US 12,111,739 B1
Data caching in reserved storage capacity of cloud provider network extensions
Ravi Mantrala, Bothell, WA (US); Peng Wu, Westborough, MA (US); Danny Wei, Seattle, WA (US); Shailesh Bhagwat, Southborough, MA (US); and Sandeep Kumar, Sammamish, WA (US)
Assigned to Amazon Technologies, Inc., Seattle, WA (US)
Filed by Amazon Technologies, Inc., Seattle, WA (US)
Filed on Sep. 30, 2022, as Appl. No. 17/937,361.
Int. Cl. G06F 11/20 (2006.01); G06F 9/50 (2006.01); G06F 11/07 (2006.01)
CPC G06F 11/2025 (2013.01) [G06F 9/5044 (2013.01); G06F 9/5077 (2013.01); G06F 11/0772 (2013.01)] 19 Claims
OG exemplary drawing
 
1. A cloud provider network extension system that is remote from a cloud provider network and facilitates interaction with the cloud provider network, the cloud provider network extension system comprising:
one or more data stores comprising:
an in-use portion storing a data volume, the data volume corresponding to persistent storage of a virtual machine instance that implements functionality associated with the cloud provider network; and
a reserved portion reserved for failover operation related to the data volume during failure of some or all of the in-use portion; and
one or more processors configured to:
retrieve at least one portion of a machine image from the cloud provider network, the machine image representing a point-in-time representation of the data volume;
cache the at least one portion of the machine image in the reserved portion of the one or more data stores;
obtain a request to read from the machine image representing the point-in-time representation of the data volume;
respond to the request by providing the at least one portion of the machine image from the reserved portion of the one or more data stores;
detect at least partial failure of the in-use portion of the one or more data stores, the at least partial failure corresponding to a failure to store at least one portion of the data volume; and
responsive to the at least partial failure of the in-use portion of the one or more data stores:
delete the at least one portion of the machine image in the reserved portion; and
replicate the at least one portion of the data volume to the reserved portion as a failover operation.