Abstract: A storage node such as a storage array or storage server has storage controllers with heterogeneous parallel processors. The CPUs may be used to perform read and write operations. GPUs are used to produce transformed data from raw data. The GPUs may be used to perform various analytic calculations. The GPUs may be interconnected via dedicated communication links.

Abstract: Operating at least one hypervisor includes running a first hypervisor as a first thread of an underlying operating system, running a second hypervisor as a second thread of the underlying operating system, loading a first guest operating system using the first hypervisor based on the first thread of the underlying operating system, loading a second guest operating system using the second hypervisor based on the second thread of the underlying operating system, and scheduling sharing of resources of the underlying system between the first hypervisor and the second hypervisor according to a scheduler of the underlying operating system, where the first hypervisor and the second hypervisor run independently of each other. The scheduler of the underlying operating system may schedule fractional time shares for the first hypervisor and the second hypervisor to access the same resource.

Abstract: Operating a hypervisor includes running a hypervisor as a thread of an underlying operating system and loading a guest operating system using the hypervisor based on the thread of the underlying operating system, where the hypervisor runs independently of the guest operating system and independently of other hypervisors running as other threads of the underlying operating system. The hypervisor may be a first hypervisor and operating a hypervisor may further include running a second hypervisor nested with the first hypervisor, where the guest operating system may be loaded using both the first hypervisor and the second hypervisor. The underlying operating system may be an operating system of a storage system.

Abstract: When a guest OS loads within the context of a container provided by the host OS, the guest OS uses PCI or other protocol to specify a virtual hardware device. The guest OS enumerates the virtual hardware device to establish the size for the BARs and establish its view of physical addresses for the memory locations. A server running in the context of the container receives read/write requests from the guest OS, maps the read/write requests to host OS physical address space, and posts responses to the virtual hardware device. Since the guest OS executes memory related operations using its own memory space, exits to the container code are not required to implement storage related actions by the Guest OS. This allows performance of an application executing in the context of the guest OS to approximate performance of an application executing in the context of the host OS.