Abstract: Disclosed are methods for estimating a time associated with shifting a first workload from a first compute environment to a second compute environment, separate from the first compute environment, estimating a likelihood of success associated with a likelihood that the first workload could successfully be shifted to the second compute environment, dividing or using the likelihood of success by the time to yield or produce a risk-adjusted shift time and, when a comparison of the shift time is longer than a maximum acceptable wait time, proceeding with a first operation associated with how to preempt the first workload by the second workload.

Abstract: The estimation of a risk factor applied to a negative potential of a focus computing resource based on proximity of the focus computing resource from other computing resources, and not just based on neighboring distance-1 computing resources. Upon determining that the risk factor applied to a negative potential of a focus computing resource is to be estimated, the negative potentials for one or more distance-1 computing resources are accessed. These negative potentials may have, in turn, been estimated based on distance and risk factor propagation from yet other higher distance computing resources. The negative potentials for the distance-1 computing resources are then used to estimate the risk factor applied to the negative potential of the focus computing resource.

Abstract: Automated estimation that a hardware or software change is correlated with one or more problem events. In addition to identifying that a change has occurred, one or more problem events are identified that have occurred after the change. Problem events are then correlated with the change using a distance measure between the change and the at least one correlated problem event. For most, if not all, of the distance measure, the greater the distance measure, the less the change is correlated with each correlated problem event. The distance measure may be a time interval, an activity-weighted interval, or any other type of distance measure.

Abstract: Restoring data, without suspending access to the data for the entire time that the data is being restored. Access is suspended only while a portion of the data is restored, before access to all of the data is permitted. A driver virtualizes any remaining unrestored data. To accomplish this, the driver maintains a list of data yet to be restored. If a request to read a portion of the data is then received, and the requested portion is on the list, the driver restores the requested portion before permitting the read request to be fulfilled. The restored portion is then removed from the list of unrestored data. If a request to write a portion of the data is received, the write is permitted, and the location just written to is removed from the list without restoring.

Abstract: Mechanisms for data source computing system(s) to provide data to data targets. The data source(s) access an identification of common groups of blocks, each group being common amongst all or a subset of the data target(s). The common groups of blocks may have different scope such that one common group of blocks is to be provided to a different set of the data targets than another common group of blocks. Optionally, a selection mechanism may be used to assign priority in initiating multicasting of the identified common groups of blocks. The data source(s) then initiate multicasting of the common groups of blocks in the appropriate order to the appropriate data targets. Since the common groups of blocks are provided in a single multicast, rather than separately, the data is provided in a more efficient manner to the data targets.

Abstract: Restoring access to running states of multiple primary computing systems onto a single computing system. The captured running states each include, or are altered to include, at least one device driver that is configured to interface with a common virtualization component that runs on the single computing system. The common virtualization component is configured to at least indirectly interface with hardware on the single computing system. The hardware potentially operates using a different interface than the device driver is configured to interface with. The system identifies a boot order for each of the primary computing systems, and then starts the running states for each of the primary computing systems in the appropriate boot order in a manner that takes advantage of the virtual environment exposed by the single computing system.

Abstract: Moving backup data within a storage hierarchy based on a calculated uniqueness of the backup data and on the estimated significance of at least a portion of the backup data. More unique and significant backup data would tend to have higher availability levels. Conversely, less unique and significant backup data would tend to have lower availability levels, or may even cause the backup data to be deleted.

Abstract: Restoring data, without suspending access to the data for the entire time that the data is being restored. Access is suspended only while a portion of the data is restored, before access to all of the data is permitted. A driver virtualizes any remaining unrestored data. To accomplish this, the driver maintains a list of data yet to be restored. If a request to read a portion of the data is then received, and the requested portion is on the list, the driver restores the requested portion before permitting the read request to be fulfilled. The restored portion is then removed from the list of unrestored data. If a request to write a portion of the data is received, the write is permitted, and the location just written to is removed from the list without restoring.

Abstract: Moving backup data within a storage hierarchy based on a calculated uniqueness of the backup data and on the estimated significance of at least a portion of the backup data. More unique and significant backup data would tend to have higher availability levels. Conversely, less unique and significant backup data would tend to have lower availability levels, or may even cause the backup data to be deleted.

Abstract: Automated derivation of an initialization ordering for computing services distributed across multiple computing systems. The initialization ordering is derived by monitoring initialization times for the computing services for one or more prior initializations of the computing services. Initialization dependencies of the computing services are then estimated based on the time of their initialization. The entire initialization ordering is then derived based on the estimated initialization dependencies.