Optimistic concurrency control: Difference between revisions

Revision as of 04:38, 17 November 2008

Latest revision as of 14:10, 30 April 2025

(174 intermediate revisions by more than 100 users not shown)

Line 1:

{{Short description|Concurrency control method}}

{{Unreferenced|date=February 2007}}

'''Optimistic concurrency control''' ('''OCC'''), also known as '''optimistic locking''', is a [[Non-lock concurrency control|non-locking concurrency control]] method applied to transactional systems such as [[relational database management systems]] and [[software transactional memory]]. OCC assumes that multiple transactions can frequently complete without interfering with each other. While running, transactions use data resources without acquiring locks on those resources. Before committing, each transaction verifies that no other transaction has modified the data it has read. If the check reveals conflicting modifications, the committing transaction rolls back and can be restarted.<ref>{{cite book | title = Expert One-on-One J2EE Design and Development | first = Rohit | last = Johnson | publisher = Wrox Press | year = 2003 | isbn = 978-0-7645-4385-2 | chapter = Common Data Access Issues | chapter-url = http://learning.infocollections.com/ebook%202/Computer/Programming/Java/Expert_One-on-One_J2EE_Design_and_Development/6266final/LiB0080.html | archive-url = https://web.archive.org/web/20111008203709/http://learning.infocollections.com/ebook%202/Computer/Programming/Java/Expert_One-on-One_J2EE_Design_and_Development/6266final/LiB0080.html | archive-date = 8 October 2011}}</ref> Optimistic concurrency control was first proposed in 1979 by [[H. T. Kung]] and John T. Robinson.<ref name="KungRobinson1981">{{Cite news| title = On Optimistic Methods for Concurrency Control | first = J. T. Robinson | last = H. T. Kung | publisher = ACM Transactions on Database Systems | year = 1981 | url = https://apps.dtic.mil/dtic/tr/fulltext/u2/a081452.pdf| archive-url = https://web.archive.org/web/20190831230313/https://apps.dtic.mil/dtic/tr/fulltext/u2/a081452.pdf| url-status = live| archive-date = August 31, 2019}}</ref>

OCC is generally used in environments with low [[Block contention|data contention]]. When conflicts are rare, transactions can complete without the expense of managing locks and without having transactions wait for other transactions' locks to clear, leading to higher throughput than other concurrency control methods. However, if contention for data resources is frequent, the cost of repeatedly restarting transactions hurts performance significantly, in which case other [[concurrency control]] methods may be better suited. However, locking-based ("pessimistic") methods also can deliver poor performance because locking can drastically limit effective concurrency even when deadlocks are avoided.

In [[computer science]], in the field of [[database]]s, '''optimistic concurrency control''', (OCC) is a [[concurrency control]] method used in [[relational database]]s without using [[Lock (computer science)|lock]]ing. It is commonly referred to as '''optimistic locking''', a reference to the non-exclusive locks that are created on the database.

[[Category:== Phases of optimistic concurrency control]] ==

Optimistic concurrency control is based on the assumption that most [[database transaction]]s don't conflict with other transactions, allowing OCC to be as permissive as possible in allowing transactions to execute.

Optimistic concurrency control transactions involve these phases:<ref name="KungRobinson1981" />

There*'''Begin''': areRecord threea phasestimestamp inmarking an OCCthe transaction:'s beginning.

*'''Modify''': Read database values, and tentatively write changes.

*'''Validate''': Check whether other transactions have modified data that this transaction has used (read or written). This includes transactions that completed after this transaction's start time, and optionally, transactions that are still active at validation time.

<ol>

*'''Commit/Rollback''': If there is no conflict, make all changes take effect. If there is a conflict, resolve it, typically by aborting the transaction, although other resolution schemes are possible. Care must be taken to avoid a [[time-of-check to time-of-use]] bug, particularly if this phase and the previous one are not performed as a single [[linearizability|atomic]] operation.

<li>'''Read''': The client reads values from the database, storing them to a private sandbox or cache that the client can then edit.</li>

<li>'''Validate''': When the client has completed editing of the values in its sandbox or cache, it initiates the storage of the changes back to the database. During validation, an algorithm checks if the changes to the data would conflict with either

* already-committed transactions in the case of ''backward validation schemes'', or

* currently executing transactions in the case of ''forward validation schemes''.

If a conflict exists, a conflict resolution algorithm must be used to resolve the conflict somehow (ideally by minimizing the number of changes made by the user) or, as a last resort, the entire transaction can be aborted (resulting in the loss of all changes made by the user).

</li>

<li>'''Write''': If there is no possibility of conflict, the transaction commits.</li>

</ol>

Optimistic concurrency is generally used in environments with a low contention for data. When conflicts are rare, validation can be done efficiently, leading to higher throughput than other concurrency control methods.{{Fact|date=February 2007}} However, if conflicts happen often, the cost of repeatedly restarting transactions hurts performance significantly; other [[non-lock concurrency control]] methods have better performance under these conditions.

==Web usage==

The [[Stateless server|stateless]] nature of [[HTTP]] makes locking infeasible for web user interfaces.{{Fact|date=February 2007}}It It'sis common for a user to start editing a record, then leave without following a "cancel" or "logout" link. If locking is used, other users who attempt to edit the same record must wait until the first user's lock expirestimes out.

[[HTTP]] does provide a form of built-in OCC. The response to an initial GET request can include an [[HTTP ETag|ETag]] for subsequent PUT requests to use in the If-Match header. Any PUT requests with an out-of-date ETag in the If-Match header can then be rejected.<ref>{{cite web | url = http://www.w3.org/1999/04/Editing/ | title = Editing the Web - Detecting the Lost Update Problem Using Unreserved Checkout | work = W3C Note | date = 10 May 1999}}</ref>

The [[Stateless server|stateless]] nature of [[HTTP]] makes locking infeasible for web user interfaces.{{Fact|date=February 2007}} It's common for a user to start editing a record, then leave without following a "cancel" or "logout" link. If locking is used, other users who attempt to edit the same record must wait until the first user's lock expires.

Some database management systems offer OCC isnatively, without requiring aspecial naturalapplication choicecode. ItFor isothers, simplethe application tocan implement andan avoidsOCC unnecessarylayer outside of the database, and avoid waiting or silently overwrittenoverwriting records. TypicallyIn such cases, the [[Form (web)|form]] presented to the usermay includesinclude a hidden field with the record's original content, a timestamp, a sequence number, or an opaque token. On submit, this is compared against the database. If it differs, the conflict resolution algorithm is invoked.

===Examples===

* [[MediaWiki]]'s edit pages use OCC. The conflict resolution algorithm is described <ref>[[w:Help:Edit conflict|hereHelp:Edit conflict]].<!-- Use interwiki syntax so that mirrors can at least have a chance to pick it up --></ref>

* [[Bugzilla]] uses OCC; conflicts[[edit conflict]]s are called "mid-air collisions".<ref>{{cite web [http| url = https://wiki.mozilla.org/Bugzilla:FAQ:Administrative_Questions#Does_Bugzilla_provide_record_locking_when_there_is_simultaneous_access_to_the_same_bug.]3F_Does_the_second_person_get_a_notice_that_the_bug_is_in_use_or_how_are_they_notified.3F | title = Bugzilla: FAQ: Administrative Questions | work = MozillaWiki | date = 11 April 2012}}</ref>

* The [[Ruby on Rails]] framework has an API for OCC.<ref>{{cite web | url = [http://api.rubyonrails.comorg/classes/ActiveRecord/Locking/Optimistic.html] | title = Module ActiveRecord::Locking | work = Rails Framework Documentation}}</ref>

* The [[Grails (framework)|Grails]] framework uses OCC in its default conventions.<ref>{{cite web | url = http://grails.org/doc/1.0.x/guide/single.html#5.3.5%20Pessimistic%20and%20Optimistic%20Locking | title = Object Relational Mapping (GORM) | work = Grails Framework Documentation | url-status = dead | archive-url = https://web.archive.org/web/20140815173309/http://grails.org/doc/1.0.x/guide/single.html#5.3.5%20Pessimistic%20and%20Optimistic%20Locking | archive-date = 2014-08-15 }}</ref>

* The [[GT.M]] database engine uses OCC for managing transactions<ref>{{cite web | url = http://tinco.pair.com/bhaskar/gtm/doc/books/pg/UNIX_manual/ch05s17.html | title = Transaction Processing | work = GT.M Programmers Guide UNIX Edition}}</ref> (even single updates are treated as mini-transactions).

* [[Microsoft]]'s [[Entity Framework]] (including Code-First) has built-in support for OCC based on a binary timestamp value.<ref>{{cite web | url = https://learn.microsoft.com/en-us/ef/core/saving/concurrency?tabs=data-annotations#optimistic-concurrency | title = Handling Concurrency Conflicts | work = Entity Framework documentation hub | date = 5 July 2023}}</ref>

* Most [[revision control]] systems support the "merge" model for concurrency, which is OCC.{{cn|date=February 2023}}

* [[Mimer SQL]] is a [[DBMS]] that only implements optimistic concurrency control.<ref>{{cite web | url = https://developer.mimer.com/article/transaction-concurrency-optimistic-concurrency-control/ | title = Transaction Concurrency - Optimistic Concurrency Control | work = Mimer Developers - Features | access-date = 22 Dec 2023}}</ref>

* [[Google App Engine]] data store uses OCC.<ref>{{cite web | url = http://code.google.com/appengine/docs/whatisgoogleappengine.html | title = The Datastore | work = What Is Google App Engine? | date = 27 August 2010}}</ref>

* The [[Apache Solr]] search engine supports OCC via the {{Mono|_version_}} field.<ref>{{cite web|url=https://lucene.apache.org/solr/guide/6_6/updating-parts-of-documents.html|title=Updating Parts of Documents|access-date=2018-06-28}}</ref>

* The [[Elasticsearch]] search engine updates its documents via OCC. Each version of a document is assigned a sequence number, and newer versions receive higher sequence numbers. As changes to a document arrive asynchronously, the software can use the sequence number to avoid overriding a newer version with an old one.<ref>{{cite web | title=Optimistic concurrency control | website=Elastic | url=https://www.elastic.co/guide/en/elasticsearch/reference/current/optimistic-concurrency-control.html | access-date=2024-02-05}}</ref>

* [[CouchDB]] implements OCC through document revisions.<ref>{{cite web | title=Technical Overview | website=Apache CouchDB Documentation | url=https://docs.couchdb.org/en/stable/intro/overview.html | access-date=2024-02-06}}</ref>

* The [[MonetDB]] [[Column-oriented DBMS|column-oriented]] [[database management system]]'s transaction management scheme is based on OCC.<ref>{{cite web | url = http://www.monetdb.org/Documentation/Manuals/SQLreference/Transactions | title = Transactions - MonetDB | date = 16 January 2013}}</ref>

* Most implementations of [[software transactional memory]] use OCC.{{citation needed|reason=Claim initially said optimistic locking, now says OCC, both claims unsourced|date=March 2019}}

* [[Redis]] provides OCC through WATCH command.<ref>{{cite web | url = http://redis.io/topics/transactions | title = Transactions in Redis }}</ref>

* [[Firebird (database server)|Firebird]] uses [[Multiversion concurrency control|Multi-generational architecture]] as an implementation of OCC for data management.{{citation needed|date = November 2020}}

* [[Amazon DynamoDB|DynamoDB]] uses conditional update as an implementation of OCC.<ref>{{cite web | url = https://docs.aws.amazon.com/amazondynamodb/latest/developerguide/WorkingWithItems.html#WorkingWithItems.ConditionalUpdate| title = Working with Items and Attributes - Conditional Writes| access-date = 2 November 2020}}</ref>

* [[Kubernetes]] uses OCC when updating resources.<ref>{{cite web | url = https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.18/#resource-operations-update | title = API Overview - Resource Operations| access-date = 3 November 2020}}</ref>

* [[YugabyteDB]] is a cloud-native database that primarily uses OCC.<ref>{{Cite web|last=Yugabyte|first=Team|title=Explicit locking {{!}} YugabyteDB Docs|url=https://docs.yugabyte.com/latest/architecture/transactions/explicit-locking/|access-date=2022-01-04|website=docs.yugabyte.com|language=en-us}}</ref>

* [[Firestore]] is a NoSQL database by [[Firebase]] that uses OCC in its transactions.

* [[Apache Iceberg]] uses OCC to update tables and run maintenance operations on them.

==See also==

* [[MediaWiki]]'s edit pages use OCC. The conflict resolution algorithm is described [[w:Help:Edit conflict|here]].<!-- Use interwiki syntax so that mirrors can at least have a chance to pick it up -->

* [[Bugzilla]] uses OCC; conflicts are called "mid-air collisions". [http://wiki.mozilla.org/Bugzilla:FAQ:Administrative_Questions#Does_Bugzilla_provide_record_locking_when_there_is_simultaneous_access_to_the_same_bug.]

*[[Server Message Block#Opportunistic Lockinglocking]]

* The [[Ruby on Rails]] framework has an API for OCC. [http://api.rubyonrails.com/classes/ActiveRecord/Locking.html]

* Most [[revision control]] systems support the "merge" model for concurrency, which is OCC.

==References==

{{Reflist}}

<div class="references">

*{{cite journal|title=On optimistic methods for concurrency control|journal=ACM Transactions on Database Systems|date=June 1981|first=H. T.|last=Kung|coauthors=John T. Robinson|volume=6|issue=2|pages=213-226|id={{doi|10.1145/319566.319567}}}}

</div>

==See also==

*[[Opportunistic Locking]]

==External links==

[[Category:concurrency control]]

*{{cite journal|title=On optimistic methods for concurrency control|journal=ACM Transactions on Database Systems|date=June 1981|first=H. T.|last=Kung|coauthorsauthor2=John T. Robinson|volume=6|issue=2|pages=213-226213–226|id={{doi|=10.1145/319566.319567}}|citeseerx=10.1.1.101.8988|s2cid=61600099 }}

* Enterprise JavaBeans, 3.0, By Bill Burke, Richard Monson-Haefel, Chapter 16. Transactions, Section 16.3.5. Optimistic Locking, Publisher: O'Reilly, Pub Date: May 16, 2006, Print {{ISBN|0-596-00978-X}},

* {{cite conference | first = Andreas | last = Hollmann | title = Multi-Isolation: Virtues and Limitations | book-title = Multi-Isolation (what is between pessimistic and optimistic locking) | pages = 8 | publisher = Happy-Guys Software GbR | date = May 2009 | ___location = 01069 Gutzkovstr. 30/F301.2, Dresden | url = http://www.andrej-hollmann.de/images/stories/informatik/multi-isolation-part-1.pdf | access-date = 2013-05-16 }}{{dead link|date=March 2018 |bot=InternetArchiveBot |fix-attempted=yes }}

{{DEFAULTSORT:Optimistic Concurrency Control}}

[[he:בקרת מקביליות אופטימית]]

<!-- Categories -->

[[ja:楽観的並行性制御]]

[[Category:Concurrency control]]

[[Category:Concurrency control algorithms]]

[[Category:Transaction processing]]