STORAGE MANAGEMENT FOR A CLUSTER OF INTEGRATED COMPUTING SYSTEMS

• US 20140380303A1
• Filed: 06/21/2013
• Published: 12/25/2014
• Est. Priority Date: 06/21/2013
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Bello, Adekunle, Yedavilli, Aruna
• Granted Patent: US 9,417,903 B2
• Time in Patent Office: Days
• US Class Current: 718/1
• CPC Class Codes: G06F 15/161 Computing infrastructure, e.g. computer clusters, blade chassis or hardware partitioning casings, cabinets, racks or drawers for data centers H05K5/00
• CPC Class Codes: G06F 15/17343 wherein the interconnection is dynamically configurable, e.g. having loosely coupled nearest neighbor architecture reconfigurable processors arrays G06F15/7867
• CPC Class Codes: G06F 16/1827 Management specifically adapted to NAS management of storage area networks [SAN] G06F3/067
• CPC Class Codes: G06F 2009/4557 Distribution of virtual machine instances; Migration and load balancing
• CPC Class Codes: G06F 2009/45579 I/O management, e.g. providing access to device drivers or storage
• CPC Class Codes: G06F 3/0604 Improving or facilitating administration, e.g. storage management
• 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 9/45558 Hypervisor-specific management and integration aspects
• CPC Class Codes: H04L 67/1097 for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS]

Abstract

Integrated computing systems with independently managed infrastructures including compute nodes and storage nodes form a cluster. Storage resource agents manage storage resources in the cluster. The resource agents identify storage requirements associated with allocation sets for resource consumers dispatched in the cluster, communicate with each other to locate inter-system storage resources that primarily satisfy locality criteria associated with resource consumer workloads, secondarily satisfy allocation set activity criteria associated with the allocation sets, and allocate the storage resources to the resource consumers to satisfy the storage requirements. The storage resource agents may base storage assignments on data placement information from a priority map. Data may be later relocated to alternate storage resources in satisfaction of cluster-wide storage policies, priority determinations, and data access rate determinations.

Claims

1. A method for managing storage resources in a cluster of integrated systems, the method comprising:

associating a plurality of communicatively interconnected storage resource agents with the cluster of integrated systems, each integrated system having an independently managed infrastructure including a plurality of compute nodes and a plurality of storage nodes;

identifying, by a first storage resource agent associated with a first integrated system in the cluster, a first storage requirement for a first resource consumer, the first storage requirement associated with an allocation set, the first resource consumer associated with a workload and dispatched on a first virtual machine hosted on a first compute node in the first integrated system;

locating, by the plurality of storage resource agents, at least one candidate storage resource in the cluster of integrated systems, each candidate storage resource associated with a locality measurement for the workload, the locality measurement satisfying locality criteria;

selecting a storage resource from the at least one candidate storage resource, the selected storage resource satisfying allocation set activity criteria, the selected storage resource on a storage node in a second integrated system in the cluster; and

allocating the selected storage resource to the first resource consumer to satisfy the first storage requirement.

2. The method of claim 1, wherein the locality criteria is selected from the group consisting of:

a requirement that the locality measurement be higher than a predetermined threshold;

a requirement that the locality measurement be among a top number of possible locality measurements, each available storage resource in the cluster associated with a possible locality measurement; and

a requirement that the locality measurement be among a top percentage of the possible locality measurements.

3. The method of claim 1, wherein the allocation set activity criteria is selected from the group consisting of:

a requirement that the allocation set be previously associated with the selected storage resource;

a requirement that the workload be previously associated with the selected storage resource; and

a requirement that the resource consumer be previously associated with the selected storage resource.

4. The method of claim 1, wherein the first resource consumer has a first cluster-wide priority, wherein a second resource consumer dispatched on a second virtual machine hosted on a second compute node in a third integrated system has a second cluster-wide priority, and wherein the selected storage resource is selected for a second storage requirement for the second resource consumer, the method further comprising:

determining, by the plurality of storage resource agents, that the first cluster-wide priority is higher than the second cluster-wide priority.

5. The method of claim 4, wherein a plurality of resource consumers is dispatched on a plurality of virtual machines hosted on the plurality of compute nodes in the cluster of integrated systems, wherein each resource consumer has a consumer priority, wherein each virtual machine has a virtual machine priority, and wherein each integrated system has a system priority, the method further comprising:

identifying a first consumer priority for the first resource consumer;

identifying a first virtual machine priority for the first virtual machine;

identifying a first system priority for the first integrated system;

identifying a second consumer priority for the second resource consumer;

identifying a second virtual machine priority for the second virtual machine;

identifying a second system priority for the second integrated system;

calculating the first cluster-wide priority based on at least one of the first consumer priority, the first virtual machine priority, and the first system priority; and

calculating the second cluster-wide priority based on at least one of the second consumer priority, the second virtual machine priority, and the second system priority.

6. The method of claim 4, wherein a cluster-wide priority map associates the first resource consumer with the first cluster-wide priority and further associates the second resource consumer with the second cluster-wide priority, the method further comprising:

accessing the first cluster-wide priority in the cluster-wide priority map; and

accessing the second cluster-wide priority in the cluster-wide priority map.

7. The method of claim 1, wherein the plurality of storage nodes in the first integrated system is inadequate for the first storage requirement.

8. The method of claim 1, wherein the allocating the selected storage resource to the first resource consumer satisfies inter-system storage criteria.

9. The method of claim 8, wherein the inter-system storage criteria is selected from the group consisting of:

a directive indicating that the first resource consumer is eligible for inter-system data storage, the directive associated with the first resource consumer;

a directive indicating that data associated with the first storage requirement is eligible for inter-system data storage, the directive associated with the first resource consumer;

a determination that the allocation set has an access rate below an access rate threshold;

a determination that the first resource consumer has a priority below a priority threshold; and

a determination that the first resource consumer has a first cluster-wide priority higher than a second cluster-wide priority, a second resource consumer having the second cluster-wide priority, the second resource consumer further having a second storage requirement, the selected storage resource selected for the second storage requirement.

10. The method of claim 1, wherein the first resource consumer has a first cluster-wide priority, and wherein the selecting the selected storage resource comprises:

determining that the selected storage resource is allocated to a second resource consumer, the second resource consumer having a second cluster-wide priority lower than the first cluster-wide priority; and

deallocating the selected storage resource from the second resource consumer.

11. The method of claim 1, further comprising:

locating, after allocating the selected storage resource to the first resource consumer, an alternate storage resource on an alternate storage node in an alternate integrated system; and

relocating the allocation set from the storage resource to the alternate storage resource, the relocating satisfying relocation criteria.

12. The method of claim 11, wherein the relocation criteria is selected from the group consisting of:

a determination that the allocation set has an access rate above an access rate threshold;

a determination that the first resource consumer has a priority above a priority threshold;

a cluster-wide storage policy of minimizing inter-system data storage; and

a cluster-wide storage policy of balancing distribution of available storage resources in the cluster.

13. A system for managing storage resources in a cluster of integrated systems, the system comprising:

a priority map identifying a plurality of resource consumers dispatched in the cluster of integrated systems, each integrated system having an independently managed infrastructure including a plurality of compute nodes and a plurality of storage nodes, the priority map associating each resource consumer with data placement information; and

a plurality of communicatively interconnected storage resource agents associated with the cluster of integrated systems, the plurality of storage resource agents adapted to;

identify storage requirements for the plurality of resource consumers, the storage requirements associated with allocation sets, the resource consumers associated with workloads;

locate inter-system storage resources in the cluster, the storage resources primarily satisfying locality criteria associated with the workloads and secondarily satisfying allocation set activity criteria;

access the priority map; and

assign the inter-system storage resources to the resource consumers to satisfy the storage requirements, the assignment based at least in part on the data placement information in the priority map.

14. The system of claim 13, wherein the allocation set activity criteria is selected from the group consisting of:

a requirement that the allocation sets be previously associated with the storage resources;

a requirement that the workloads be previously associated with the storage resources; and

a requirement that the resource consumers be previously associated with the storage resources.

15. The system of claim 13, wherein the data placement information for each resource consumer includes a directive and a cluster-wide priority, and wherein the inter-system storage resource assignments comply with the directive and the cluster-wide priority of each resource consumer in the priority map.

16. A computer program product for managing storage resources in a cluster of integrated systems, the computer program product comprising a computer-readable storage medium having program code embodied therewith, the program code executable by a computer system to perform a method comprising:

associating a plurality of communicatively interconnected storage resource agents with the cluster of integrated systems, each integrated system having an independently managed infrastructure including a plurality of compute nodes and a plurality of storage nodes;

identifying, by a first storage resource agent associated with a first integrated system in the cluster, a first storage requirement for a first resource consumer, the first storage requirement associated with an allocation set, the first resource consumer associated with a workload and dispatched on a first virtual machine hosted on a first compute node in the first integrated system;

locating, by the plurality of storage resource agents, at least one candidate storage resource in the cluster of integrated systems, each candidate storage resource associated with a locality measurement for the workload, the locality measurement satisfying locality criteria;

selecting a storage resource from the at least one candidate storage resource, the selected storage resource satisfying allocation set activity criteria, the selected storage resource on a storage node in a second integrated system in the cluster; and

allocating the selected storage resource to the first resource consumer to satisfy the first storage requirement, the allocating further satisfying inter-system storage criteria.

17. The computer program product of claim 16, wherein the first resource consumer has a first cluster-wide priority, wherein a second resource consumer dispatched on a second virtual machine hosted on a second compute node in a third integrated system has a second cluster-wide priority, and wherein the selected storage resource is selected for a second storage requirement for the second resource consumer, the method further comprising:

determining, by the plurality of storage resource agents, that the first cluster-wide priority is higher than the second cluster-wide priority.

18. The computer program product of claim 16, wherein the inter-system storage criteria is selected from the group consisting of:

a directive indicating that the first resource consumer is eligible for inter-system data storage, the directive associated with the first resource consumer;

a directive indicating that data associated with the first storage requirement is eligible for inter-system data storage, the directive associated with the first resource consumer;

a determination that the allocation set has an access rate below an access rate threshold;

a determination that the first resource consumer has a priority below a priority threshold; and

a determination that the first resource consumer has a first cluster-wide priority higher than a second cluster-wide priority, a second resource consumer having the second cluster-wide priority, the second resource consumer further having a second storage requirement, the storage resource selected for the second storage requirement.

19. The computer program product of claim 16, the method further comprising:

locating, after allocating the selected storage resource to the first resource consumer, an alternate storage resource on an alternate storage node in an alternate integrated system; and

relocating the allocation set from the storage resource to the alternate storage resource, the relocating satisfying relocation criteria.

20. The computer program product of claim 19, wherein the relocation criteria is selected from the group consisting of:

a determination that the allocation set has an access rate above an access rate threshold;

a determination that the first resource consumer has a priority above a priority threshold;

a cluster-wide storage policy of minimizing inter-system data storage; and

a cluster-wide storage policy of balancing distribution of available storage resources in the cluster.