30 results about "Multiversion concurrency control" patented technology

Various embodiments include one or more of systems, methods, software, and data structures that allow dual access to concurrent data in a database management system including uncommitted updated data. Some embodiments include a dual snapshot of data in a multi-version concurrency control style database, but the embodiments described herein are equally applicable to virtually any other type of database and database management system.

The embodiment of the invention discloses a multi-version concurrency control method in a database and a database system, which are mainly applied to the technical field of databases. The database system is provided with a data page chain of a page; the data page chain comprises a page compass of each version of the page; the page compass of one version of the page is used for pointing to another version of the page before last operation of the version of the page, so that the page can be directly rolled back to the page which needs to be read based on the data page chain of the page which is requested to be read if the time stamp of the current version of the page is greater than the time stamp including a reading service in a data reading request when pages in the database are read, a user is convenient to know the page in the database at any time and is convenient to inquire data of each version of the page in the database. Furthermore, the database system can realize roll-back of records efficiently through combination of the data page chain and a record chain, and can realize reading consistency.

A transaction creates a wait for dependency on a version in a main memory database implementing a multi-version concurrency control scheme. The wait for dependency allows the transaction to update the version while other transactions are reading the version. The multi-version concurrency control scheme also allows commit dependencies to be implementing concurrently with wait for dependencies. Commit dependencies allow a transaction to read an updated version before the updated version is committed.

The invention discloses a multi-version concurrency control method and device for a database, a server and a storage medium. The method comprises the steps of blocking a transaction number of a current transaction in an exclusive lock mode when the current transaction is started; if a current operation of the current transaction is a write operation, generating a transaction save point; searchingfor a target data page corresponding to the write operation, and locking the target data page in the exclusive lock mode; executing the current operation according to a preset rule. Through this blocking mechanism, the number of locks is not directly related to the number of records modified by transactions, the number of the locks can be greatly decreased, blocking conflicts and blocking resourceconsumption are reduced, and the concurrency efficiency of a database management system is improved.

The invention discloses a distributed transaction processing system using a multi-type replica in a decentralized schema. The distributed transaction processing system comprises a transaction interface module, a transaction processing module and a transaction memory module, wherein the transaction interface module comprises an outward interface sub-module and a transaction preprocessing sub-module, the transaction processing module comprises a multi-type replica sub-module, a read transaction processing module, a replica group transaction state sub-module, a read request distribution sub-module, a replica repair sub-module, a multiversion concurrency control (MVCC) read sub-module, a local writing transaction processing sub-module, a local writing transaction paxos replica consistency sub-module, a local writing transaction commit sub-module, a global writing transaction processing sub-module, a main sub-transaction paxos replica consistency sub-module, a secondary sub-transaction paxos replica consistency sub-module and a global writing transaction commit sub-module. The distributed transaction processing system can solve the problems in an existing system that the reading and writing environment is limited, reading and writing availability can not be configured autonomously according to requirements of application, a global transaction depends on locking and cost is high.

Throughput is preserved in a distributed system while maintaining concurrency by pushing a commit wait period to client commit paths and to future readers. As opposed to servers performing commit waits, the servers assign timestamps, which are used to ensure that causality is preserved. When a server executes a transaction that writes data to a distributed database, the server acquires a user-level lock, and assigns the transaction a timestamp equal to a current time plus an interval corresponding to bounds of uncertainty of clocks in the distributed system. After assigning the timestamp, the server releases the user-level lock. Any client devices, before performing a read of the written data, must wait until the assigned timestamp is in the past.

The invention discloses a distributed transaction processing system which comprises a computing layer, a storage layer and a scheduling layer. The computing layer is used for receiving an SQL request of a client, converting the SQL request into a KV read-write request, transmitting the KV read-write request to the storage layer, and initiating two-stage transaction submission based on the SQL request, wherein the storage layer is used for providing a read-write interface for two-stage transaction submission initiated by the calculation layer and providing storage service; and the scheduling layer is used for maintaining meta-information of the whole cluster, including KV data distribution information and the like, providing global time service for the computing layer and the storage layer, and storing distributed information submitted in two stages of transactions. When the transaction is submitted, the transaction is submitted through a two-stage submission protocol, and a multi-version concurrency control mechanism is adopted, so that mutual blocking of reading and writing is avoided.

The invention discloses a multi-version concurrency control method and system based on virtual global clock synchronization. The method comprises the following steps: reading data; calculating and verifying a version timestamp; and writing the value written into a tuple in a set into a tuple storage position in a database, and updating the read timestamp and the write timestamp of the related tuple. The system comprises: a data reading module used for reading and copying all tuples involved in a current transaction; a timestamp calculation module used for calculating a submission timestamp ofthe current transaction according to the currently designed tuple read timestamp and write timestamp; a timestamp verification module used for verifying whether the timestamp obtained through calculation is valid or not; and a data writing module used for writing the successfully verified timestamp and the modified tuple into the database storage. The method has the following beneficial effects: the access tuples are checked, and the timestamp of each transaction is inertially calculated when the transaction is submitted, so that the limited sorting condition can be avoided.

The invention provides a metadata snapshot method and device, and relates to the technical field of cloud computing, in particular to the technical field of cloud storage. According to the specific implementation scheme, in response to the arrival of the current snapshot moment, generating a second basic version number of the current snapshot moment based on a first basic version number of a binary tree at the previous snapshot moment searched in a database; in the process from the current snapshot moment to the next snapshot moment, updating the search binary tree based on the updated metadata when the metadata in the database is updated, and generating an updated version number for the search binary tree updated each time; and in response to the arrival of the next snapshot moment, performing snapshot operation on the search binary trees corresponding to all version numbers generated between the current snapshot moment and the next snapshot moment to generate snapshot information. According to the method and the device, a multi-version concurrency control mode is adopted, data of previous versions are not affected while the data are modified, the accuracy and consistency of the data are guaranteed, and snapshot data can be obtained quickly and efficiently.

A system, method, and apparatus for optimizing database transactions in a multi-version concurrency control system are disclosed. The method includes receiving a pending transaction; acquiring predictive execution data associated with the pending transaction, wherein the predictive execution data includes a primary key value of target data required by the pending transaction and a value of the target data in the predictive execution; retrieving a current value of the target data from a target secondary database based on the primary key value of the target data; determining whether the current value of the target data affects service logic of the pending transaction based on whether the current value of the target data is different from the value of the target data; and recording in the target secondary database, the current value of the target data as version data corresponding to the pending transaction upon determining that the current value of the target data does not affect the service logic of the pending transaction.

Systems and methods provide multi-version concurrency control of database records with uncommitted transactions. The system and methods may include receiving a query regarding a transaction counter number. When a transaction header identifying data is updated with the assigned transaction counter number, the updated transaction header may be used in identifying data as an instance of multiversion concurrency control information by the received query. A key lookup may be performed, and when the key lookup encounters an uncommitted row for a transaction number, a corresponding transaction header identifying data to identify a data array element to determine whether the transaction number is committed, and determining the transaction counter number when the transaction number is committed. The transaction counter number may be stamped on the row version record when it is determined that the row is committed.

The invention discloses a web-based distributed concurrency control method, and the method is characterized by comprising a concurrency control server and a cluster server, wherein the concurrency control server is used for generating a token and sending the token to the cluster server; the cluster server is used for requesting for the token from the concurrency control server and returning back the token to the concurrency control server; and by setting a timer and a limit value, interaction between the cluster server and the concurrency control server is completed according to the limit value. The invention provides the web-based distributed concurrency control method, interactions in a concurrency control process is reduced through a pre-allocation method, performance of cluster concurrency control is promoted, and an unbalance phenomenon in cluster concurrency control is balanced through dynamic recovery and reallocation.

The embodiment of the invention provides a multi-version concurrency control method of a database and a database system, and relates to the technical field of databases. The method comprises the steps of, when a target transaction is started, generating a transaction ID for the target transaction, wherein the transaction ID comprises a thread number of a working thread starting the target transaction and a current local count value of the working thread; obtaining a logic snapshot of a database, and searching for a target visible version of the target transaction on a version chain corresponding to a to-be-operated target data record according to the logic snapshot; and reading or updating the data record on the version chain of the target data record based on the target visible version. According to the technical scheme, the isolation level of snapshot isolation can be provided, and the operation efficiency of transactions is improved. Compared with an existing method, the concurrency control method has lower overhead, the transaction processing capacity of the database can be remarkably improved, and the highest speed-up ratio can reach 4.24 times.

The invention discloses a multi-version concurrent storage method and device, relates to the field of artificial intelligence, in particular to a computer vision and deep learning technology, and is specifically applicable to an infrastructure scene. According to the specific implementation scheme, the method comprises the steps of firstly responding to a received update request of a user, obtaining a target object corresponding to the update request, and then determining a request identifier corresponding to the update request; finally determining an update version corresponding to the target object based on metadata of the target object and the request identifier, and updating a storage version of the target object. The brand-new multi-version concurrent control storage method is realized, and different versions can be stored based on an update request of a user.

The invention provides a method and device for achieving audit data storage under multi-version concurrency control, and the method comprises the steps: employing a fixed transaction number as a tuplehead to be spliced with to-be-inserted audit data under the condition that the to-be-inserted audit data exists, and forming an audit data tuple; wherein the fixed transaction number is a submitted transaction number; and storing the audit data tuples in an audit data table in a database. The transaction number does not need to be applied as in the prior art, so that all fixed connections can seethe audit tuples marked by the fixed transaction number, the corresponding audit data tuples can also be quickly inserted into the audit data table of the database, and the transaction does not needto be frequently submitted as in the prior art.