Patent Application
20130325829
The present invention relates generally to pattern matching between sequences of transactions, and more particularly to multi-character elements in which the elements are interleaved differently in each sequence.
Databases are routinely upgraded to new versions, or new software patches are applied on existing versions, or the database is migrated to a new database system. In each of these situations, it is typical to compare the performance of a benchmark transaction workload in the new database environment as compared to the old database environment. A benchmark transaction workload is a sequence of different transaction types. In a typical database environment, each transaction can be a sequence of one or more Structured Query Language (SQL) statements. To compare the performances of the benchmark transaction workloads, corresponding instances of transactions in the new and old database environments are matched. Matching and comparing corresponding transactions is necessary because there will typically be workloads occurring in the database environments that are extraneous to the benchmark workloads, and performance of the same transaction type varies during the workload execution since the underlying data is continuously modulated. The benchmark transaction workload comparison typically involves comparing the transaction logs of the old and new database environments.
A simple form of matching involves identifying the occurrences of a short character string inside a longer character string. A simple approach to this problem is advancing the shorter string through the longer string one character at a time, and determining if there is a match between the shorter string and the corresponding characters of the longer string. If there is not a match, the shorter string is advanced by one character in the longer string.
Another well-known matching problem involves sequence alignment. This problem is perhaps best known in the context of DNA matching. A full DNA sequence can be characterized as a series of shorter four-character sequences, where each character of the a shorter sequence can be one of A, C, G, or T, representing different nucleotide bases. Various techniques have been developed to determine the alignment of two DNA sequences that will give the highest level of correspondence between the two sequences.
Neither of these known matching solutions address matching transactions in the more complicated environment usually found associated with databases.
Embodiments of the present invention provide a system, method, and program product for identifying matching transactions between two log files. Each transaction includes one or more programming statements, each programming statement includes one or more tokens and an indicator indicating if the programming statement is the end-of-transaction programming statement of the transaction, each log file contains records recording the execution on a computer of the programming statements, each record includes a transaction identifier, the tokens, and the end-of-transaction indicator indicating if the record is the end-of-transaction record of the transaction of the corresponding executed programming statement. A computer, in response to determining that a least recently read record of a fixed number of sequentially read records from a first log file is not associated with an indication of match, removes references to the transaction identifier and the programming statement associated with the least recently read record from a first table that includes associations of programming statements to the transaction identifiers to which they are associated, and associations of transaction identifiers to their associated programming statements. The computer, in response to determining that the least recently read record is an end-of-transaction record, removes references to the transaction identifier associated with the record from the first table. The computer reads a next sequential record from the first log file into the fixed number of sequentially read records. The computer adds information to the first table, based on the transaction identifier and the programming statement information contained in the next sequential record read from the first log file, associating the programming statement to the transaction identifier, and associating the transaction identifier to the programming statement. The computer reads from a second log file the record at a position in the second log file a defined fixed offset of records away from the position in the second log file corresponding to the position in the first log file of the next sequential record read from the first log file. The computer determines if the record read from the second log file matches a record in the fixed number of sequentially read records. The computer, in response to determining that the record read from the second log file matches a record in the fixed number of sequentially read records, adds information to a second table that contains associations of transaction identifiers contained in records read from the second log file to transaction identifiers contained in records read from the first log file, based on information contained in the first table. The computer, in response to determining that the record read from the second log file does not match a record in the fixed number of sequentially read records, removes references to the transaction identifier contained in the record read from the second log file from the second table. The computer, in response to determining that the record read from the second log file matches a record in the fixed number of sequentially read records and that the record read from the second log file is an end-of-transaction record, identifies a transaction from the first log file that matches a transaction from the second log file based on information contained in the second table.
Embodiments of the invention are directed to heuristic approaches for identifying matching transactions between two transaction log files containing sequences of SQL records, where each transaction includes one or more SQL records. In embodiments of the invention, the first transaction log file is called the “capture” file, and the second transaction log file is called the “replay” file. The heuristic algorithms of the present invention advance down the replay file one record at a time and attempt to find matching records and transactions in the capture file.
Certain assumptions are made with respect to the SQL records in the transaction log files. These assumptions are typically true for records in a transaction log file. The heuristic algorithms and the various assumptions of the invention may result in a small fraction of matches not being identified. However, there will typically be enough actual matches identified such that the missed matches are relatively insignificant.
Embodiments of the present invention will now be described in detail with reference to the accompanying drawings.
In preferred embodiments of the invention, user computing device 110 can be, for example, a mainframe or mini computer, a laptop, or netbook personal computer (PC), or a desktop computer. Transaction matching system 100 is shown as being wholly implemented on user computing device 110. However, transaction matching system 100 may operate in a distributed environment in which one or more of its components are implemented across a plurality of computing devices that communicate over a network, such as a local area network (LAN) or a wide area network (WAN) such as the Internet. For example, benchmark analysis system 140 may operate on a separate computing device having sufficient capabilities to support only the operation of benchmark analysis system 140. In general, transaction matching system 100 can execute on any computing device 110, or combination of computing devices, satisfying desired implementation requirements, and as described in relation to
Transaction processing system 120 includes transaction manager 122, log manager 124, and transaction log file 126. Transaction manager 122 manages the processes that execute transactions against database 132 via database management system 130. Transaction manager 122 also manages all transactions so as to maintain data consistency in database 132. This is accomplished through the use of log manager 124. Log manager 124, among its other activities, records each transaction operation of a transaction workload, such as the execution of SQL statements in a transaction, in a SQL record to a transaction log file 126.
Database management system 130 includes database 132, which may reside, for example, on tangible storage device 830 (see
Benchmark analysis system 140 operates generally to analyze different executions of a benchmark transaction workload, and provide systems and applications programmers and systems administrators information to determine, for example, the most efficient organization of a database 132, or of a transaction workload, or for determining the most efficient database management system 130 or transaction processing system 120. The information that benchmark analysis system 140 processes is derived from transaction log file 126. For example, the transaction log file 126 information pertaining to two different executions of a benchmark transaction workload are stored on disk, such as tangible storage device 830, after each benchmark workload completes, and this information is made available to benchmark analysis system 140 for analysis.
Benchmark analysis system 140 includes transaction matching program 142, which is the focus of the present invention. Transaction matching program 142, the operation of which will be described in more detail below, operates generally to identify matching transactions between different executions of a benchmark transaction workload. These matching transactions are then analyzed by benchmark analysis system 140 to provide the information stated above.
Embodiments of the invention are described with respect to the components and their functionality as presented in
Although the same benchmark transaction workload may be executed twice in the same database environment, the transaction log files 126 of these executions may be different for several reasons. For example, otherwise matching SQL statement clauses that include current date or time host variables may produce different log records. There may also be other workloads executing on the database systems that produce extraneous log records that will be intermixed in the transaction log file 126 with the benchmark transaction workload records. In addition, SQL records for a given transaction may appear in different orders in different log files. If the transaction workload is executed on different database systems, the transaction file 126 log records may reflect that various system variables may have different names between the two database systems.
Because of these and other differences between the transaction log files 126, the heuristic approaches of embodiments of the invention use various assumptions. For example, one assumption is that a record in the replay file will find a match in the capture file, if there is a match, within a certain range or “match window”. This assumption serves to recognize that records can be out of order between the capture and replay files, and that there may be extraneous records in the replay file for which there is no matching record in the capture file. The assumption also serves to limit, or bound, the number of compare operations and thus limit the computer resources consumed by the compares. The trade-off for this assumption is that if there is a capture record outside of the match window that does match the current replay record, this match will not be found and the replay record will be flagged as extraneous. This and other assumptions will be discussed in more detail below.
A benchmark transaction workload typically includes a sequence of database transactions. Each database transaction will typically comprise a sequence of multiple SQL statements. When a series of database transactions are executed, for example, by transaction processing system 120, the SQL statements of a transaction are typically executing in an interleaved manner with the SQL statements of other database transactions of the benchmark transaction workload. Thus, although the database transactions may execute in sequence, their underlying SQL statements can be interleaved with SQL statements of other transactions, and the corresponding transaction log file 126 records will be similarly interleaved. Further, different executions of the same benchmark transaction workload can produce different execution sequences of the underlying SQL transactions. This could be due to such factors as I/O scheduling by the operating system, I/O delay, network delay, locking and latching within the database, etc. Although the execution of the underlying SQL statements of different transactions may be interleaved, the SQL statements in a given transaction will execute in order, and will appear in the transaction log file 126 in order.
SQL statements belonging to a database transaction might also be changed when executed in different database environments. This could result, for example, from the way certain database environments optimize SQL statements prior to execution. In embodiments of the invention, another assumption is made that, with regard to a changed SQL statement, each occurrence of the changed statement will be the same. For example, if database environment optimizes a certain SQL statement, then every occurrence of the SQL statement will be optimized in the same way. In certain embodiments, if a changed SQL statement in the replay file can at least be partially matched to its corresponding SQL statement in the capture file, and this partial match occurs a certain number of times, this correspondence is recorded and subsequent occurrences of the changed SQL statement are automatically changed to the unchanged version of the SQL statement, thus reducing the resources associated with processing partial SQL statement matches.
A benchmark transaction log file 126 may also contain extraneous SQL records. These are records in the capture and replay files that cannot be matched to records in the other file. Extraneous records may result from transactions executing in a database environment that are not part of the benchmark transaction workload.
As mentioned above, the same SQL statement executed at different times in the same database environment, or different database environments, can produce different transaction log file 126 records. In embodiments of the invention, these different log records can be identified as partial matches, and, depending on how close the match is, can be considered as matches.
With respect to SQL statement matching, each SQL statement is considered an ordered sequence of tokens. SQL record in the transaction log files 126 are compared token by token rather than comparing SQL record as complete character strings. A SQL token is a word or character that can be identified meaningfully when the SQL statement is parsed, or interpreted. For example, each token can typically be considered a keyword, an identifier, a quoted identifier, a constant, or one of several special character symbols. SQL statement clauses can include several tokens. Token by token comparisons allow for meaningful determinations of partial matches, and identifications of complete matches, even though not all corresponding token values are identical. For example, there might be some differences in the host variables between log files. If the only difference between two SQL record is differences in the host variables, this could be considered a high-scoring partial match. In preferred embodiments of the invention, tokens matching is done in a Boolean manner, i.e., either a pair of corresponding tokens match exactly or they don't match at all. For instance, there is no partial matching between two tables named “Order” and “Orders”. This token based string comparison also helps to remove the differences that arise due to change in comments, change in schema name, etc. For example, comments can be ignored or stripped from the record during the compare step.
An assumption with regard to SQL record matching is that if a replay SQL record partially matches a capture SQL record, and the SQL record includes multiple host variables, then at least one of the host variables in the matching records should match. This assumption is one way to increase the likelihood that a partial match between a replay transaction and a capture transaction with a match score above the match score threshold is in fact an actual match. If the potentially matching records do not have a common host variable, then a mismatch is declared.
In preferred embodiments of the invention, the similarity between two SQL records is determined by assigning a numerical value to a match score. For example, if all tokens of a SQL record match those of another, the match score is 1. If no tokens match, the match score is 0. If the first token, which will typically be the statement or command type, doesn't match, then the entire record may be considered as not matching. A partial match score can be based, for example, on the percentage of tokens that match. If most tokens match, except for host variable tokens, a percentage increase may be added to the match score. In preferred embodiments a match score threshold is defined, below which a mismatch is declared. For example, a match threshold of 80% may be defined. Because a transaction workload often involves repeating sequences of transactions, it is possible that a replay file record may partially match more than one capture file records with a match score above the mismatch threshold.
As mentioned above, embodiments of the invention use a match window as an assumption to limit the number of records that are compared in the capture file to a replay file record. The capture file match window is expressed as a 2K+1 window, where K is the number of records searched before and after the capture file record that corresponds to the replay record that is being matched. For example, if capture records having log file index locations 80 to 120 are searched for a match to a replay record at log file index location 100, K equals 20 and the match window is 41 records. Generally, embodiments of the invention step down the replay file one record at a time, and search the corresponding match window in the capture file for matching records. If a matching capture file record is not found in the match window, the replay record is marked as extraneous. If the match window advances beyond a capture file record that has not been matched, the capture file record is considered extraneous. In certain embodiments of the invention, the match window is not centered on the corresponding record number of the record being read in the replay file, but is a fixed offset number of records away. The match window can be implemented, for example, as a circular buffer with a length of 2K+1.
The number of records in the match window is typically based on an input to the algorithm, for example 501 records, and can be based on such considerations as the number of extraneous records in a log file, and actual statistics related to distance between matching records. The window size can be adjusted to accommodate performance and accuracy considerations.
Both the replay records and the capture records as illustrated include the underlying SQL statements for the following two transactions:
In this example, capture record A1 at position 1 has a better match score than record A2 at position 3, because record A2 is from a different transaction, and the match between replay record A1 at position 1 and capture record A1 at position 1 is recorded (step 206), as indicated in
This case illustrates that match window 300 has advanced to be centered at capture file position 5 (step 210), and capture file record M5 at position 2 is dropped from match window 300 (step 212, “yes” branch) before it was matched with any replay records. As a result, capture file record M5 at position 2 has been marked as extraneous (step 214). As illustrated, it appears that replay record M5 at position 10 may have matched capture file record M5 at position 2. However, because the match window is five records, these two records will not be compared to each other.
In this case, a match is recorded between replay record N5 at position 5, and capture record N5 at position 7. This match is shown for consistency in the algorithm as illustrated. In operation, embodiments of the invention also include an advancing window for transactions. If the current replay record being compared cannot be a SQL record in any of the transactions in the transaction window, the record would be marked as extraneous. This will be explained more fully below.
With respect to transaction matching generally, a capture transaction will match a replay transaction if all the SQL record of the capture transaction match all the SQL record of the replay transaction. However, as stated above, partial matching of SQL records are possible. Thus, if a replay transaction has more than one potential matching capture transactions, the capture transaction with the highest match score is considered the best match. For example, the capture transaction having the greatest sum of SQL record match scores with relation to a specific replay transaction is considered the best candidate for a transaction match.
An assumption with regard to transaction matching is that at least one of the underlying SQL statements of a replay transaction will have an exact match with the corresponding SQL statement of a potentially matching capture statement. This assumption is one way to increase the likelihood that a declared match between a replay transaction and a capture transaction is in fact an actual match. These exact SQL record matches are recorded in an exact match table, and can be used to improve the likelihood of identifying an actual matching transaction from the potential matching transactions in replay table XACT→C-XACT. For example, replay table R-XACT→C-XACT may indicate that a replay transaction has more than one potentially matching capture transactions after the end-of-transaction record for the replay transaction has been read. If one or more potentially matching capture transactions have SQL records that match exactly with corresponding SQL record in the replay transaction, as indicated in the exact match table, then the other potentially matching capture transactions can be eliminated from the table.
Window 500 represents the capture file match window. In this example, K equals 1, and the full match window size is 3. Several tables used by transaction matching program 142 are represented in
Also associated with transactions in the capture file is the Committed Xacts buffer. This is a buffer of 2K+1 records, and is used to determine if a capture transaction for which the end-of-transaction SQL record has been read, indicating that transaction manager 122 has performed a “commit operation” for the transaction, finds a corresponding replay end-of-transaction SQL record within 2K+1 replay log file records. This buffer can be implemented as a circular buffer having a length of 2K+1, and may be implemented using the match window circular buffer.
Associated with transactions in the replay file is the R-XACT→C-XACT table, which is used by transaction matching program 142 to associate partial replay transaction to potential matching capture transactions. For example, two replay SQL records for a replay transaction have been read by transaction matching program 142, and the end-of-transaction SQL record for this transaction has not yet been read. Several records from the capture file have also been read from match window 500 relating to different transactions. There may be several capture transactions that include at least two SQL records corresponding to the order and the type of the two replay SQL records. These capture transactions would be included in the replay R-XACT→C-XACT table. When the end-of-transaction SQL record for a replay transaction is read, indicating that the final SQL record for this transaction has been read, this table is used by transaction matching program 142 to identify the remaining potential matching capture transactions for the replay transaction. Information to populate this table comes from entries in the capture file SQL→XACT and XACT→SQL tables, described above.
Additional tables used by transaction matching program 142 include the Transaction Matches table, which is used to record instances where a single capture transaction is identified as matching a replay transaction. The Exact SQL Matches table is used to identify exact matches between replay transactions and capture transactions.
The Declared XACT Match table is used to identify matches between recurring pairs of replay and capture transactions in which the match would be exact, except for the same recurring differences in their underlying SQL tokens. This would be the case if a capture SQL statement was changed by transaction processing system 120 prior to submission by transaction manager 122. As stated above, an assumption is made that if such a capture SQL statement was changed by transaction processing system 120 prior to submission by transaction manager 122, every occurrence of the statement will be changed in the same way. If such a partial transaction match occurs a certain number of times, an entry is recorded in the table that maps the partially matching capture SQL record to the matching replay SQL record. Entries in this table may be used by transaction matching program 142 to modify subsequent occurrences of the partially matching capture SQL record to exactly match the corresponding replay SQL record upon the partially matching capture SQL record's entry into the match window. This allows transaction matching program 142 to quickly declare the match without spending computing resources to arrive at the same decision.
As the initial buffers and pointers are established, no unmatched or non-extraneous capture records are dropped from match window 500 (step 402, “no” branch), and no capture transactions are pushed from the Committed Xacts buffer (step 406, “no” branch). Capture records A1 and B1c have entered the match window (because this is the first iteration of the algorithm). The associations between the records and the capture transactions to which they belong are recorded in capture tables SQL→XACT and XACT→SQL (step 410). In this case, both A11 and B1c2 are associated with capture transaction 1 (TC1). Because capture record B1c is an end-of-transaction record for TC1 (step 412, “yes” branch), capture transaction TC1 is recorded in position 1 of the Committed Xacts buffer (step 414).
Transaction matching program 142 then performs a comparison of replay record 1 to the unmatched and non-extraneous capture records in the match window (step 416). As stated previously, one of the assumptions is that although the order of the SQL records in the capture file may be different than the order of the corresponding SQL records in the replay file, a record in the replay file will find a match in the capture file, if there is a match, within a match window of 2K+1 records of the corresponding capture file index position. Thus, only records in the match window are compared to the current replay record to identify a match. Further, in the preferred embodiment of the invention, the other assumptions described above relating to SQL record matching apply with respect the operation of transaction matching program 142.
In this example, replay record A1 at position 1 (replay A11) is an exact match to capture record A1 at position 1 (capture A11). Because the match score between replay A11 and capture A11 is greater than the match score threshold (step 418, “yes” branch), indicating at least a partial match, an entry is added to replay table R-XACT→C-XACT (step 420). In this case, capture table SQL XACT indicates that capture A11 is associated with capture transaction 1 (TC1), and an entry associating replay transaction 1 (TR1) to TC1 is added to replay table R-XACT→C-XACT. This entry indicates that, based on the capture and replay records read so far, partial transaction TR1 has a potential match to partial transaction TC1. Capture record A1 at position 1 is also removed from match window 500 (step 422) so it will not be compared with subsequent replay records (see step 416), or flagged as extraneous (see step 404).
In this example, replay A11 is an exact match to capture A11 (step 424, “yes” branch), so an entry is added to the Exact SQL Matches table (step 426). Because the match is not less than the match score threshold (step 428, “no” branch), and the replay record is not an end-of-transaction record (step 432, “no” branch), processing returns to step 400.
Replay record A2 at position 2 is then compared to the unmatched and non-extraneous capture records in match window 500 (step 416). In this case, the match between replay A22 and capture A23 has a match score greater than the match score threshold (step 418, “yes” branch). Replay table R-XACT→C-XACT is then updated to include an entry indicating the potential capture transactions that partial replay transaction T2 might match (step 420). In this case, as indicated in the SQL→XACT table, a record of non-end-of-transaction type “A” appears in capture transactions TC1 and TC2. Because not enough capture or replay record have been read to narrow the potential matches, the entry indicates that partial transaction TR2, of which only a record of type “A” has been read, has potential matches with capture transactions TC1 and TC2, both of which include a non-end-of-transaction type “A”. Capture record A2 at position 3 is also removed from match window 500 (step 422) so it will not be compared with subsequent replay records (see step 416), or flagged as extraneous (see step 404).
In this example, replay A22 is an exact match to capture A23 (step 424, “yes” branch), so an entry is added to the Exact SQL Matches table (step 426). Because the match is not less than the match score threshold (step 428, “no” branch), and the replay record is not an end-of-transaction record (step 432, “no” branch), processing returns to step 400.
No unmatched or non-extraneous capture records are dropped from match window 500 (step 402, “no” branch). Although the position previously occupied by capture record A1 at position 1 has been dropped from match window 500, record A11 was removed from window 500 because it was found to match a replay record with a match score greater than the match score threshold (see step 422). No capture transactions are pushed from the Committed Xacts buffer (step 406, “no” branch). Capture record B34 is recorded in the SQL→XACT and XACT→SQL capture tables (step 410). Capture B34 is not an end-of-transaction record (step 412, “no” branch).
Replay record B1c at position 3 is then compared to the unmatched and non-extraneous capture records in match window 500 (step 416). In this example, the match between replay B1c 3 and capture B1c 2 has a match score greater than the match score threshold (step 418, “yes” branch). Replay table R-XACT→C-XACT is updated to include information resulting from this match (step 420). In this case, the table already includes the potential match between TR1 and TC1, and the replay transaction TR1 is updated to include the just read replay B1c 3 record. Capture record B1c 3 is also removed from match window 500 (step 422). In this example, replay B1c 3 is not an exact match to capture B1c 2 (step 424, “no” branch), and the match score is not less than the match score threshold (step 428, “no” branch).
In certain embodiments, if the same non-exact but greater than match threshold match between a replay SQL record and a capture SQL record is found a certain number of times, an entry is added to the Declared SQL Matches table. As described above, the assumption is that this partial match is recurring because an otherwise exactly matching SQL statement has been changed, for example optimized, in one of the database environments that produced either the capture or replay log files, and every occurrence of the SQL statement has been changed in the same way. In these embodiments, each new capture record that is read is compared against entries in the Declared SQL Matches table. If a matching entry is found, the newly read capture record is changed to match the corresponding replay record.
Replay B1c 3 is an end-of-transaction record (step 432, “yes” branch). Transaction matching program 142, beginning with replay table R-XACT→C-XACT, then identifies the best match between the replay transaction for which the end-of-transaction record was just read, and the potential capture transaction matches to this replay transaction contained in the table (step 434). In this case, replay table R-XACT→C-XACT indicates that capture transaction TC1 is the only potential capture transaction match to replay transaction TR1. Thus, the TR1-to-TC1 match is recorded in the Transactions Matches table, and references to TR1, TC1, and their underlying SQL records are removed from the capture and replay tables (step 436). This includes removing the TCI entry from the capture Committed Xacts table, so that during the next iteration of the algorithm, the entry is not pushed from the table, which would result in all references to TC1 being flagged as extraneous. In a preferred embodiment, if replay table R-XACT→C-XACT had indicated that more than one capture transaction was a potential match to replay transaction TR1, it would be determined if any of the SQL records of TR1 exactly matched any of the SQL records in any of the potentially matching capture transactions, based on the entries in the Exact SQL Matches table. If any exact matches were found, those potentially matching capture transactions without exactly matching SQL records would be removed from replay table R-XACT→C-XACT. If there were still more than one potentially matching capture transaction, then a best match would be determined by comparing replay transaction TR1 to each potentially matching capture transaction token by token, and declaring a match between TR1 and the potentially matching capture transaction whose match operation resulted in the highest match score. Processing then returns to step 400.
In this case, replay table R-XACT→C-XACT indicated only a single matching capture transaction for the replay transaction for which the end-of-transaction record had just been read (see step 434). It is possible that a replay transaction for which the end-of-transaction record has just been read has more than one potential matching capture transactions. For example, two complete capture transactions of the same transaction type are read, and the end-of-transaction record for a corresponding replay transaction is read. In this example, all three transactions are the same transaction type, and all three have had all their underlying SQL statement records read. Here, it may be possible to eliminate one or more of the potential matching transactions by taking the order of the underlying SQL records into account. As stated above, one of the assumptions is that the SQL records for a transaction will appear in a log file in the correct order, although SQL records from different transactions may be interleaved. Thus, comparing the ordered replay transaction to the potentially matching ordered capture transactions may eliminate one or more capture transactions having the same underlying SQL statements, but in a different order. If the potentially matching capture transactions in fact have underlying SQL transactions in the same order as the replay transaction, then the best match to the replay transaction may be identified by finding the candidate capture transaction having the highest sum of SQL record match scores.
Similar to the previous iterations of the algorithm, no unmatched or non-extraneous capture records are dropped from match window 500 (step 402, “no” branch) and no capture transactions are pushed from the Committed Xacts buffer (step 406, “no” branch). Capture record D2c 7 is recorded in the SQL→XACT and XACT→SQL capture tables (step 410). Capture D2c 7 is an end-of-transaction record (step 412, “yes” branch), and an entry for capture transaction TC2 is added to capture table Committed Xacts at position 1.
Replay record D2c 6 is then compared to the unmatched and non-extraneous capture records in match window 500 (step 416). In this example, the match between replay D2c6 and capture D2c 7 has a match score greater than the match score threshold (step 418, “yes” branch). Replay table R-XACT→C-XACT is updated to include information resulting from this match (step 420). In this case, the table already includes the potential match between TR2 and TC2, and the transactions are updated to include the just read replay D2c6 and capture D2c 7 records. Capture record D2c 7 is also removed from match window 500 (step 422). In this example, replay D2c 6 is not an exact match to capture D2c 7 (step 424, “no” branch), and the match score is not less than the match score threshold (step 428, “no” branch).
Replay D2c 6 is an end-of-transaction record (step 432, “yes” branch). Transaction matching program 142, beginning with replay table R-XACT→C-XACT, then identifies the best match between the replay transaction for which the end-of-transaction record was just read, and the potential capture transaction matches to this replay transaction contained in the table (step 434). In this case, replay table R-XACT→C-XACT indicates that capture transaction TC2 is the only potential capture transaction match to replay transaction TR2. Thus, the TR2-to-TC2 match is recorded in the Transactions Matches table, and references to TR2, TC2, and their underlying SQL records are removed from the capture and replay tables (step 436). Processing then returns to step 400.
The rest of the replay and capture SQL records would be processed in a similar fashion, ultimately identifying all matching transactions. As mentioned above, information relating to the matching transactions can then be used by benchmark analysis system 140 to compare execution statistics of the two runs of the benchmark transaction workload.
The simplified example of
A capture SQL record is considered extraneous, generally, when the capture SQL record is not matched to a replay SQL record while the capture SQL record is within a certain number of records of the capture record having the same index as the replay record currently being processed. This is more easily understood in terms of match window 500, and the Committed Xacts capture table, described above. A first way that a capture SQL record is considered extraneous is when it is dropped from match window 500 before it is matched to a replay record (see step 402, “yes” branch). In this case, all entries in the capture tables SQL XACT and XACT→SQL that include the dropped capture record are removed. Entries in the replay table R-XACT→C-XACT that reference the capture transaction of which the dropped capture SQL record belongs are also removed (see step 404). In this manner, all references to the capture transaction of which the dropped capture SQL record belongs are removed from the tables.
A second way in which a capture transaction can be considered as extraneous is when the capture transaction is pushed from the capture Committed Xacts buffer (see step 406, “yes” branch). This occurs when an end-of-transaction capture SQL record is not matched to a corresponding replay end-of-transaction capture SQL record within 2K+1 replay records, which is the depth of the Committed Xacts buffer. Similar to the two ways described just above, when a capture transaction entry is dropped from the Committed Xacts buffer, references to the capture transaction are removed from the capture tables SQL→XACT and XACT→SQL, and from the replay table R-XACT→C-XACT (see step 408).
Data processing system 800, 900 is representative of any electronic device capable of executing machine-readable program instructions. Data processing system 800, 900 may be representative of a smart phone, a computer system, PDA, or other electronic devices. Examples of computing systems, environments, and/or configurations that may represented by data processing system 800, 900 include, but are not limited to, personal computer systems, server computer systems, thin clients, thick clients, hand-held or laptop devices, multiprocessor systems, microprocessor-based systems, network PCs, minicomputer systems, and distributed cloud computing environments that include any of the above systems or devices.
User computing device 110 includes internal components 800 and external components 900 illustrated in
Internal components 800 also include a R/W drive or interface 832 to read from and write to one or more portable computer-readable tangible storage devices 936 such as a CD-ROM, DVD, memory stick, magnetic tape, magnetic disk, optical disk or semiconductor storage device. The programs 122, 124, and 142 in user computing device 110 can be stored on one or more of the respective portable computer-readable tangible storage devices 936, read via the respective R/W drive or interface 832 and loaded into the respective hard drive 830.
Internal components 800 also include network adapters or interfaces 836 such as a TCP/IP adapter cards, wireless wi-fi interface cards, or 3G or 4G wireless interface cards or other wired or wireless communication links. The programs 122, 124, and 142 in user computing device 110 can be downloaded to user computing device 110 from an external computer via a network (for example, the Internet, a local area network or other, wide area network) and network adapters or interfaces 836. From the network adapters or interfaces 836, the programs 122, 124, and 142 in user computing device 110 are loaded into hard drive 830. The network may comprise copper wires, optical fibers, wireless transmission, routers, firewalls, switches, gateway computers and/or edge servers.
External components 900 can include a computer display monitor 920, a keyboard 930, and a computer mouse 934. External components 900 can also include touch screens, virtual keyboards, touch pads, pointing devices, and other human interface devices. Internal components 800 also include device drivers 840 to interface to computer display monitor 920, keyboard 930 and computer mouse 934. The device drivers 840, R/W drive or interface 832 and network adapter or interface 836 comprise hardware and software (stored in storage device 830 and/or ROM 824).
Aspects of the present invention have been described with respect to block diagrams and/or flowchart illustrations of methods, apparatus (system), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer instructions. These computer instructions may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions/acts specified in the flowchart and/or block diagram block or blocks.
The aforementioned programs can be written in any combination of one or more programming languages, including low-level, high-level, object-oriented or non object-oriented languages, such as Java, Smalltalk, C, and C++. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer, or entirely on a remote computer or server. In the latter scenario, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider). Alternatively, the functions of the aforementioned programs can be implemented in whole or in part by computer circuits and other hardware (not shown).
Based on the foregoing, computer system, method and program product have been disclosed in accordance with the present invention. However, numerous modifications and substitutions can be made without deviating from the scope of the present invention. Therefore, the present invention has been disclosed by way of example and not limitation.
This application is a continuation of U.S. patent application Ser. No. 13/483,778 filed on May 30, 2012, the entire content and disclosure of which is incorporated herein by reference.
| Number | Date | Country | |
|---|---|---|---|
| Parent | 13483778 | May 2012 | US |
| Child | 13906374 | US |
