Distributed persistent memory using asynchronous streaming of log records

Abstract

Technologies for distributed durable data replication include a computing device having persistent memory that stores a memory state and an update log. The computing device isolates a host partition from a closure partition. The computing device may sequester one or more processor cores for use by the closure partition. The host partition writes transaction records to the update log prior to writing state changes to persistent memory. A replication service asynchronously transmits log records to a remote computing device, which establishes a replica update log in persistent memory. If the host partition fails, the closure partition transmits remaining log records from the update log to the remote computing device. The update log may be quickly replayed when recovering the computing device from failure. The remote computing device may also replay the replica update log to update a remote copy of the state data. Other embodiments are described and claimed.

1 . A computing device for distributed data durability, the computing device comprising:
 a persistent memory to store a replica update log; and   a replication receipt module to:
 receive a transaction record corresponding to a persistent memory state change from another computing device; 
 store the transaction record in the replica update log; 
 replay the transaction record to apply the persistent memory state change to a remote state data copy of the computing device in response to storing of the transaction record; and 
 remove the transaction record from the replica update log in response to replaying of the transaction record. 
   
     
     
         2 . The computing device of  claim 1 , wherein to replay the transaction record comprises to replay the transaction record to apply the persistent memory state change to the remote state data copy stored in the persistent memory of the computing device. 
     
     
         3 . The computing device of  claim 1 , wherein to replay the transaction record comprises to replay the transaction record to apply the persistent memory state change to the remote state data copy stored in a data storage device of the computing device. 
     
     
         4 . The computing device of  claim 1 , wherein to receive the transaction record comprises to receive the transaction record from a host partition of the other computing device. 
     
     
         5 . The computing device of  claim 1 , wherein to receive the transaction record comprises to receive the transaction record from a closure partition of the other computing device. 
     
     
         6 . The computing device of  claim 1 , wherein:
 to receive the transaction record comprises to receive the transaction record from a host partition of the other computing device; and   the replication receipt module is further to:
 receive a second transaction record corresponding to a second persistent memory state change from a closure partition of the other computing device; 
 store the second transaction record in the replica update log; 
 replay the second transaction record to apply the second persistent memory state change to the remote state data copy in response to storing of the second transaction record; and 
 remove the second transaction record from the replica update log in response to replaying of the second transaction record. 
   
     
     
         7 . The computing device of  claim 1 , wherein the replication receipt module comprises a polling-based packet transport service to receive the transaction record. 
     
     
         8 . A method for distributed data durability, the method comprising:
 receiving, by a computing device, a transaction record corresponding to a persistent memory state change from another computing device;   storing, by the computing device, the transaction record in a replica update log stored in a persistent memory of the computing device;   replaying, by the computing device, the transaction record to apply the persistent memory state change to a remote state data copy of the computing device in response to storing the transaction record; and   removing, by the computing device, the transaction record from the replica update log in response to replaying the transaction record.   
     
     
         9 . The method of  claim 8 , wherein replaying the transaction record comprises replaying the transaction record to apply the persistent memory state change to the remote state data copy stored in the persistent memory of the computing device. 
     
     
         10 . The method of  claim 8 , wherein replaying the transaction record comprises replaying the transaction record to apply the persistent memory state change to the remote state data copy stored in a data storage device of the computing device. 
     
     
         11 . The method of  claim 8 , wherein receiving the transaction record comprises receiving the transaction record from a closure partition of the other computing device. 
     
     
         12 . The method of  claim 8 , wherein receiving the transaction record comprises receiving the transaction record from a host partition of the other computing device, the method further comprising:
 receiving, by the computing device, a second transaction record corresponding to a second persistent memory state change from a closure partition of the other computing device;   storing, by the computing device, the second transaction record in the replica update log;   replaying, by the computing device, the second transaction record to apply the second persistent memory state change to the remote state data copy in response to storing the second transaction record; and   removing, by the computing device, the second transaction record from the replica update log in response to replaying the second transaction record.   
     
     
         13 . The method of  claim 8 , wherein receiving the transaction record comprises receiving the transaction record using a polling-based packet transport service of the computing device. 
     
     
         14 . One or more computer-readable storage media comprising a plurality of instructions that in response to being executed cause a computing device to:
 receive a transaction record corresponding to a persistent memory state change from another computing device;   store the transaction record in a replica update log stored in a persistent memory of the computing device;   replay the transaction record to apply the persistent memory state change to a remote state data copy of the computing device in response to storing the transaction record; and   remove the transaction record from the replica update log in response to replaying the transaction record.   
     
     
         15 . The one or more computer-readable storage media of  claim 14 , wherein to replay the transaction record comprises to replay the transaction record to apply the persistent memory state change to the remote state data copy stored in the persistent memory of the computing device. 
     
     
         16 . The one or more computer-readable storage media of  claim 14 , wherein to replay the transaction record comprises to replay the transaction record to apply the persistent memory state change to the remote state data copy stored in a data storage device of the computing device. 
     
     
         17 . The one or more computer-readable storage media of  claim 14 , wherein to receive the transaction record comprises to receive the transaction record from a host partition of the other computing device. 
     
     
         18 . The one or more computer-readable storage media of  claim 14 , wherein to receive the transaction record comprises to receive the transaction record from a closure partition of the other computing device. 
     
     
         19 . The one or more computer-readable storage media of  claim 14 , wherein to receive the transaction record comprises to receive the transaction record from a host partition of the other computing device, and wherein the one or more computer-readable storage media further comprises a plurality of instructions that in response to being executed cause the computing device to:
 receive a second transaction record corresponding to a second persistent memory state change from a closure partition of the other computing device;   store the second transaction record in the replica update log;   replay the second transaction record to apply the second persistent memory state change to the remote state data copy in response to storing the second transaction record; and   remove the second transaction record from the replica update log in response to replaying the second transaction record.   
     
     
         20 . The one or more computer-readable storage media of  claim 14 , wherein to receive the transaction record comprises to receive the transaction record using a polling-based packet transport service of the computing device.