes, and the time interval and the time threshold are compared:

[0102] 15 minutes are greater than 10 minutes, and the paragraph is marked as a marked paragraph for backup optimization.

[0103] For determining which paragraphs have been completed and are worth being included in the backup optimization process during the document editing process. This determination is based on two core dimensions:

[0104] Word threshold: the paragraph must have sufficient length to avoid frequent saving of temporary input or fragmented text;

[0105] Time threshold: the paragraph must not be modified within a certain time, indicating that the editing is tending to be completed;

[0106] The paragraph identification generation module is used to extract the binary of the first and last characters of each sentence of the marked paragraph and calculate the XOR value, calculate the XOR value of each sentence and the creation time of the marked paragraph, generate a binary set of the marked paragraph, and generate a hash value of the marked paragraph based on the binary set, including:

[0107] For any one paragraph, the paragraph in the embodiment is "The paragraph identification is unique information for distinguishing different paragraphs. It should avoid repetition as much as possible to enhance the recognition accuracy after content change.";

[0108] The paragraph is divided into two sub-sentences according to the period, n=2:

[0109] The first sentence is: "The paragraph identification is unique information for distinguishing different paragraphs.";

[0110] The second sentence is: "It should avoid repetition as much as possible to enhance the recognition accuracy after content change."

[0111] For the first sentence :

[0112] First character: segment→ Unicode encoding = 6BB5 (hexadecimal) = 27573 (decimal);

[0113] Last character: sense→ Unicode encoding = 606F = 24687;

[0114] Since the period will increase the repeatability between single sub-sentences if used as the tail character of the sub-sentence, in the complete marked paragraph, the tail character before the period of each sub-sentence is used to calculate the sub-sentence identification.

[0115] Base XOR value: 27573 XOR 24687 = 5186, XOR is XOR operation;

[0116] For the second sentence (S2):

[0117] First character: it → Unicode = 5B83 = 23427;

[0118] Last character: degree → Unicode = 5EA6 = 24230;

[0119] Base XOR value: 23427 XOR 24230 = 1429;

[0120] Get paragraph creation timestamp: 2025 / 6 / 18 15:00:00 → Unix timestamp T = 1755548400;

[0121] Calculate the sub-sentence identification of each sub-sentence as:

[0122] Sub-sentence 1: Sub-sentence identification = 5186 XOR 1755548400 = 1755551650;

[0123] Sub-sentence 2: Sub-sentence identification = 1429 XOR 1755548400 = 1755549693;

[0124] Get the binary identification set of the paragraph: {1755551650, 1755549693};

[0125] In the present application, the binary identification set is calculated for each paragraph in the backup document, but only the hash value is calculated for the marked paragraph, because the marked paragraph is defined as a completed paragraph and will not be easily modified, so the hash value is fixed and used to uniquely identify the paragraph;

[0126] Using the hash value to uniquely identify the paragraph avoids the increase of user's wrong order or the influence of adding or deleting paragraphs on the order of paragraphs, that is, ignoring the influence of paragraph order, simplifying the operation process.

[0127] Using the existing technology hash function, the hash value of the paragraph is generated from the sub-sentence identification set: assuming that a common hash method (such as SHA-256, MD5 or CRC32) is used:

[0128] For example, 1755551650-1755549693 is concatenated into a string, and then hashed:

[0129] Hash("1755551650-1755549693") → example CRC32 result = `A2F61CDE`;

[0130] Even if the content of two paragraphs is only changed by one sentence (such as changing a word at the end of a sentence), the first and last characters will be different, the final sentence identifier will be different, and the hash value will be completely different. By adding a timestamp and participating in XOR, repeated paragraphs created at different times can also be distinguished to prevent identifier conflicts caused by repeated content. The hash value can be used in multiple modules such as version comparison, change detection, deletion judgment, etc. to reduce the computational burden of full paragraph comparison.

[0131] The paragraph deletion detection module is used to determine whether the hash value of the marked paragraph exists. When it is detected that the hash value of the marked paragraph does not exist, it is determined that the marked paragraph is deleted, and the marked paragraph is backed up and optimized, including:

[0132] Suppose the current paragraph P1 is a marked paragraph that exists in 5 backup documents (in order of saving):

[0133] Paragraph P1 is "Paragraph identifier is the unique information used to distinguish different paragraphs. It should be avoided as much as possible to enhance the recognition accuracy after the content is changed.";

[0134] The version number in the version version number saves the record after paragraph P1 is identified as a marked paragraph;

[0135] The following times are for 2025 / 6 / 18:

[0136] Version number V3 save time 15:10:00 contains paragraph hash value A2F61CDE (paragraph P1);

[0137] Version number V4 save time 15:15:00 contains paragraph hash value A2F61CDE (paragraph P1);

[0138] Version number V5 save time 15:20:00 contains paragraph hash value A2F61CDE (paragraph P1);

[0139] Version number V6 save time 15:25:00 contains paragraph hash value no (paragraph P1 is deleted);

[0140] Version number V7 save time 15:30:00 contains paragraph hash value no (paragraph P1 does not appear);

[0141] Arrange the backup documents in order of saving;

[0142] Count the hash values of all marked paragraphs in each backup document;

[0143] Record any one backup document as the target document as;

[0144] Set the verification number to 2;

[0145] The last backup document of the target document is obtained according to the storage order and is recorded as a forward reference document V4;

[0146] The backward reference document set is obtained by acquiring a number of backup documents after the target document according to the storage order, and is recorded as V6 and V7;

[0147] The hash values of the marked paragraphs in the forward reference document, the target document, and the backward reference document set are compared;

[0148] If the hash value of the marked paragraph existing in the target document exists in the forward reference document but does not exist in all the backward reference document sets, it is determined that the marked paragraph is deleted in the target document;

[0149] The marked paragraph deleted in the target document V5 is recorded as a target paragraph.

[0150] The deletion paragraph tracking module is configured to slice the marked paragraph into a plurality of slice sentences, calculate a slice identifier of the slice sentence according to a first character and a last character of the slice sentence and a creation time of the marked paragraph, and match the slice identifier with a backup marked paragraph, and the deletion paragraph tracking module comprises:

[0151] Paragraph P1, creation time: 2025 / 06 / 18 15:00:00, creation timestamp (T): 1755548400, total number of characters: 44;

[0152] The following is the character and position sequence number:

[0153]

[00] paragraph

[01] fall

[02] mark

[03] identification

[04] is

[05] used

[06] to

[07] distinguish

[08] different

[09] paragraphs

[10] fall

[11] the

[12] only

[13] information

[14] of

[15] it

[16] should

[17] be

[18] avoided

[19] as

[20] much

[21] as

[22] possible

[23] to

[24] avoid

[25] duplication

[26] of

[27] content

[28] to

[29] enhance

[30] the

[31] accuracy

[32] of

[33] content

[34] changes

[35] to

[36] identify

[37] the

[38] degree

[39] of

[40] precision

[41]

[42] .

[0154] The slice length is set to 19, and the slice length is used to limit the number of characters of each slice sentence;

[0155] The slice overlap length is set to 10, and the slice overlap length is used to limit the number of overlapping characters between adjacent two slice sentences;

[0156] → Adjacent length of adjacent slices: 19-10+1=10 characters;

[0157] The generated slice numbers are as follows:

[0158] Slice number is C1, starting position is 0, character content is "The paragraph identifier is the unique information used to distinguish different paragraphs. It", the positioned first character is "段", and the positioned last character is "它";

[0159] Slice number is C2, starting position is 10, character content is "unique information for the same paragraph. It should尽量避免重复,以", the positioned first character is "同", and the positioned last character is "以";

[0160] Slice number is C3, starting position is 20, character content is "should尽量避免重复,以enhance the recognition after content change", the positioned first character is "应", and the positioned last character is "别";

[0161] Slice number is C4, starting position is 30, character content is "enhance the recognition accuracy after content change.", the positioned first character is "增", and the positioned last character is ".";

[0162] Finally, only 4 slices are generated because starting from index 30 and taking 20 more characters backward is just enough; starting from index 40 will be less than 20 characters, so no more slices are generated.

[0163] The reason for using the full stop as the positioned last character in the operation is that the goal of slicing sentences is to keep the character granularity as stable as possible and track the position range; it doesn't necessarily match the semantic sentence and is a kind of character range sliding; at this time, any character is part of the structure, even the full stop, which is also used for positioning; if the last character is deleted, it may cause the repetition of two slice identifiers and reduce the distinguishability.

[0164] Generate the slice identifier for each slice:

[0165] Content of C1: The paragraph identifier is the unique information used to distinguish different paragraphs. It;

[0166] First character: "段" → Unicode = 27573, Unicode is the binary encoding of the prior art;

[0167] Last character: "它" → Unicode = 24687;

[0168] Exclusive OR part: 27573 ⊕ 24687 = 5154;

[0169] Slice identifier: 5154 ⊕ 1729312345 = 1729317499, ⊕ is the exclusive OR operator;

[0170] Content of C2: unique information for the same paragraph. It should尽量避免重复,以;

[0171] First character: "same" -> Unicode=21516;

[0172] Tail character: "with" -> Unicode=20197;

[0173] XOR part: 21516⊕20197=1985;

[0174] Slice identifier: 1985⊕1729312345=1729314336;

[0175] C3 content: Should try to avoid duplication to enhance the recognition after content change;

[0176] First character: "should" -> Unicode=24212;

[0177] Tail character: "else" -> Unicode=21035;

[0178] XOR part: 24212⊕21035=3823;

[0179] Slice identifier: 3823⊕1729312345=1729316166;

[0180] C4 content: Enhance the recognition accuracy after content change;

[0181] First character: "increase" -> Unicode=22686;

[0182] Tail character: "." -> Unicode=12290;

[0183] XOR part: 22686⊕12290=10656;

[0184] Slice identifier: 10656⊕1729312345=1729323001;

[0185] Use the slice identifier to match the backup marked paragraph, specifically including:

[0186] Get the save time of all backup documents, and represent the save time as a timestamp;

[0187] Compare the timestamp of the target paragraph creation time with the timestamp of each backup document save time, and get the backup file corresponding to the maximum timestamp earlier than the target paragraph timestamp as the backup document of the first save target paragraph;

[0188] Mark the backup document of the first save target paragraph as the starting backup document;

[0189] Save the documents of the other two versions of the incomplete paragraph P1 respectively:

[0190] Version number V1 keeps the time stamp of the time 1755548300. The backup document contains 3 paragraphs, and the time stamps of the creation time of each paragraph are [first paragraph 1729311900, second paragraph 1755548410, third paragraph 1755548405] respectively;

[0191] Version number V2 keeps the time stamp of the time 1755548500. The backup document contains 3 paragraphs, and the time stamps of the creation time of each paragraph are [first paragraph 1729312343, second paragraph 1755548410, third paragraph 1755548405] respectively;

[0192] The time stamp of the creation time of the paragraph P1 is 1755548400, which is earlier than the maximum time stamp of the time stamp of the paragraph P1, and the backup file corresponding to the maximum time stamp is the backup document of version number V1.

[0193] The target document is V5.

[0194] The backup documents of version numbers V1, V2, V3, V4, and V5 form a paragraph associated document set.

[0195] For any one backup document in the paragraph associated document set:

[0196] The time error threshold is set to 10 seconds on the time stamp.

[0197] The time stamp of the creation time of the second paragraph in the backup document of version number V1 is 1755548410, and the corresponding creation time interval is [1755548400, 1755548420]. The time stamp of the creation time of the third paragraph is 1755548405, and the corresponding creation time interval is [1755548395, 1755548415].

[0198] The time stamp of the creation time of the second paragraph in the backup document of version number V2 is 1755548410, and the corresponding creation time interval is [1755548400, 1755548420]. The time stamp of the creation time of the third paragraph is 1755548405, and the corresponding creation time interval is [1755548395, 1755548415].

[0199] The time stamps are not necessarily completely consistent due to network delay or different system time calculation accuracy.

[0200] The time stamps of the creation time of the paragraph P1 are all in the creation time interval of the second paragraph and the third paragraph in the backup files of version number V1 and version number V2.

[0201] The sentence identification of the candidate paragraph is compared with all the slice identifications of the target paragraph.

[0202] The equal determination threshold is set to 2;

[0203] The number of clause identifiers equal to the slice identifier in the second paragraph in the backup document of version number V1 is equal to 2, corresponding to C1 and C2 respectively, and the second paragraph is "The paragraph identifier is unique information for distinguishing different paragraphs. It should avoid repetition as much as possible, to" just correspond to the slice clause: then the second paragraph in the backup document of version number V1 is an evolving paragraph;

[0204] The number of clause identifiers equal to the slice identifier in the second paragraph in the backup document of version number V2 is equal to 3, corresponding to C1, C2 and C3 respectively, and the second paragraph is "The paragraph identifier is unique information for distinguishing different paragraphs. It should avoid repetition as much as possible to enhance the identification after the content changes" then the second paragraph in the backup document of version number V2 is an evolving paragraph;

[0205] The version number V3, the version number V4 and the version number V5 all contain the complete paragraph P1;

[0206] The deleted paragraph optimal saving module is used to determine the optimal way to save the marked paragraph according to the word count of the backup marked paragraph and the deleted marked paragraph, comprising:

[0207] Get the word count of all evolving paragraphs in the backup document set of the paragraph associated document, calculate the sum, and record it as evolving total word count = 30+40+44+44+44=202;

[0208] Get the word count of the target paragraph, and compare the word count with the evolving total word count:

[0209] If the word count is less than (evolving total word count / 2)=202 / 2=101, 44<101, then calculate the time length from the backup document saving time to the current time, and record it as the first time length = 1400 seconds;

[0210] Get the number of times the starting backup document is queried, and record it as the first number = 30 times;

[0211] The time length from the target document saving time to the current time is the second time length = 300 seconds;

[0212] Get the number of times the target document is queried, and record it as the second number = 15 times;

[0213] Calculate the first number divided by the first time length, and record it as the first query frequency = 30 / 1400≈0.024;

[0214] Calculate the second number divided by the second time length, and record it as the second query frequency = 15 / 300=0.5;

[0215] At this time, the first query frequency is less than the second query frequency, the target paragraph in the target document V5 is retained, and all remaining evolution paragraphs in the paragraph associated document set are deleted.

[0216] Hash values or binary identifiers are used to locate and identify paragraphs, instead of word-by-word comparison:

[0217] More efficient: Hash value comparison is a fixed-length operation (e.g. 256 bits), which is much faster than word-by-word comparison, especially in large text segments.

[0218] Save storage and indexing is more convenient: Hash values or binary identifiers have fixed length, suitable for use as index keys, while word content is variable length and difficult to index.

[0219] Fault tolerance: Evolution paragraphs that are "similar but not exactly the same" can be identified through slicing and matching thresholds, while word-by-word comparison cannot handle slight changes in content.

[0220] Good privacy: Hash values are not reversible, and can achieve content desensitization to protect user data privacy.

[0221] It should be noted that in this paper, relationship terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between the entities or operations. Moreover, the term "includes", "contains" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or equipment.

[0222] The above is only the preferred embodiment of the present application, it should be noted that for those skilled in the art, without departing from the technical principles of the present application, a number of improvements and refinements can be made, which should also be considered as the protection scope of the present application.

Claims

1. A monitoring operating system for a data disaster recovery backup control system, characterized in that, include: The paragraph management module records the creation time and last update time of each paragraph in the document. It categorizes paragraphs based on set word count and time thresholds, and marks backed-up and optimized paragraphs as marked paragraphs. The paragraph identifier generation module is used to extract the binary representation of the first and last characters of each sentence in the marked paragraph and calculate the XOR value. It then calculates the XOR value of each sentence with the creation time of the marked paragraph to generate the binary set of the marked paragraph and generates the hash value of the marked paragraph based on the binary set. The paragraph deletion detection module is used to determine whether the hash value of the marked paragraph exists. When the hash value of the marked paragraph does not exist, it is determined that the marked paragraph has been deleted, and the marked paragraph is backed up and optimized. The paragraph tracking module is used to slice marked paragraphs into multiple slice sentences. It calculates the slice identifier of the slice sentence based on the first and last characters of the slice sentence and the creation time of the marked paragraph, and uses the slice identifier to match the backed-up marked paragraph. The optimal way to save deleted paragraphs is determined based on the number of words in the backed-up and deleted paragraphs.

2. The monitoring operating system for a data disaster recovery backup control system according to claim 1, characterized in that, The paragraph management module records the creation and last update times of each paragraph in the document, and categorizes paragraphs based on set word count and time thresholds, including: Set an automatic save interval, which is used to automatically save the document at the automatic save interval when the user creates the document, and generate a backup document of the document; Set a word count threshold; Set a time threshold, which is used to determine whether the paragraph has been created; For any paragraph in the document: Count the number of words in a paragraph and compare it to a word count threshold: If the word count is greater than or equal to the word count threshold, then calculate the time interval between the paragraph's last update time and the current time, and compare the time interval with the time threshold: If the time interval is greater than or equal to the time threshold, the paragraph is marked as a tagged paragraph for backup optimization.

3. The monitoring operating system for a data disaster recovery backup control system according to claim 1, characterized in that, The paragraph identifier generation module is used to extract the binary representation of the first and last characters of each sentence in the marked paragraph and calculate their XOR value. It then calculates the XOR value of each sentence with the creation time of the marked paragraph to generate a binary set of the marked paragraph. Based on this binary set, it generates a hash value for the marked paragraph, including: For any paragraph: Divide the paragraph into multiple clauses according to the period. ,in, Let be the i-th clause, and n be the number of clauses. : Get Sentences The first character Calculate the binary representation of the first character. ,in, The binary representation of the first character. Used to calculate the binary representation of a character; Get Sentences The last character ; Calculate the binary representation of the last character. ,in, The binary representation of the last character; calculate and XOR operation: ,in, It is the XOR symbol. For clauses The basic XOR value; clauses The creation time of the paragraph is represented as a timestamp. ; The XOR operation is performed to calculate the timestamp and the base XOR value: ,in, Used as a clause marker; For any marked paragraph: Obtain the clause identifiers of all clauses in the marked paragraph to obtain the binary set. ; Calculate the hash value of the marked paragraph by applying a hash function to the binary set. , , This is a hash function.

4. The monitoring operating system for a data disaster recovery backup control system according to claim 3, characterized in that, The paragraph deletion detection module is used to determine whether the hash value of the marked paragraph exists. When the hash value of the marked paragraph does not exist, it is determined that the marked paragraph has been deleted, and the marked paragraph backup optimization is performed, including: Arrange backup documents in the order they were saved; Calculate the hash value of all marked paragraphs in each backup document; Designate any backup document as the target document; Set the number of verifications; The previous backup document of the target document is obtained in the order of its saving and is recorded as the forward reference document; After obtaining the target document in the order of saving, verify the number of backup documents to form a backward reference document set; Compare the hash values ​​of marked paragraphs in the forward reference document, target document, and backward reference document sets: If the hash value of a marked paragraph in the target document exists in the forward reference document but not in any of the backward reference document sets, then the marked paragraph is considered to have been deleted from the target document. The paragraph marked as deleted in the target document will be denoted as the target paragraph.

5. The monitoring operating system for a data disaster recovery backup control system according to claim 4, characterized in that, The deleted paragraph tracking module is used to slice the marked paragraph into multiple slice sentences, including: Set the slice length, which is used to limit the number of characters in each slice clause; Set the slice overlap length, which is used to limit the number of characters overlapping between two adjacent slice clauses; Get the first character of the target paragraph and denote it as the first character to be positioned. S1. Starting from the first character of the positioning, select the nth character of the slice length according to the character arrangement order of the target paragraph, and denot it as the last character of the positioning; S2. Determine if the last character of the positioning position is empty: If the last positioned character is not empty, then the sequence of characters between the first positioned character, the last positioned character, and the first positioned character and the last positioned character is recorded as a slice clause; S3. Calculate the slice length - slice overlap length + 1, and record the result as the adjacent length; Starting from the first character, select the adjacent character of length according to the character arrangement order of the target paragraph, and use it as the new first character. Repeat S1-S3. S4. If the last character of the positioning is empty: Obtain the last character of the target paragraph, and combine the first character of the target paragraph with the last character and the character sequence between the first character of the target paragraph and the last character to form a segmented sentence.

6. The monitoring operating system for a data disaster recovery backup control system according to claim 5, characterized in that, The step of calculating the slice identifier of a slice sentence based on the first and last characters of the slice sentence and the creation time of the marked paragraph, and using the slice identifier to match the backed-up marked paragraph, includes: Get the binary representation of the first and last characters of the slice clause, calculate the XOR value of the binary representation of the first and last characters, and denot it as the slice XOR value. Obtain the creation time of the target paragraph and represent the creation time as a timestamp; Calculate the XOR value of the timestamp and the slice XOR value, and denote it as the slice identifier; Use slice identifiers to match the marked segments in the backup, specifically including: Get the save time of all backup documents and represent the save time as a timestamp; Compare the timestamp of the target paragraph creation time with the timestamp of each backup document saving time, and obtain the backup file corresponding to the largest timestamp earlier than the target paragraph's timestamp as the backup document for the first time the target paragraph is saved; Record the first backup document containing the target paragraph as the starting backup document; Retrieve all backup documents between the starting backup document and the target document according to the order in which they were saved, and combine them with the starting backup document and the target document to form a paragraph-related document set.

7. The monitoring operating system for a data disaster recovery backup control system according to claim 6, characterized in that, The step of calculating the slice identifier of the slice sentence based on the first and last characters of the slice sentence and the creation time of the marked paragraph, and using the slice identifier to match the backed-up marked paragraph, also includes: For any backup document in the paragraph-related document set: Set a time error threshold; The time before the creation time of each paragraph in the backup document is recorded as the lower creation limit time, and the time after the creation time is recorded as the upper creation limit time. The creation time interval is composed of the upper limit creation time and the lower limit creation time. Compare the creation time of the target paragraph with the creation time range of each paragraph in the backup document; Obtain the creation time interval of the target paragraph, and record the paragraphs in the backup document that correspond to the creation time interval as candidate paragraphs; Compare the clause markers of the candidate paragraph with all the segment markers of the target paragraph; Set an equality threshold; When the number of clause identifiers in a candidate paragraph that are equal to the slice identifier is greater than or equal to the equality judgment threshold, the candidate paragraph is identified as an evolved paragraph of the target paragraph.

8. The monitoring operating system for a data disaster recovery backup control system according to claim 7, characterized in that, The optimal method for saving deleted paragraphs is used to determine the optimal way to save the deleted paragraphs based on the number of words in the backed-up and deleted paragraphs, including: Get the word count of all backup documents in the paragraph-related document set, calculate the sum, and record it as the total word count of evolution; Get the word count of the target paragraph and compare it with the total word count of the evolved paragraph: If the number of characters is less than (total number of characters in the evolution / 2), then calculate the duration from the time the initial backup document was saved to the current time, and record it as the first duration; Get the number of times the initial backup document was queried, and record it as the first query. The duration from when the target document was saved to the current time is defined as the second duration. The number of times the target document was queried is recorded as the second query. Calculate the first number and divide it by the first duration; record this as the first query frequency. Calculate the second number and divide it by the second duration; this is recorded as the second query frequency. If the first query frequency is greater than the second query frequency, the evolved paragraph in the starting backup document will be replaced with the target paragraph, and all remaining evolved paragraphs in the paragraph-associated document set will be deleted. If the first query frequency is less than or equal to the second query frequency, the target paragraph in the target document is retained, and all remaining evolved paragraphs in the paragraph-related document set are deleted.

Images