Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Method and Apparatus for Serialization of Data Synchronization with respect to new operations

IP.com Disclosure Number: IPCOM000019340D
Original Publication Date: 2003-Sep-11
Included in the Prior Art Database: 2003-Sep-11
Document File: 2 page(s) / 43K

Publishing Venue

IBM

Abstract

The following describes a programming method that allows the synchronization of data across a remote geographically mirrored set of systems to be performed concurrently with application updates to the data on the source system (production copy).

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 52% of the total text.

Page 1 of 2

  Method and Apparatus for Serialization of Data Synchronization with respect to new operations

  With remote logical geographic mirroring, at times, the data between the copy on the source system (production copy) and the copy on the target system (mirror copy) may become out of sync. When this occurs, synchronization processing is required in order to get the two copies back in sync. This synchronization processing can be very time consuming. In addition, care needs to be taken in order to ensure that no updates to the data occur while the data is being synchronized in order to prevent the loss of the updates and other data integrity concerns. The user may be hesitant to use the support if their entire set of data must be off-line while the data is being synchronized.

The solution is to divide up the entire virtual address range which is to be synchronized into sub ranges in order to keep track of which operations are to only be performed on the production copy (since they will be synchronized at a later time), which operation are to be mirrored over to the mirror copy (since they have already been synchronized), and which operations must block and wait momentarily (since they are currently in the processes of being synchronized over to the mirror copy). With this mechanism, the end user is allowed to access and update their data while synchronization of the data is occurring in the background. This mechanism also allows for the potential adjusting of the priority of the synchronization processing which is occurring in the background depending on the relative importance to the end user of being synchronized versus current use of the data. First, the data which is to be synchronized can first be broken into two parts which each require synchronization, the allocated storage for the data along with its various attributes (i.e., the data extent information) and the actual data within the extent itself. Synchronization processing will divide the overall virtual address range into five distinct ranges. As synchronization processing begins, all five virtual address ranges are set to the beginning of the overall virtual address range and they proceed through the virtual address range as synchronization processing is performed until all of the ranges meet at the other end of the virtual address range. These ranges are as follows: The first range is of data and extent information which is yet to be processed. Any new operations (of any type) which occur within this range only need to be performed on the production copy since they will be synchronized over to the mirror copy later on as synchronization processing proceeds. When...