Method for a caching mechanism for a central synchronization server

• US 7,539,778 B2
• Filed: 06/25/2002
• Issued: 05/26/2009
• Est. Priority Date: 06/30/2001
• Status: Expired due to Term

• Current Assignee: King.com, Ltd (Microsoft Corporation)
• Original Assignee: International Business Machines Corporation
• Inventors: Hepper, Stefan, Hansmann, Uwe, Heintzman, Douglas, Stober, Thomas, Purakayastha, Apratim
• Primary Examiner(s): Pwu; Jeffrey
• Assistant Examiner(s): Taha; Shaq
• Application Number:
  US10/480,943
• Publication Number:
  US 20040153576A1
• Time in Patent Office: 2,527 Days
• Field of Search: 709/248, 709/250, 709/224, 709/225, 709/201
• US Class Current: 709/248
• CPC Class Codes: G06F 16/9574 of access to content, e.g. by caching

Abstract

The present invention discloses a new synchronization architecture for synchronization of data between different clients by using a central synchronization server linked to a Back End data store which additionally provides a cache for permanently buffering incoming updates into a permanent store by assigning an unique cache identifier (ID). Write conflicts between the synchronization server writing new entries to the cache and updates replicated from backend to cache are solved using a blocking mechanism based on the cache IDs, so that the backend updates are blocked as long as incoming updates from the clients having the same cache ID are not completely written into the cache during a synchronization session. The present invention is preferably suited for a synchronization architecture having a high number of clients connected with the central synchronization server as blocking of the Back End data store, and the connection and the transport to the Back End data store are minimized.

Claims

1. A method for replication an update from a Back End data store to a cache comprising the steps of:

receiving an update from said Back End data store by a central synchronization server containing update data, time stamp, and backend ID,

translating backend data format of a backend update into cache data format,

mapping backend ID to cache ID by using a mapping table, and

buffering of said update from said Back End data store into said cache only when the time stamp is newer than the time stamp of an update buffered in said cache having the same cache ID;

wherein buffering of updates from said Back End data store into said cache are blocked as long as updates from clients having the same cache IDs are not completely written into said cache during a synchronization session.

2. Method according to claim 1, comprising the further step:

generating a new cache ID when mapping backend ID to cache ID is not successful, and

inserting said new cache ID into the mapping table.