Fault recovery management in a cloud computing environment

Patent metadata

Field	Value
Publication number	US-10061652-B2
Application number	US-201615220320-A
Country	US
Kind code	B2
Filing date	Jul 26, 2016
Priority date	Jul 26, 2016
Publication date	Aug 28, 2018
Grant date	Aug 28, 2018

Abstract

Technologies for managing fault recovery in a cloud computing environment may be used after faults of various sizes, including faults which put total functioning capacity below subscribed capacity. Computing services have repair priorities. A fault recovery manager selects a higher priority service whose capacity is below a minimum availability, and chooses a lower priority service still above its minimal availability, and reassigns capacity from the lower priority service to the higher priority service without depriving the lower priority service of operability. Capacity reassignment continues at least until the higher priority service is at or above minimal availability, or the lower priority service is at minimal availability. Lower priority services may also be terminated entirely to free up resources for higher priority services. New deployments may be prevented until all services are at or above minimal availability. Spare capacity may be reserved against demand fluctuations or further faults.

First claim

What is claimed is:

1. A computing technology method for fault recovery management by a tenant or an infrastructure provider in a cloud computing environment having computational resources which support instances of computing services, the method comprising:
selecting a higher priority computing service which has a current number of working instances that is below a minimal availability threshold of the higher priority computing service, the higher priority computing service also having a repair priority;
choosing a lower priority computing service which has a repair priority that is lower than the higher priority computing service's repair priority, the lower priority computing service also having a current number of working instances that is above a minimal availability threshold of the lower priority computing service;
reducing the current number of working instances of the chosen lower priority computing service, thereby freeing for reassignment at least a portion of at least one computational resource of the cloud computing environment;
increasing the number of working instances of the higher priority computing service, thereby reassigning at least part of the freed computational resource; and
repeating at least said reducing and increasing at least until one of the following conditions occurs: (i) the number of working instances of the higher priority computing service is no longer below the minimal availability threshold of the higher priority computing service, or (ii) further reducing would put the number of working instances of the lower priority computing service below the minimal availability threshold of the lower priority computing service.

2. The method of claim 1 , wherein the cloud computing environment comprises a data center, wherein each computing service either belongs to an infrastructure provider or else belongs to at least one tenant, and wherein the method further comprises preventing deployment to the data center of any additional computing service of any one or more tenants until each computing service deployed in the data center is either terminated or suspended or has a number of working instances which is not below a minimal availability threshold of the computing service.

3. The method of claim 1 , wherein the method further comprises calculating availability gaps for at least two computing services in the cloud computing environment, wherein an availability gap of a computing service is calculated as the minimal availability threshold of the computing service minus the current number of working instances of the computing service, and wherein selecting a higher priority computing service includes favoring selection of computing services with larger availability gaps over selection of computing services with smaller availability gaps, at least for a given repair priority.

4. The method of claim 1 , wherein the method further comprises:
calculating availability gaps for at least two computing services in the cloud computing environment, wherein an availability gap of a computing service is calculated as the minimal availability threshold of the computing service minus the current number of working instances of the computing service;
sorting the at least two computing services according to their availability gaps and their repair priorities; and
wherein selecting a higher priority computing service includes selecting from among the sorted computing services.

5. The method of claim 1 , wherein said selecting, choosing, reducing and increasing are repeated until each computing service has at least a minimal availability number of working instances, in that for each computing service in the cloud computing environment the number of working instances of the computing service is not below the minimal availability threshold of the computing service.

6. The method of claim 5 , wherein the cloud computing environment has unassigned computational resources remaining after the number of working instances of each non-terminated and non-suspended computing service is not below the minimal availability threshold of the computing service, and the method further comprises:
selecting a highest priority computing service whose repair priority is at least at high as the repair priority of any other computing service; and
increasing the number of working instances of the highest priority computing service, thereby reassigning at least part of the unassigned computational resources.

7. The method of claim 5 , wherein the cloud computing environment has unassigned computational resources remaining after the number of working instances of each non-terminated and non-suspended computing service is not below the minimal availability threshold of the computing service, and the method further comprises at least one of the following:
dedicating at least a portion of the unassigned computational resources to a reserve, thereby providing a buffer against demand fluctuations or faults; or
allowing a previously prevented deployment of at least one additional computing service in the cloud computing environment, thereby assigning at least part of the unassigned computational resources.

8. The method of claim 1 , wherein said selecting, choosing, reducing and increasing are repeated until no computing service has more than a minimal availability number of working instances, but at least one computing service still has less than a minimal availability number of working instances, and the method further comprises:
terminating all working instances of a computing service whose repair priority is at least at low as the repair priority of any other computing service; and
increasing the number of working instances of at least one computing service which had less than the minimal availability number of working instances before the increasing.

9. A computer-readable storage medium configured with executable instructions to perform a method for fault recovery management in a cloud computing environment including at least one availability zone having computational resources which support computing services, the method comprising:
determining that a current total functioning capacity of the availability zone is less than a total subscribed capacity of all the computing services in the availability zone;
selecting a higher priority computing service which has a current assigned capacity that is below a minimal availability threshold of the higher priority computing service, the higher priority computing service also having a repair priority;
choosing a lower priority computing service which has a repair priority that is lower than the higher priority computing service's repair priority, the lower priority computing service also having a current assigned capacity that is above a minimal availability threshold of the lower priority computing service;
reducing the current assigned capacity of the chosen lower priority computing service, thereby freeing for reassignment at least a portion of at least one computational resource of the cloud computing environment;
increasing the assigned capacity of the higher priority computing service, thereby reassigning at least part of the freed computational resource; and
repeating at least said reducing and increasing at least until one of the following conditions occurs: (i) the assigned capacity of the higher priority computing service is no longer below the minimal availability threshold of the higher priority computing service, (ii) the assigned capacity of the higher priority computing service is at least a subscribed capacity of the higher priority computing service, or (iii) further reducing would put the assigned capacity of the lower priority computing service below the minimal availability threshold

Assignees

• Microsoft Technology Licensing Llc

Inventors

• Johnson James E
• Russinovich Mark

Classifications

• G06F11/3442

  for planning or managing the needed capacity · CPC title

• G06F11/2033

  switching over of hardware resources · CPC title

• G06F11/0709

  in a distributed system consisting of a plurality of standalone computer nodes, e.g. clusters, client-server systems · CPC title

• G06F11/142 Primary

  Reconfiguring to eliminate the error (group management mechanisms in a peer-to-peer network H04L67/1044) · CPC title

• G06F2201/805

  Real-time · CPC title