Efficient state tracking for clusters

• US 9,875,159 B2
• Filed: 03/01/2013
• Issued: 01/23/2018
• Est. Priority Date: 01/16/2012
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Keller, Michael
• Primary Examiner(s): Aponte, Francisco
• Application Number:
  US13/782,221
• Publication Number:
  US 20130185379A1
• Time in Patent Office: 1,789 Days
• Field of Search: 709213
• CPC Class Codes: G06F 11/1438 Restarting or rejuvenating
• CPC Class Codes: G06F 2201/84 Using snapshots, i.e. a logical point-in-time copy of the data

Abstract

Exemplary method, system, and computer program product embodiments for efficient state tracking for clusters are provided. In one embodiment, by way of example only, in a distributed shared memory architecture, an asynchronous calculation of deltas and the views is performed while concurrently receiving client requests and concurrently tracking the client requests times. The results of the asynchronous calculation may be applied to each of the client requests that are competing for data of the same concurrency during a certain period with currently executing client requests. Additional system and computer program product embodiments are disclosed and provide related advantages.

Claims

1. A method for efficient state tracking for clusters by a processor device in a distributed shared memory architecture, the method comprising:

performing an asynchronous calculation of deltas while concurrently receiving client requests and concurrently tracking client requests times;

responding to each of the client requests for data of the same concurrency during a certain period with currently executing client requests with updated views based upon results of the asynchronous calculation; and

bounding a latency for the client requests by a time necessitated for the asynchronous calculation of at least two of the deltas;

wherein a first state snapshot is atomically taken while simultaneously calculating the at least two of the deltas, and each of the client requests received during the certain period are served with the updated views of the asynchronously calculated at least two of the deltas, and wherein upon the time necessitated for the asynchronous calculation of the at least two of the deltas expires, the first state snapshot is disregarded and a second state snapshot is atomically taken for the asynchronous calculation of the deltas for the updating of the views of subsequent client requests.

2. The method of claim 1, further including, in conjunction with the responding, executing each of the client requests occurring after the certain period on the updated views, wherein all deltas and views are updated.

3. The method of claim 1, further including, in conjunction with the performing, folding the deltas in response to receiving client requests.

4. The method of claim 1, further including, performing the asynchronous calculation for the deltas by a direct comparison between one of a plurality of snapshot states and the views, wherein the plurality of snapshot states include at least one of a most recent snapshot and a previous snapshot.

5. The method of claim 1, further including performing the asynchronous calculation on demand.

6. The method of claim 1, further including performing the asynchronous calculation only upon receipt of the client requests for the deltas calculated since a specified client request.