Distributed computing management

• US 8,832,156 B2
• Filed: 06/15/2009
• Issued: 09/09/2014
• Est. Priority Date: 06/15/2009
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Heil, Andreas, Calsyn, Martin, Brandle, Alexander, Lyutsarev, Vassily
• Primary Examiner(s): Alam, Shahid
• Assistant Examiner(s): Sultana, Nargis
• Application Number:
  US12/484,773
• Publication Number:
  US 20100318565A1
• Time in Patent Office: 1,912 Days
• Field of Search: 707/798, 707/803
• US Class Current: 707/798
• CPC Class Codes: G06F 11/323 Visualisation of programs or trace data
• CPC Class Codes: G06F 16/9024 Graphs; Linked lists G06F16/9027 takes precedence
• CPC Class Codes: G06F 9/5066 Algorithms for mapping a plurality of inter-dependent sub-tasks onto a plurality of physical CPUs mappping at compile time, see G06F8/451
• CPC Class Codes: H04L 41/22 comprising specially adapted graphical user interfaces [GUI]

Abstract

Management of distributed computing systems is required, for example, to carry out activities using shared resources such as computational tasks, in-silico experiments and other tasks. In an embodiment a distributed computing graph represents processes executing in the distributed computing system in order to carry out the activity and this graph is available to entities in the distributed computing system. For example, a user interface at an entity in the distributed computing system enables a user to view a local representation of the distributed computing graph and to control the activity using that local representation which maps dynamically to and from the underlying distributed computing graph. In examples, the local representation of the distributed computing graph enables live control of the activity without the need for compilation. In examples the distributed computing graph is shared through the distributed computing system using peer-to-peer distribution or using a replication mechanism.

Claims

1. A method at an entity in a distributed computing system, the method comprising:

providing an input arranged to connect to the distributed computing system and access a distributed computing graph associated with an activity to be carried out in the distributed computing system, the distributed computing graph representing currently executing processes in the distributed computing system;

arranging a processor to provide a dynamic mapping between the distributed computing graph and a data structure at the entity holding a local view of the distributed computing graph such that changes in the distributed computing graph are reflected in the local view; and

providing a user interface enabling control of the activity by modifying the local view based on input received at the user interface which modifies the local view, the local view indicating which processes represented by the distributed computing graph are executed at the entity and which processes are executed remote of the entity,

the method further comprising;

receiving input at the user interface which modifies the local view and wherein the method comprises using the dynamic mapping between the local view and the distributed computing graph to update the distributed computing graph according to the modified local view in order to control the activity;

receiving user input specifying a second local view, attempting to connect to a distributed computing graph corresponding to the second local view and if successful updating the second local view according to the corresponding distributed computing graph;

or

receiving user input specifying the second local view, attempting to connect to the distributed computing graph corresponding to the second local view and if unsuccessful starting a new distributed computing graph corresponding to the second local view.

2. A method as claimed in claim 1 wherein the step of accessing the distributed computing graph comprises accessing a plurality of nodes connected by links where at least some of the nodes represent any of computational processes and data sources at one or more entities in the distributed computing system.

3. A method as claimed in claim 1 which comprises arranging the processor to provide the dynamic mapping such that changes in the local view are reflected in the distributed computing graph.

4. A method as claimed in claim 1 which comprises providing a peer-to-peer content distribution mechanism at the entity for distribution of the distributed computing graph within the distributed computing system.

5. A method as claimed in claim 1 which comprises arranging the input to access the distributed computing graph using a replication mechanism for replicating the distributed computing graph at the entities in the distributed computing system.

6. A method as claimed in claim 1 which further comprises receiving input at the user interface which modifies the local view and wherein the method comprises using the dynamic mapping between the local view and the distributed computing graph to update the distributed computing graph according to the modified local view in order to control the activity.

7. A method as claimed in claim 1 which further comprises receiving user input specifying the second local view;

attempting to connect to the distributed computing graph corresponding to the second local view and if successful updating the second local view according to the corresponding distributed computing graph.

8. A method as claimed in claim 1 which further comprises receiving user input specifying the second local view;

attempting to connect to the distributed computing graph corresponding to the second local view and if unsuccessful starting the new distributed computing graph corresponding to the second local view.

9. A method at an entity in a distributed computing system having a plurality of entities in communication with one another, the method comprising:

providing an input arranged to connect to the distributed computing system and access a distributed computing graph associated with an activity to be carried out using a plurality of the entities in the distributed computing system, the distributed computing graph representing currently executing processes in the distributed computing system;

arranging a processor to provide a dynamic mapping between the distributed computing graph and a data structure at the entity holding a local view of the distributed computing graph such that changes in the distributed computing graph are reflected in the local view;

providing a user interface enabling control of the activity by modifying the local view based on input received at the user interface which modifies the local view, the local view indicating which processes represented by the distributed computing graph are executed at the entity and which processes are executed remote of the entity; and

providing a peer-to-peer content distribution mechanism at the entity for accessing the distributed computing graph at the input from the distributed computing system,

the method further comprising;

receiving input at the user interface which modifies the local view and wherein the method comprises using the dynamic mapping between the local view and the distributed computing graph to update the distributed computing graph according to the modified local view in order to control the activity;

receiving user input specifying a second local view, attempting to connect to a distributed computing graph corresponding to the second local view and if successful updating the second local view according to the corresponding distributed computing graph;

or

receiving user input specifying the second local view, attempting to connect to the distributed computing graph corresponding to the second local view and if unsuccessful starting a new distributed computing graph corresponding to the second local view.

10. A method as claimed in claim 9 wherein the step of accessing the distributed computing graph comprises accessing a plurality of nodes connected by links where at least some of the nodes represent any of computational processes and data sources at one or more entities in the distributed computing system.

11. A method as claimed in claim 9 which comprises arranging the processor to provide the dynamic mapping such that changes in the local view are reflected in the distributed computing graph.

12. A method as claimed in claim 9 which further comprises receiving input at the user interface which modifies the local view and wherein the method comprises using the dynamic mapping between the local view and the distributed computing graph to update the distributed computing graph according to the modified local view in order to control the activity.

13. A method as claimed in claim 9 which further comprises receiving user input specifying the second local view;

attempting to connect to the distributed computing graph corresponding to the second local view and if successful updating the second local view according to the corresponding distributed computing graph.

14. A method as claimed in claim 9 which further comprises receiving user input specifying the second local view;

attempting to connect to the distributed computing graph corresponding to the second local view and if unsuccessful starting the new distributed computing graph corresponding to the second local view.

15. A distributed computing system comprising:

a plurality of entities in communication with one another, at least one of the Plurality of entities including a processor, using a communications network, each entity comprising at least one computing resource; and

a distributed computing graph, stored in the distributed computing system, the graph being associated with an activity to be carried out in the distributed computing system, the distributed computing graph representing processes associated with the activity at least some of which are currently executing in the distributed computing system, wherein;

each of the entities is arranged to provide a dynamic mapping between the distributed computing graph and a data structure at that entity holding a local view of the distributed computing graph such that changes in the distributed computing graph are reflected in the local view; and

at least some of the entities are arranged to provide a user interface which enables a user to control the activity by modifying the local view,

the at least some of the entities being further arranged to;

receive input at the user interface which modifies the local view and wherein the method comprises using the dynamic mapping between the local view and the distributed computing graph to update the distributed computing graph according to the modified local view in order to control the activity;

receive user input specifying a second local view, attempting to connect to a distributed computing graph corresponding to the second local view and if successful updating the second local view according to the corresponding distributed computing graph;

or

receive user input specifying the second local view, attempting to connect to the distributed computing graph corresponding to the second local view and if unsuccessful starting a new distributed computing graph corresponding to the second local view.

16. A distributed computing system as claimed in claim 15 wherein at least some of the entities are arranged to provide a user interface based on the local view at that entity;

the local view indicating which processes represented by the distributed computing graph are executed at the entity and which processes are executed remote of the entity.

17. A distributed computing system as claimed in claim 15 wherein each entity is arranged to provide the dynamic mapping such that changes in the local view are implemented in the distributed computing graph.

18. A distributed computing system as claimed in claim 15 wherein each entity comprises a peer-to-peer content distribution mechanism for distribution of the distributed computing graph within the distributed computing system.

19. A distributed computing system as claimed in claim 15 wherein each entity comprises a replication mechanism for replicating the distributed computing graph at the entities in the distributed computing system.