Resilient Message Passing Applications

• US 20120159505A1
• Filed: 12/20/2010
• Published: 06/21/2012
• Est. Priority Date: 12/20/2010
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: MURA, Marco, DALAL, Danny
• Granted Patent: US 9,384,052 B2
• Time in Patent Office: Days
• US Class Current: 718/104
• CPC Class Codes: G06F 2209/547 Messaging middleware
• CPC Class Codes: G06F 9/5027 the resource being a machine, e.g. CPUs, Servers, Terminals

Abstract

A message passing system may execute a parallel application on multiple compute nodes. Each compute node may perform a single workload on at least two physical computing resources. Messages may be passed from one compute node to another, and each physical computing resource assigned to a compute node may receive and process the messages. In some embodiments, the compute nodes may be virtualized so that a message passing system may only detect a single compute node and not the multiple underlying physical computing resources.

Claims

1. A system comprising:

a plurality of compute nodes;

an application execution environment that exposes said computing nodes to an application;

for each of said compute nodes, at least two physical computing resources, each of said physical computing resources that perform identical computing tasks for said compute node;

said application execution environment that further passing messages between compute nodes, said messages being generated by an application executing on said plurality of compute nodes.

2. The system of claim 1, said application execution environment that further:

detects that a first physical computing resource is unavailable; and

removes said first physical computing resource from receiving messages.

3. The system of claim 1, a first compute node being a virtualized compute node.

4. The system of claim 3, said virtualized compute node having at least two of said physical computing resources assigned to said virtualized compute node.

5. The system of claim 4, said virtualized compute node having a computing resource manager that:

receives a first message for said compute node and passes said first message to identical application threads executing on said at least two physical computing resources.

6. The system of claim 5, said virtualized computing resource manager that further:

receives a second message from a first of said at least two physical computing resources and passes said second message to a second compute node, said second compute node comprising a plurality of physical computing resources; and

receives a third message from a second of said at least two physical computing resources and does not transmit said third message, said third message having a work payload similar to a work payload of said second message.

7. The system of claim 6 having at least one virtualized compute node and at least one non-virtualized compute node.

8. The system of claim 1 further comprising:

an application set up mechanism that;

receives an application to execute, said application having a plurality of execution threads, each of said execution thread being operable on a different computing node, said application further using a message passing system to pass messages from one execution thread to another;

determines a number of computing nodes to be assigned to said application;

for each of said computing resources, identifies at least two physical computing resources to execute an execution thread and configures a computing resource manager for said at least two physical computing resources; and

causes said execution threads to be started on each of said computing resources.

9. The system of claim 8, said computing resource manager being executed on each of said physical computing resources.

10. The system of claim 8, said computing resource manager being executed on a first computing resource and not on one of said at least two physical computing resources for at least one of said computing resource.

11. The system of claim 10, said computing resource manager being executed on a physical computing resource that is not one of said at least two physical computing resources.

12. A method for operating a first compute node in a message passing environment for executing a parallel processing application, said method comprising:

receiving a first message from a second compute node, said message being addressed to said first compute node;

identifying a plurality of physical computing resources, each of said plurality of computing resources executing a first execution thread for said parallel processing application;

transferring said first message to each of said physical computing resources;

receiving a second message from a first physical computing resource, said second message having an address for a third compute node; and

transferring said second message to said third compute node.

13. The method of claim 12 further comprising:

receiving a third message from a second physical computing resource, said third message having an address for said third compute node;

determining that said third message and said second message are similar; and

not transmitting said third message.

14. The method of claim 12, each of said compute nodes being virtualized compute nodes.

15. A system comprising:

multiple physical computing resources;

an application execution environment that;

exposes a first set of compute nodes to an application, each of said compute nodes executing a workflow portion of said application;

transmits communications from a first compute node to a second compute node;

for each of said compute nodes, provides at least two computing resources, each of said computing resources executing a similar workflow, and forwards communications received by a compute node to each of said computing resources associated with said compute node;

determines that a first message from said first compute node to said second compute node has not been received and retransmits said first message.

16. The system of claim 15, at least one of said compute nodes being a virtualized compute node.

17. The system of claim 15, one of said physical computing resources being designated a primary physical computing resource and a second of said physical computing resources being designated a duplicate physical computing resource.

18. The system of claim 15, said application execution environment that further:

detects that a third physical computing resource has become unavailable and removes said third physical computing resource from receiving communications.

19. The system of claim 15, said application execution environment that further:

queries each of said physical computing resources to determine whether or not each physical computing resource received a message, said application execution environment that further sends said message when said message had not been received.

20. The system of claim 19, said application execution environment that further:

detects a failed communication by setting a timer at a first communication attempt for a first message and detecting expiration of said timer prior to detecting a successful receipt of said first message.