Resource migration negotiation
Resource migration negotiation is disclosed. A request is received, from a remote physical node in a plurality of physical nodes, for a resource. An operating system is run collectively across the plurality of physical nodes. The request includes information pertaining to a guest thread running on the remote physical node. Based at least in part on at least some of the information included in the request, it is determined whether to send the requested resource or reject the request. A response is provided based at least in part on the determination.
1. A computer system, comprising:
a plurality of physically interconnected computing nodes, wherein a guest operating system is run on a virtual environment that is defined by a set of hyper-kernels running on the plurality of interconnected computing nodes;
wherein a request for a resource made by a hyper-kernel on a first computing node is received at a second computing node that owns the resource, wherein the request made by the hyper-kernel on the first computing node includes metadata information associated with the requested resource, and wherein the request for the resource was made by the hyper-kernel on the first computing node based at least in part on an evaluation of a set of requestor-side cost functions;
wherein in response to receiving the request for the resource made by the hyper-kernel on the first computing node, a hyper-kernel on the second computing node determines, at least in part by evaluating a set of owner-side cost functions, whether to send the resource or reject the request, wherein evaluating the set of owner-side cost functions by the hyper-kernel on the second computing node comprises evaluating the metadata information included in the request made by the hyper-kernel on the first computing node; and
wherein the hyper-kernel on the second computing node provides a response based at least in part on the determination of whether to send the resource or reject the request.
2. The computer system recited in claim 1 wherein the hyper-kernel on the second computing node determines, based at least in part on the evaluating of the set of owner-side cost functions, a cost of sending the resource and a cost of rejecting the request.
3. The computer system recited in claim 2 wherein the cost of sending the resource is determined to be less than the cost of rejecting the request, and wherein the response comprises sending the resource to the first computing node.
4. The computer system recited in claim 2 wherein the cost of rejecting the request is determined to be less than the cost of sending the resource, and wherein the response comprises sending, to the first computing node, a message indicating rejection of the request.
5. The computer system recited in claim 4 wherein the message includes a reason for rejection of the request.
6. The computer system recited in claim 4 wherein based at least in part on the message indicating rejection of the request, the hyper-kernel on the first computing node migrates a virtual processor to the second computing node.
7. The computer system recited in claim 2 wherein an owner-side cost function included in the set of owner-side cost functions comprises one or more owner-side cost terms.
8. The computer system recited in claim 7 wherein the requested resource comprises a portion of memory, and wherein an owner-side cost term is associated with an indication of whether the portion of memory is in use by a non-moveable device.
9. The computer system recited in claim 7 wherein the request is associated with a requesting guest thread running on the first computing node, and wherein an owner-side cost term included in the owner-side cost function is associated with an efficiency of the requesting guest thread.
10. The computer system recited in claim 7 wherein the requested resource comprises a portion of memory, and wherein an owner-side cost term is associated with a warmth of the portion of memory.
11. A method, comprising:
receiving a request for a resource made by a hyper-kernel on a first computing node, wherein the request is received at a second computing node that owns the resource, wherein the request made by the hyper-kernel on the first computing node includes metadata information associated with the requested resource, wherein the request for the resource was made by the hyper-kernel on the first computing node based at least in part on an evaluation of a set of requestor-side cost functions, wherein the first computing node and the second computing node are included in a plurality of physically interconnected computing nodes, and wherein a guest operating system is run on a virtual environment that is defined by a set of hyper-kernels running on the plurality of interconnected computing nodes;
in response to receiving the request for the resource made by the hyper-kernel on the first computing node, determining, by a hyper-kernel on the second computing node, whether to send the resource or reject the request, wherein whether to send the resource or reject the request is determined at least in part by evaluating a set of owner-side cost functions, wherein evaluating the set of owner-side cost functions by the hyper-kernel on the second computing node comprises evaluating the metadata information included in the request made by the hyper-kernel on the first computing node; and
providing, by the hyper-kernel on the second computing node, a response based at least in part on the determination of whether to send the resource or reject the request.
12. The method of claim 11 further comprising, based at least in part on the evaluating of the set of owner-side cost functions, determining, by the hyper-kernel on the second computing node determines, a cost of sending the resource and a cost of rejecting the request.
13. The method of claim 12 wherein the cost of sending the resource is determined to be less than the cost of rejecting the request, and wherein the response comprises sending the resource to the first computing node.
14. The method of claim 12 wherein the cost of rejecting the request is determined to be less than the cost of sending the resource, and wherein the response comprises sending, to the first computing node, a message indicating rejection of the request.
15. The method of claim 14 wherein the message includes a reason for rejection of the request.
16. The method of claim 14 wherein based at least in part on the message indicating rejection of the request, the hyper-kernel on the first computing node migrates a virtual processor to the second computing node.
17. The method of claim 12 wherein an owner-side cost function included in the set of owner-side cost functions comprises one or more owner-side cost terms.
18. The method of claim 17 wherein the requested resource comprises a portion of memory, and wherein an owner-side cost term is associated with an indication of whether the portion of memory is in use by a non-moveable device.
19. The method of claim 17 wherein the request is associated with a requesting guest thread running on the first computing node, and wherein an owner-side cost term included in the owner-side cost function is associated with an efficiency of the requesting guest thread.
20. The method of claim 17 wherein the requested resource comprises a portion of memory, and wherein an owner-side cost term is associated with a warmth of the portion of memory.