Abstract: Multi-model fine turning and inference includes batching, by a base executor, a plurality of requests received from a plurality of client executors into a request batch. Each request specifies input data and requests offload processing of the input data by a selected layer of a plurality of layers of a base model of the base executor. The base executor processes the requests of the request batch through the selected layer to generate, for each request of the request batch, an output corresponding to the request. Each output is transmitted from the base executor to the client executor that submitted the request corresponding to the output. The outputs generated in response to the batched requests enable the one or more client executors to perform specific tasks.

Abstract: A method, computer program product, and computer system for using a shared mixture of experts (MoE) architecture to implement inference with respect to a machine learning model. A gate mechanism receives, from N clients, N requests and an identification of N MoE models. N is at least 2. The gate mechanism selects, from N sets of experts, E experts to process the N requests. The N sets of experts collectively include at least one duplicative expert that is common to more than one set of the N sets of experts. Each duplicative expert has been deduplicated by being stored only once in the one or more graphic processing units (GPUs). The N requests are routed to the E experts. The E experts are executed to generate N respective responses to the N requests. Each response of the N responses is transmitted to the respective client of the N clients.

Abstract: A selected recipe operating successfully on the second orchestration container for a transferred target application is confirmed. The recipes for transferring the target application to the second orchestration container are formed. The recipe for transferring the target application to the second orchestration container is executed. The recipe for transferring the target application to the second orchestration container is re-executed. A selected recipe based on a determination that the selected recipe meets or exceeds a threshold for operation on the second orchestration container.

Abstract: A selected recipe operating successfully on the second orchestration container for a transferred target application is confirmed. The recipes for transferring the target application to the second orchestration container are formed. The recipe for transferring the target application to the second orchestration container is executed. The recipe for transferring the target application to the second orchestration container is re-executed. A selected recipe based on a determination that the selected recipe meets or exceeds a threshold for operation on the second orchestration container.

Abstract: An embodiment includes detecting, by a Recovery System, a cache metric where the cache metric comprises a probability of access to a data object for each of a plurality of time segments of a failover log, where the probability of access to the data object for each of the plurality of time segments is based on an access time of the data object contained in the failover log. The embodiment also includes responsive to receiving the cache metric, caching by the Recovery System of the data object based on the cache metric where the data object is cached in a standby cache of the Recovery System.

Abstract: An embodiment enables a first machine to receive an asynchronous data update from a second machine, the asynchronous data update comprising a first stream of identifiers and a second stream of data, wherein the first stream of identifiers comprises a first identifier corresponding to a first data segment in the second stream of data. An embodiment analyzes, as a part of the asynchronous data update, the first stream of identifiers. An embodiment determines at a first time, responsive to the analyzing, that the first data segment corresponding to the first identifier has not arrived at the first machine by the first time. An embodiment updates, responsive to the determining at the first time, a record at the first machine with the first identifier. An embodiment updates the record at a second time, responsive to receiving the first data segment at the first machine at the second time.

Abstract: User-defined information relating to recovery of an application is obtained, based on a selected event. Data relating to a set of candidate systems available to deploy the application is obtained. The data includes one or more system statistics for the set of candidate systems, and the one or more system statistics include, at least, one or more recovery point actuals and historical data for the set of candidate systems. A scoring of the set of candidate systems is performed based on the user-defined information and the data that is obtained. A candidate system is selected from the set of candidate systems based on the scoring. Deployment of the application on the candidate system that is selected is initiated.

Abstract: An embodiment enables a first machine to receive an asynchronous data update from a second machine, the asynchronous data update comprising a first stream of identifiers and a second stream of data, wherein the first stream of identifiers comprises a first identifier corresponding to a first data segment in the second stream of data. An embodiment analyzes, as a part of the asynchronous data update, the first stream of identifiers. An embodiment determines at a first time, responsive to the analyzing, that the first data segment corresponding to the first identifier has not arrived at the first machine by the first time. An embodiment updates, responsive to the determining at the first time, a record at the first machine with the first identifier. An embodiment updates the record at a second time, responsive to receiving the first data segment at the first machine at the second time.

Abstract: An embodiment includes detecting, by a Recovery System, a cache metric where the cache metric comprises a probability of access to a data object for each of a plurality of time segments of a failover log, where the probability of access to the data object for each of the plurality of time segments is based on an access time of the data object contained in the failover log. The embodiment also includes responsive to receiving the cache metric, caching by the Recovery System of the data object based on the cache metric where the data object is cached in a standby cache of the Recovery System.

Abstract: A computer-implemented method according to one embodiment includes creating a log of one or more events occurring during a performance of continuous data protection (CDP) for an instance of data; in response to identifying a conclusion of the performance of CDP for the instance of data, determining a value for the one or more events within the log; and conditionally creating and storing a snapshot of the instance of data, based on a comparison of the value for the one or more events to a predetermined threshold.

Abstract: A computer-implemented method according to one embodiment includes creating a log of one or more events occurring during a performance of continuous data protection (CDP) for an instance of data; in response to identifying a conclusion of the performance of CDP for the instance of data, determining a value for the one or more events within the log; and conditionally creating and storing a snapshot of the instance of data, based on a comparison of the value for the one or more events to a predetermined threshold.