Multiplexing users and enabling virtualization on a hybrid system

• US 8,914,528 B2
• Filed: 12/27/2011
• Issued: 12/16/2014
• Est. Priority Date: 06/24/2010
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Chan, Yuk Lung, Krishnamurthy, Rajaram B., Parris, Carl Joseph
• Primary Examiner(s): BILGRAMI, ASGHAR H
• Application Number:
  US13/337,746
• Publication Number:
  US 20120102138A1
• Time in Patent Office: 1,085 Days
• Field of Search: 709/229, 713/150
• US Class Current: 709/229
• CPC Class Codes: G06F 12/0862 with prefetch
• CPC Class Codes: G06F 12/0868 Data transfer between cache memory and other subsystems, e.g. storage devices or host systems
• CPC Class Codes: G06F 15/177 Initialisation or configuration control processor initialisation G06F9/4405
• CPC Class Codes: G06F 16/172 Caching, prefetching or hoarding of files
• CPC Class Codes: G06F 16/182 Distributed file systems
• CPC Class Codes: G06F 21/602 Providing cryptographic facilities or services
• CPC Class Codes: G06F 2212/264 Remote server
• CPC Class Codes: G06F 2212/602 Details relating to cache prefetching
• CPC Class Codes: G06F 3/0608 Saving storage space on storage systems
• CPC Class Codes: G06F 3/061 Improving I/O performance
• CPC Class Codes: G06F 3/0644 Management of space entities, e.g. partitions, extents, pools
• CPC Class Codes: G06F 3/065 Replication mechanisms
• CPC Class Codes: G06F 3/0653 Monitoring storage devices or systems
• CPC Class Codes: G06F 3/067 Distributed or networked storage systems, e.g. storage area networks [SAN], network attached storage [NAS]
• CPC Class Codes: G06F 3/0685 Hybrid storage combining heterogeneous device types, e.g. hierarchical storage, hybrid arrays
• CPC Class Codes: G06F 3/0688 Non-volatile semiconductor memory arrays
• CPC Class Codes: H04L 41/5054 Automatic deployment of services triggered by the service manager, e.g. service implementation by automatic configuration of network components
• CPC Class Codes: H04L 47/24 Traffic characterised by specific attributes, e.g. priority or QoS
• CPC Class Codes: H04L 67/568 Storing data temporarily at an intermediate stage, e.g. caching

Abstract

A method, hybrid server system, and computer program product, support multiple users in an out-of-core processing environment. At least one accelerator system in a plurality of accelerator systems is partitioned into a plurality of virtualized accelerator systems. A private client cache is configured on each virtualized accelerator system in the plurality of virtualized accelerator systems. The private client cache of each virtualized accelerator system stores data that is one of accessible by only the private client cache and accessible by other private client caches associated with a common data set. Each user in a plurality of users is assigned to a virtualized accelerator system from the plurality of virtualized accelerator systems.

Claims

1. A method, with a hybrid server system comprising a server system and at least one accelerator system, for supporting multiple users in an out-of-core processing environment, the method comprising:

partitioning, with at least one processor, at least one accelerator system in a plurality of accelerator systems into a plurality of virtualized accelerator systems;

configuring a private client cache on each virtualized accelerator system in the plurality of virtualized accelerator systems, wherein the private client cache of each virtualized accelerator system stores data that is one of accessible by only the private client cache and accessible by other private client caches associated with a common data set; and

assigning each user in a plurality of users to a virtualized accelerator system from the plurality of virtualized accelerator systems.

2. The method of claim 1, further comprising:

determining that a data set associated with a first virtualized accelerator system in the plurality of virtualized accelerator systems is also associated with a second virtualized accelerator system; and

transferring, based on determining that the data set is associated with both the first virtualized accelerator system and the second virtualized accelerator system, data between the first virtualized accelerator system and the second virtualized accelerator system.

3. The method of claim 2, further comprising:

determining that the data set is associated with a first user assigned to the first virtualized accelerator system and a second user assigned to the second virtualized accelerator system, wherein the first user and the second user are different from each other; and

monitoring, based on determining that the data set is associated with the first and second users, data traffic at the first virtualized accelerator system and the second virtualized accelerator system.

4. The method of claim 3, wherein the monitoring comprises:

monitoring, by the private client cache at the first virtualized accelerator system, data traffic on the private client cache of the second virtualized accelerator system; and

monitoring, by the private client cache at the second virtualized accelerator system, data traffic on the private client cache of the first virtualized accelerator system.

5. The method of claim 1, further comprising:

receiving a request for data from a first virtualized accelerator system in the plurality of virtualized accelerator systems;

broadcasting the request to a set of other virtualized accelerator systems in the plurality of virtualized accelerator systems; and

receiving, based on the broadcasting, at least a portion of the data requested by the first virtualized accelerator system from at least a second virtualized accelerator system in the set of other virtualized accelerator systems.

6. The method of claim 1, wherein the configuring comprises:

initially configuring each private client cache with a first cache size;

determining that cache space of least one private client cache is being utilized; and

increasing the at least one private client cache to a second cache size.

7. The method of claim 6, wherein increasing the at least one private client cache comprises:

determining that a first virtualized accelerator system comprising the at least one private client cache is requesting to increase a size of the at least one private client cache;

determining that a second virtualized accelerator system comprising an additional private client cache is requesting to increase a size of the additional private client cache;

identifying a priority associated with each of the first virtualized accelerator system and the second virtualized accelerator system; and

increasing the size of the at least one private client cache based on the first virtualized private client cache having a priority that is higher than the priority of the second virtualized accelerator system.

8. The method of claim 1, further comprising:

determining that a first user state associated with a first user in the plurality of users is substantially similar to a second user state associated with second user in the plurality of users, wherein the first user is assigned to a first virtualized accelerator system in the plurality of virtualized accelerator systems, and wherein the second user is assigned to a second virtualized accelerator system in the plurality of virtualized accelerator systems; and

based on the first user state and the second user state being substantially similar, joining the first virtualized accelerator system and the second virtualized accelerator system.

9. The method of claim 1, further comprising:

determining that a first user state associated with a first user in the plurality of users is substantially similar to a second user state associated with second user in the plurality of users, wherein the first user and the second user are assigned to a first virtualized accelerator system and a second virtualized accelerator system that have been logically joined; and

based on the first user state and the second user state being substantially similar, logically separating the first virtualized accelerator system and the second virtualized accelerator system.

10. A hybrid server system in an out-of-core processing environment, the hybrid server system comprising:

at least one accelerator system, comprising at least one processor, wherein the accelerator system is partitioned into a plurality of virtualized accelerator systems,

wherein each virtualized accelerator system in the plurality of accelerator systems comprises a private client cache, wherein the private client cache of each virtualized accelerator system stores data that is one of accessible by only the private client cache and accessible by other private client caches associated with a common data set, and wherein a different user is assigned to a virtualized accelerator system in the plurality of accelerator systems.

11. The accelerator system of claim 10, wherein a first virtualized accelerator system in the plurality of virtualized accelerator systems is configured to:

determine that a data set associated with the first virtualized accelerator system is also associated with a second virtualized accelerator system in the plurality of virtualized accelerator systems; and

transfer, based on determining that the data set is associated with both the first virtualized accelerator system and the second virtualized accelerator system, data between the first virtualized accelerator system and the second virtualized accelerator system.

12. The accelerator system of claim 11, wherein the first virtualized accelerator system is further configured to:

determine that the data set is associated with a first user assigned to the first virtualized accelerator system and a second user assigned to the second virtualized accelerator system, wherein the first user and the second user are different from each other; and

monitoring, based on determining that the data set is associated with the first and second users, data traffic at the second virtualized accelerator system.

13. The accelerator system of claim 10, further comprising:

a virtualized accelerator directory agent configured to

receive a request for data from a first virtualized accelerator system in the plurality of virtualized accelerator systems; and

broadcast the request to a set of other virtualized accelerator systems in the plurality of virtualized accelerator systems,

wherein at least a portion of the data requested is sent from at least a second virtualized accelerator system in the set of other virtualized accelerator systems to the first virtualized accelerator system based on the broadcasting.

14. The hybrid server system of claim 10, further comprising:

a server system communicatively coupled to the accelerator system, wherein the server system is configured to;

initially configure each private client cache with a first cache size;

determine that cache space of least one private client cache is being utilized; and

increase the at least one private client cache to a second cache size,

wherein server system is configured to increase the at least one private client cache by;

determining that a first virtualized accelerator system comprising the at least one private client cache is requesting to increase a size of the at least one private client cache;

determining that a second virtualized accelerator system comprising an additional private client cache is requesting to increase a size of the additional private client cache;

identifying a priority associated with each of the first virtualized accelerator system and the second virtualized accelerator system; and

increasing the size of the at least one private client cache based on to the first virtualized private client cache having a priority that is higher than the priority of the second virtualized accelerator system.

15. A computer program product for managing multiple users in an out-of-core processing environment, the computer program product comprising:

a non-transitory storage medium readable by a processing circuit and storing instructions for execution by the processing circuit for performing a method comprising;

partitioning, with at least one processor, at least one accelerator system in a plurality of accelerator systems into a plurality of virtualized accelerator systems;

configuring a private client cache on each virtualized accelerator system in the plurality of virtualized accelerator systems, wherein the private client cache of each virtualized accelerator system stores data that is one of accessible by only the private client cache and accessible by other private client caches associated with a common data set; and

assigning each user in a plurality of users to a virtualized accelerator system from the plurality of virtualized accelerator systems.

16. The computer program product of claim 15, wherein the method further comprises:

determining that a data set associated with a first virtualized accelerator system in the plurality of virtualized accelerator systems is also associated with a second virtualized accelerator system; and

transferring, based on determining that the data set is associated with both the first virtualized accelerator system and the second virtualized accelerator system, data between the first virtualized accelerator system and the second virtualized accelerator system.

17. The computer program product of claim 16, wherein the method further comprises:

determining that the data set is associated with a first user assigned to the first virtualized accelerator system and a second user assigned to the second virtualized accelerator system, wherein the first user and the second user are different from each other; and

monitoring, based on determining that the data set is associated with the first and second users, data traffic at the first virtualized accelerator system and the second virtualized accelerator system.

18. The computer program product of claim 15, wherein the method further comprises:

receiving a request for data from a first virtualized accelerator system in the plurality of virtualized accelerator systems;

broadcasting the request to a set of other virtualized accelerator systems in the plurality of virtualized accelerator systems; and

receiving, based on the broadcasting, at least a portion of the data requested by the first virtualized accelerator system from at least a second virtualized accelerator system in the set of other virtualized accelerator systems.

19. The computer program product of claim 15, wherein the configuring comprises:

initially configuring each private client cache with a first cache size;

determining that cache space of least one private client cache is being utilized; and

increasing the at least one private client cache to a second cache size.

20. The computer program product of claim 19, wherein increasing the at least one private client cache comprises:

determining that a first virtualized accelerator system comprising the at least one private client cache is requesting to increase a size of the at least one private client cache;

determining that a second virtualized accelerator system comprising an additional private client cache is requesting to increase a size of the additional private client cache;

identifying a priority associated with each of the first virtualized accelerator system and the second virtualized accelerator system; and

increasing the size of the at least one private client cache based on the first virtualized private client cache having a priority that is higher than the priority of the second virtualized accelerator system.