Common contiguous memory region optimized virtual machine migration within a workgroup
Embodiments of the invention relate to scanning, by a first processor in a work group, a memory associated with the first processor for data. The first processor updates a first data structure to include at least a portion of the data based on the scanning. The first processor transmits a representation of the first data structure to one or more peer processors of the first processor included in the work group using a dedicated link. The first processor receives a representation of a second data structure associated with at least one of the one or more peer processors of the first processor. The first processor updates the first data structure based on the received representation of the second data structure.
1. A method comprising:
scanning, by a first processor in a work group, a memory associated with the first processor for data;
updating, by the first processor, a first data structure to include at least a portion of the data based on the scanning, the first data structure comprising a plurality of entries that each include a first indicator that indicates whether a fraction of a page of memory is being referenced and a second indicator that indicates a size of a chunk of the contiguous memory representing the fraction of the page size when the second indicator is taken together with the first indicator;
transmitting, by the first processor, a representation of the first data structure to one or more peer processors of the first processor included in the work group using a dedicated link;
receiving, by the first processor, a representation of a second data structure comprising at least the first indicator and the second indicator associated with at least one of the one or more peer processors of the first processor; and
updating, by the first processor, the first data structure based on the received representation of the second data structure.
2. The method of claim 1 , wherein the representation of the first data structure corresponds to a first mnemonic, and wherein the representation of the second data structure corresponds to a second mnemonic.
3. The method of claim 2 , wherein the first mnemonic provides an indication of the first processor and the first data structure, and wherein the second mnemonic provides an indication of the at least one of the one or more peer processors of the first processor and the second data structure.
4. The method of claim 1 , further comprising:
listening, by the first processor, for a migration request;
completing, by the first processor, a migration based on the migration request; and
updating, by the first processor, the first data structure based on the migration.
5. The method of claim 4 , wherein the migration comprises a migration of a virtual machine from at least one of the one or more peer processors of the first processor to the first processor.
6. The method of claim 5 , wherein updating the first data structure based on the migration comprises adding data that is associated with the virtual machine that was not present at the first processor prior to the migration.
7. The method of claim 4 , further comprising:
subsequent to updating the first data structure based on the migration, transmitting, by the first processor, a second representation of the first data structure to the one or more peer processors of the first processor included in the work group.
8. The method of claim 1 , further comprising:
transmitting, by the first processor, a page size associated with the first processor over the dedicated link to the one or more peer processors of the first processor.
9. The method of claim 1 , wherein the dedicated link comprises a Transmission Control Protocol/Internet Protocol (TCP/IP) link, the method further comprising:
transmitting, by the first processor, the representation of the first data structure over the TCP/IP link.
10. The method of claim 1 , wherein each of a plurality of entries in the first data structure further comprises at least one of:
a third indicator that indicates the payload of the chunk of the memory,
a fourth indicator that identifies the first processor, and
a fifth indicator that indicates a number of times that the payload is replicated on the first processor.
11. The method of claim 1 , further comprising:
generating, by the first processor, a score that reflects a difference between data stored by the first processor and data stored by the at least one of the one or more peer processors of the first processor; and
sharing the score with the one or more peer processors of the first processor included in the work group.
12. The method of claim 1 , wherein the updating of the first data structure to include the at least a portion of the data based on the scanning comprises adding data to the first data structure that was not present in the first data structure prior to performing the scanning.