Geographic co-location service for cloud computing
First Claim
1. A method utilizing a computing processor and memory for geographically organizing the storage of data and hosted services in a distributed computing environment, the method comprising:
- receiving one or more allocation constraints for an affinity group, the affinity group being a grouping of storage accounts and computation accounts with an assigned affinity;
identifying, based in part on the one or more allocation constraints and a weighted coin-toss accounting for a worst-case capacity of resources at different geo-domains, a geo-domain for access to an account from the geo-domain and for physical storage of the account, the account being associated with the affinity group, the account comprising a storage account for storing data used in association with the hosted service and a computation account for storing code of the hosted service;
identifying a stamp of the geo-domain; and
storing the account in the stamp,wherein the worst-case capacity of resources is defined as a maximum resource load to which a given geo-domain would experience when all currently stored accounts at the geo-domain maximize a measured resource.
3 Assignments
0 Petitions
Accused Products
Abstract
Embodiments of the present invention relate to systems, methods, and computer storage media for geographically organizing the storage of data and hosted services in a distributed computing environment. An allocation constraint is utilized to identify and select a geo-region in which an affinity group that is comprised of a hosted service and/or data is to be physically stored. Additionally, a geo-domain within the geo-region is identified for hosting the hosted service and/or data of the affinity group. A stamp, such as a storage stamp or a computation stamp, is identified on the geo-domain for storing the hosted service and/or the data. The hosted service is stored in an identified computation stamp and the data is stored in an identified storage stamp. In an additional exemplary embodiment, the identification of a geo-domain involves analyzing resource information related to potential geo-domains within a geo-region.
77 Citations
12 Claims
-
1. A method utilizing a computing processor and memory for geographically organizing the storage of data and hosted services in a distributed computing environment, the method comprising:
-
receiving one or more allocation constraints for an affinity group, the affinity group being a grouping of storage accounts and computation accounts with an assigned affinity; identifying, based in part on the one or more allocation constraints and a weighted coin-toss accounting for a worst-case capacity of resources at different geo-domains, a geo-domain for access to an account from the geo-domain and for physical storage of the account, the account being associated with the affinity group, the account comprising a storage account for storing data used in association with the hosted service and a computation account for storing code of the hosted service; identifying a stamp of the geo-domain; and storing the account in the stamp, wherein the worst-case capacity of resources is defined as a maximum resource load to which a given geo-domain would experience when all currently stored accounts at the geo-domain maximize a measured resource. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12)
-
Specification