Browse Prior Art Database

Providing A Multi-Process Apply Using A Shared, Consumer/Producer Work Queue

IP.com Disclosure Number: IPCOM000237945D
Publication Date: 2014-Jul-23
Document File: 2 page(s) / 58K

Publishing Venue

The IP.com Prior Art Database

Abstract

A method and system is disclosed for providing a multi-process apply using a shared, consumer/producer work queue.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

Page 01 of 2

/Producer Work Queue

Producer Work Queue

Disclosed is a method and system for providing a multi-process apply using a shared, consumer/producer work queue.

In a high volume, homogeneous Database management System (DBMS) replication environment, where the Recovery Point Objective (RPO) and Recovery Time Objective (RTO) is contained to the second or sub-second range, a particular emphasis is placed on performance and efficiency. The elimination of unnecessary in-stream processing is an iterative process, where the ultimate bottleneck is generally found in the application of captured changes to a target DBMS. Most replication products, therefore, implement a high-performance apply server using a dependency analysis node to identify independent transactions and dependent transactions. The transactions are identified such that a pool of parallel writer nodes apply the independent transactions against the target DBMS in an efficient manner. In a typical node configuration for replication apply engines, the nodes are usually implemented as threads in a single monolithic process, or an address space.

The disclosed method and system creates an extended consumer producer model store or set of stores that employs lock-less serialization techniques. According to embodiments of the present invention, the lockless serialization techniques are shared between an apply server address space and a transaction manager address space. A single control pipe is used for basic coordination; while all bulk data movement is achieved using a consumer producer lock-less store. Dependency analysis nodes reside in the apply server address space and writer nodes reside in the transaction manager address space creating a distributed processing model, for a replication apply function, that is highly efficient. A high performance, concurrent, B+ tree enables a consumer producer environment between two foreign address spaces by using backing stores in shared memory.

Consumer

Consumer/

Providing A Multi - -Process Apply Using A Shared

Process Apply Using A Shared , ,

A topology of the method and system is illustrated in the below figure, according to embodiments of the present invention.

Figure

As shown in the figure, the apply server receives inbound tran...