Technologies for dynamically sharing remote resources across remote computing nodes

Patent metadata

Field	Value
Publication number	US-2019004858-A1
Application number	US-201715636969-A
Country	US
Kind code	A1
Filing date	Jun 29, 2017
Priority date	Jun 29, 2017
Publication date	Jan 3, 2019
Grant date	—

Abstract

Technologies for dynamically sharing remote resources include a computing node that sends a resource request for remote resources to a remote computing node in response to a determination that additional resources are required by the computing node. The computing node configures a mapping of a local address space of the computing node to the remote resources of the remote computing node in response to sending the resource request. In response to generating an access to the local address, the computing node identifies the remote computing node based on the local address with the mapping of the local address space to the remote resources of the remote computing node and performs a resource access operation with the remote computing node over a network fabric. The remote computing node may be identified with system address decoders of a caching agent and a host fabric interface. Other embodiments are described and claimed.

First claim

1 . A computing node for remote resource usage, the computing node comprising:
one or more processors; one or more memory devices having stored therein a plurality of instructions that, when executed by the one or more processors, cause the computing node to:
determine whether additional resources are required by the computing node;
send a resource request for remote resources to a remote computing node in response to a determination that the additional resources are required by the computing node;
configure a mapping of a local address space of the computing node to the remote resources of the remote computing node in response to sending of the resource request; and
generate an access to a local address in the local address space in response to configuration of the mapping; and
a network interface controller to:
identify the remote computing node based on the local address with the mapping of the local address space to the remote resources of the remote computing node in response to generation of the access to the local address; and
perform a resource access operation with the remote computing node over a network fabric in response to identification of the remote computing node.

2 . The computing node of claim 1 , wherein the plurality of instructions, when executed, further cause the computing node to monitor for available resources advertised by the remote computing node over the network fabric.

3 . The computing node of claim 1 , wherein to configure the mapping of the local address space to the remote resources comprises to configure one or more system address decoders of the computing node to map the local address space to the remote resources of the remote computing node.

4 . The computing node of claim 3 , wherein to configure one or more system address decoders comprises to:
configure a first system address decoder of a caching agent of the computing node to map the local address space to a host fabric interface of the computing node; and configure a second system address decoder of the host fabric interface to map the local address space to the remote resources of the remote computing node.

5 . The computing node of claim 3 , wherein to configure the mapping of the local address space to the remote resources further comprises to send a hot-plug notification to an operating system of the computing node in response to configuration of the one or more system address decoders.

6 . The computing node of claim 1 , further comprising a host fabric interface, wherein the host fabric interface comprises the network interface controller.

7 . The computing node of claim 1 , further comprising a machine learning resource borrower to determine whether additional resources are required with a machine learning algorithm.

8 . The computing node of claim 7 , wherein the machine learning algorithm is based on one or more of an application need, a service level objective, a data center policy, a dynamic condition, or a resource demand history.

9 . The computing node of claim 1 , wherein the plurality of instructions, when executed, further cause the computing node to:
receive a release request, from the remote computing node, to release resources; and remove the mapping of the local address space to the remote resources in response to a receipt of the release request.

10 . The computing node of claim 9 , wherein the plurality of instructions, when executed, further cause the computing node to reflow data from the remote resources of the remote computing node to local resources of the computing node in response to removal of the mapping.

11 . The computing node of claim 1 , wherein the additional resources comprises memory resources, the remote resources comprises memory resources of the remote computing node, and the local address space comprises a memory address space of the computing node.

12 . The computing node of claim 1 , wherein the additional resources comprises storage resources, the remote resources comprises storage resources of the remote computing node, and the local address space comprises a storage address space of the computing node.

13 . One or more machine-readable storage media comprising a plurality of instructions stored thereon that, in response to being executed, cause a computing node to:
determine whether additional resources are required by the computing node; send a resource request for remote resources to a remote computing node in response to determining that the additional resources are required by the computing node; configure a mapping of a local address space of the computing node to the remote resources of the remote computing node in response to sending the resource request; generate an access to a local address in the local address space in response to configuring the mapping; identify the remote computing node based on the local address using the mapping of the local address space to the remote resources of the remote computing node in response to generating the access to the local address; and perform a resource access operation with the remote computing node over a network fabric in response to identifying the remote computing node.

14 . The one or more machine-readable storage media of claim 13 , wherein the plurality of instructions, when executed, further cause the computing node to monitor for available resources advertised by the remote computing node over the network fabric.

15 . The one or more machine-readable storage media of claim 13 , wherein to configure the mapping of the local address space to the remote resources comprises to configure one or more system address decoders of the computing node to map the local address space to the remote resources of the remote computing node.

16 . The one or more machine-readable storage media of claim 15 , wherein to configure the one or more system address decoders comprises to:
configure a first system address decoder of a caching agent of the computing node to map the local address space to a host fabric interface of the computing node; and configure a second system address decoder of the host fabric interface to map the local address space to the remote resources of the remote computing node.

17 . The one or more machine-readable storage media of claim 15 , wherein to configure the mapping of the local address space to the remote resources further comprises to send a hot-plug notification to an operating system of the computing node in response to configuring the one or more system address decoders.

18 . The one or more machine-readable storage media of claim 13 , wherein:
to identify the remote computing node based on the local address using the mapping comprises to identify, by a host fabric interface of the computing node, the remote computing node based on the local address using the mapping; and to perform the resource access operation with the remote computing node over the network fabric comprises to perform, by the host fabric interface, the resource access operation with the remote computing node.

19 . The one or more machine-readable storage media of claim 13 , wherein to determine whether additional resources are required comprises to determine whether additional resources are required with a machine learning algorithm.

20 . The one or more machine-readable storage media of claim 19 , wherein the machine learning algorithm is based on one or more of an application need, a service level objective, a data center policy, a dynamic condition, or a resource demand history.

Assignees

• Intel Corp

Inventors

• Bernat Francesc Guim
• Doshi Kshitij A
• Barragan Daniel Rivas
• Gonzalez Alejandro Duran
• Servat Harald

Classifications

• G06F9/5016 Primary

  the resource being the memory · CPC title

• G06F9/5005 Primary

  to service a request · CPC title

• H04L61/103

  across network layers, e.g. resolution of network layer into physical layer addresses or address resolution protocol [ARP] · CPC title

• H04W8/22

  Processing or transfer of terminal data, e.g. status or physical capabilities · CPC title

• H04L45/745

  Address table lookup; Address filtering · CPC title