About: Multiple granularity locking is a research topic. Over the lifetime, 8 publications have been published within this topic receiving 25 citations.
TL;DR: A direct graph model called the holder/wire-transaction waited-by graph (H/W-TWBG) is introduced to capture the precise status of systems in terms of deadlock to address the deadlock problem in sequential transaction processing.
Abstract: The authors address the deadlock problem in sequential transaction processing where the strict two-phase locking and the multiple granularity locking protocol with five lock modes are used. The scheduling policy honors lock requests in a first-in-first-out basis except for lock conversions. As a basic tool, a direct graph model called the holder/wire-transaction waited-by graph (H/W-TWBG) is introduced to capture the precise status of systems in terms of deadlock. The properties of H/W-TWBG are presented. Based on H/W-TWBG, the identification principles of the victim candidates are established in a deadlock cycle, and a periodic deadlock detection and resolution algorithm which has a reasonable time and storage complexity is preserved. One important feature of the deadlock resolution scheme is that some deadlocks can be resolved without aborting any transaction. >
TL;DR: This paper develops a new multiple granularity locking protocol specifically for CSCW, which provides versionable object locking, guarantees conflict freedom, prevents update loss and ensures greater concurrency than other locking protocols.
Abstract: Locking is a commonly used mechanism in Computer Supported Cooperative Work (CSCW) for coordinating concurrent users and implementing various styles of information sharing. Part of a sharing style's definition is the granularity of shared objects. Since this varies among different applications, it is desirable to provide a range of different lock granules. In this paper, I first examine existing multiple granularity locking protocols and I evaluate their suitability for CSCW. My main contribution is the development of a new such protocol specifically for CSCW, which provides versionable object locking, guarantees conflict freedom, prevents update loss and ensures greater concurrency than other locking protocols. Its suitability for CSCW is justified by a comparative performance analysis.
TL;DR: An efficient locking technique is proposed by integrating multi version with multiple granularity in the hierarchical structure which allows several requesting transactions to be executed in parallel by serving them an appropriate version to read while the write operation of some other transaction is in progress.
Abstract: Transaction performs lock and unlocking operations on the data items required in its execution. These operations on the data items are important in order to maintain consistency of the database as the data may be accessed by concurrently executing transactions. Designing locking and unlocking mechanism on the data items it involves fine tuning which involves the following factors a) level of granularity b) appropriate version to be locked c) supporting the compatibility mode. In this paper, two locking protocols are covered namely Multi version and Multi granularity locking protocol. The Multiple granularity locking protocol specify at which level locks can be applied, if the lock is applied on exact level of database, then definitely system performance can be improved and basically explore how can be increase the concurrency of MGL locking protocol by considering suitable example. An efficient locking technique is proposed by integrating multi version with multiple granularity in the hierarchical structure. This allows several requesting transactions to be executed in parallel by serving them an appropriate version to read while the write operation of some other transaction is in progress.
TL;DR: A multiple-granularity locking protocol, in which the locks are optional and they convey the meanings of access mode, locking scope, and locking effect, which allows synchronous collaboration by guaranteeing a conflict-free environment and avoiding update loss while it also supports version control.
Abstract: In distributed Web authoring, shared documents can be accessed concurrently by multiple authors who must be coordinated to avoid conflicts. The current Web standard for distributed authoring and versioning uses a two-phase locking to coordinate concurrent access. As the degree to which authors work concurrently may vary though among cooperative sessions, it is necessary to extend the aforementioned standard so as to support a multitude of lock granularity levels. In this paper, we first examine related protocols from the database literature, and then, we comment on their suitability for distributed authoring in the World Wide Web. Our main contribution is a multiple-granularity locking protocol, in which the locks are optional and they convey the meanings of access mode, locking scope, and locking effect. This protocol allows synchronous collaboration by guaranteeing a conflict-free environment and avoiding update loss while it also supports version control. Specifically, by identifying and timestamping object versions, the protocol preserves author intention and operation causality, which were possible so far with operational transformation only. The protocol's efficiency, finally, is demonstrated by a real test with human users and evaluated with simulation experiments, which reveal significant advantages over other protocols of this kind.
TL;DR: In this paper, a multi-version granular structure is proposed to support simultaneous execution of share and update transactions by providing a suitable version at each requesting mode, and a detailed description of different cases that exhibits their work by converting some of the non-compatible modes into the compatible mode of operations.
Abstract: The paper presents a multi-version granular structure, an extension of the multiple granularity locking mechanisms. The existing hierarchical locking carries out its operation at an either coarse or fine granular level that does not support enhanced concurrency on different locking modes. The core problem lies in the locking waiting time that a read or a write operation undergoes when the conflict occurs for the same set of resources in the given hierarchical structure. To overcome such limitations, we proposed multi-versioning to support simultaneous execution of share and update transactions by providing a suitable version at each requesting mode. We further cover a detailed description of different cases that exhibits our work by converting some of the non-compatible modes into the compatible mode of operations.