Abstract: Persistent data, such as disks associated with virtual machines, are migrated between and across data stores using mirrored input-output operations (IOs). The migration technique employs a data mover to copy selected regions of the virtual machine disk to a target data store while the virtual machine is executing. If the virtual machine issues a write IO, the write IO is selectively mirrored on the target data store depending on whether or not the write IO is in a region of the virtual machine disk that has already been copied.

Abstract: Memory pages that are allocated to a memory consumer and continue to be accessed by the memory consumer are included in a free list, so that they may be immediately allocated to another memory consumer as needed during the course of normal operation without preserving the original contents of the memory page. When a memory page in the free list is accessed to perform a read, a generation number associated with the memory page is compared with a stored copy. If the two match, the read is performed on the memory page. If the two do not match, the read is not performed on the memory page.

Abstract: Memory pages that are allocated to a memory consumer and continue to be accessed by the memory consumer are included in a free list, so that they may be immediately allocated to another memory consumer as needed during the course of normal operation without preserving the original contents of the memory page. When a memory page in the free list is accessed to perform a read, a generation number associated with the memory page is compared with a stored copy. If the two match, the read is performed on the memory page. If the two do not match, the read is not performed on the memory page.

Abstract: A method for transferring guest physical memory from a source host to a destination host during live migration of a virtual machine (VM) involves (a) transmitting pages of the guest physical memory from the source host to the destination host over a network connection, (b) transferring state information from the source host to the destination host, (c) while performance benefits regarding continued access to the guest physical memory on the source host persist, using the transferred state information to run the VM on the destination host in place of running the VM on the source host, and (d) while the VM is running on the destination host, writing pages of the guest physical memory from the source host to a shared datastore such that the destination host can retrieve the written guest physical pages from the shared datastore.

Abstract: Memory pages that are allocated to a memory consumer and continue to be accessed by the memory consumer are included in a free list, so that they may be immediately allocated to another memory consumer as needed during the course of normal operation without preserving the original contents of the memory page. When a memory page in the free list is accessed to perform a read, a generation number associated with the memory page is compared with a stored copy. If the two match, the read is performed on the memory page. If the two do not match, the read is not performed on the memory page.

Abstract: A disclosed method includes distinguishing static and dynamic persistent data, copying the static persistent data at the source data store to the destination data store on a basis of frequencies of updates, copying the dynamic persistent data to the destination data store, updating a bitmap data structure during the copying of the dynamic persistent data to indicate which blocks of the dynamic persistent data have been modified during the copying step, identifying the blocks that have been modified during the copying step using the bitmap data structure, and iteratively copying the identified blocks to the destination data store. The method may further include determining when the number of the identified blocks is below a predetermined threshold and suspending the virtual machine prior to the next copying step.

Abstract: A method for transferring guest physical memory from a source host to a destination host during live migration of a virtual machine (VM) involves (a) transmitting pages of the guest physical memory from the source host to the destination host over a network connection, (b) transferring state information from the source host to the destination host, (c) while performance benefits regarding continued access to the guest physical memory on the source host persist, using the transferred state information to run the VM on the destination host in place of running the VM on the source host, and (d) while the VM is running on the destination host, writing pages of the guest physical memory from the source host to a shared datastore such that the destination host can retrieve the written guest physical pages from the shared datastore.

Abstract: A method for transferring guest physical memory from a source host to a destination host during live migration of a virtual machine (VM) involves creating a file on a shared datastore, the file on the shared datastore being accessible to both the source host and the destination host. Pages of the guest physical memory are transferred from the source host to the destination host over a network connection and pages of the guest physical memory are written to the file so that the destination host can retrieve the written guest physical pages from the file.

Abstract: A disclosed method includes distinguishing static and dynamic persistent data, copying the static persistent data at the source data store to the destination data store on a basis of frequencies of updates, copying the dynamic persistent data to the destination data store, updating a bitmap data structure during the copying of the dynamic persistent data to indicate which blocks of the dynamic persistent data have been modified during the copying step, identifying the blocks that have been modified during the copying step using the bitmap data structure, and iteratively copying the identified blocks to the destination data store. The method may further include determining when the number of the identified blocks is below a predetermined threshold and suspending the virtual machine prior to the next copying step.

Abstract: Techniques are disclosed for managing memory within a virtualized system that includes a memory compression cache. Generally, the virtualized system may include a hypervisor configured to use a compression cache to temporarily store memory pages that have been compressed to conserve memory space. A “first-in touch-out” (FITO) list may be used to manage the size of the compression cache by monitoring the compressed memory pages in the compression cache. Each element in the FITO list corresponds to a compressed page in the compression cache. Each element in the FITO list records a time at which the corresponding compressed page was stored in the compression cache (i.e. an age). A size of the compression cache may be adjusted based on the ages of the pages in the compression cache.

Abstract: Persistent data of virtual machines are migrated between and across data stores using a method that provides for a bounded disk size, low performance overhead, arbitrarily low downtimes, atomic switch-over, and ability to build optimizations based around block level heuristics. The method includes the steps of copying the persistent data at the source data store to the destination data store, updating a bitmap data structure during the copying step to indicate which blocks of the persistent data have been modified during the copying step, identifying the blocks that have been modified during the copying step using the bitmap data structure, and copying the identified blocks to the destination data store. The method may further include the step of determining that the number of the identified blocks is below a predetermined threshold and suspending the virtual machine prior to the second copying step.

Abstract: Techniques for migrating persistent data of virtual machines between and across data stores are optimized using special tracking data structures and monitoring methods. Special tracking data structures include an incremental change block tracking bitmap that indicate what blocks have been modified during a copy operation. The determination of whether any one block has been modified during the copy operation is based on whether or not the copy operation has progressed past that block. Another special tracking data structure is a Bloom filter, which provides a space-efficient data structure for keeping track of dirtied blocks. In addition, heat-based optimization techniques are applied so that blocks that are frequently updated are filtered and not transferred to the destination data store until the last iteration of the migration process.

Abstract: Persistent data, such as disks associated with virtual machines, are migrated between and across data stores using mirrored input-output operations (IOs). The migration technique employs a data mover to copy selected regions of the virtual machine disk to a target data store while the virtual machine is executing. If the virtual machine issues a write IO, the write IO is selectively mirrored on the target data store depending on whether or not the write IO is in a region of the virtual machine disk that has already been copied.

Abstract: Techniques are disclosed for managing memory within a virtualized system that includes a memory compression cache. Generally, the virtualized system may include a hypervisor configured to use a compression cache to temporarily store memory pages that have been compressed to conserve memory space. A “first-in touch-out” (FITO) list may be used to manage the size of the compression cache by monitoring the compressed memory pages in the compression cache. Each element in the FITO list corresponds to a compressed page in the compression cache. Each element in the FITO list records a time at which the corresponding compressed page was stored in the compression cache (i.e. an age). A size of the compression cache may be adjusted based on the ages of the pages in the compression cache.

Abstract: A method for transferring guest physical memory from a source host to a destination host during live migration of a virtual machine (VM) involves creating a file on a shared datastore, the file on the shared datastore being accessible to both the source host and the destination host. Pages of the guest physical memory are transferred from the source host to the destination host over a network connection and pages of the guest physical memory are written to the file so that the destination host can retrieve the written guest physical pages from the file.

Abstract: Persistent data of virtual machines are migrated between and across data stores using a method that provides for a bounded disk size, low performance overhead, arbitrarily low downtimes, atomic switch-over, and ability to build optimizations based around block level heuristics. The method includes the steps of copying the persistent data at the source data store to the destination data store, updating a bitmap data structure during the copying step to indicate which blocks of the persistent data have been modified during the copying step, identifying the blocks that have been modified during the copying step using the bitmap data structure, and copying the identified blocks to the destination data store. The method may further include the step of determining that the number of the identified blocks is below a predetermined threshold and suspending the virtual machine prior to the second copying step.

Abstract: Techniques for migrating persistent data of virtual machines between and across data stores are optimized using special tracking data structures and monitoring methods. Special tracking data structures include an incremental change block tracking bitmap that indicate what blocks have been modified during a copy operation. The determination of whether any one block has been modified during the copy operation is based on whether or not the copy operation has progressed past that block. Another special tracking data structure is a Bloom filter, which provides a space-efficient data structure for keeping track of dirtied blocks. In addition, heat-based optimization techniques are applied so that blocks that are frequently updated are filtered and not transferred to the destination data store until the last iteration of the migration process.