Techniques for managing a distributed computing environment using event digests

Patent metadata

Field	Value
Publication number	US-10747640-B2
Application number	US-201715635270-A
Country	US
Kind code	B2
Filing date	Jun 28, 2017
Priority date	Nov 8, 2016
Publication date	Aug 18, 2020
Grant date	Aug 18, 2020

Abstract

Techniques and apparatus for managing a distributed computing environment using event digests are described. In one embodiment, for example, an apparatus may include at least one memory, and logic for a system manager, at least a portion of the logic comprised in hardware coupled to the at least one memory, the logic to determine a workload to schedule, access an event digest associated with a plurality of compute hosts, the event digest comprising event digest values determined using out-of-band information, determine metrics from the event digest, generate at least one host weight for at least a portion of the plurality of compute hosts based on the metrics, identify at least one candidate host from the portion of the plurality of compute hosts based on the at least one host weight, and schedule the workload on the at least one candidate host. Other embodiments are described and claimed.

First claim

What is claimed is:

1. An apparatus, comprising:
at least one memory; and
logic for a system manager, at least a portion of the logic comprised in hardware coupled to the at least one memory, the logic to:
determine a workload to schedule,
access an event digest associated with a plurality of compute hosts, the event digest comprising event digest values determined using out-of-band information, at least a portion of the event digest values determined using out-of-band information during a provisioning state of a first compute host of the plurality of compute hosts,
determine metrics from the event digest,
compute at least one host weight for at least a portion of the plurality of compute hosts based on the metrics,
identify at least one candidate host from the portion of the plurality of compute hosts based on the at least one host weight, and
schedule the workload on the at least one candidate host.

2. The apparatus of claim 1 , the system manager comprising one of a cloud controller or a cloud orchestrator, the provisioning state to comprise a state prior to booting of an operating system on the first compute host, the metrics comprising a sleep state or a power state of the first compute host during the provisioning state.

3. The apparatus of claim 1 , the metrics comprising at least one of processor power, C0 residency, efficient performance indicator (EPI), uncore clocks, memory bandwidth, or core frequency.

4. The apparatus of claim 1 , the logic to:
obtain in-band information responsive to identifying a plurality of candidate hosts, and
determine a single candidate host based on the in-band information.

5. The apparatus of claim 1 , the logic to compute the at least one host weight based on an aggregation of a plurality of fields of the event digest.

6. The apparatus of claim 1 , the logic to compute the at least one weight to indicate a resource availability of the plurality of compute hosts.

7. The apparatus of claim 1 , the logic to compute the at least one weight to indicate a compatibility of the at least one compute hosts to execute the workload.

8. The apparatus of claim 1 , the logic to access the event digest via an out-of-band communication channel comprising one of an intelligent platform management interface (IPMI) communication channel or a platform environment control interface (PECI), the logic to compute a respective weight for each of the plurality of compute hosts.

9. The apparatus of claim 1 , the logic to:
monitor for an error condition based on the metrics,
detect the error condition based on the monitored metrics, and
obtain in-band metrics based on the detection of the error condition to determine a cause of the error condition.

10. The apparatus of claim 1 , the logic to:
monitor for an error condition based on the metrics,
detect the error condition based on the monitored metrics, and
obtain in-band metrics based on the detection of the error condition to determine a remedial action for the error condition.

11. A method, comprising:
determining a workload to schedule;
accessing an event digest associated with a plurality of compute hosts, the event digest comprising event digest values determined using out-of-band information, at least a portion of the event digest values determined using out-of-band information during a provisioning state of a first compute host of the plurality of compute hosts;
determining metrics from the event digest;
computing at least one host weight for at least a portion of the plurality of compute hosts based on the metrics;
identifying at least one candidate host from the portion of the plurality of compute hosts based on the at least one host weight; and
scheduling the workload on the at least one candidate host.

12. The method of claim 11 , the event digest accessed via a system manager operating on one of a cloud controller or a cloud orchestrator, the provisioning state to comprise a state prior to booting of an operating system on the first compute host, the metrics comprising a sleep state or a power state of the first compute host during the provisioning state.

13. The method of claim 11 , the metrics comprising at least one of processor power, C0 residency, efficient performance indicator (EPI), uncore clocks, memory bandwidth, or core frequency.

14. The method of claim 11 , comprising:
obtaining in-band information responsive to identifying a plurality of candidate hosts, and
determining a single candidate host based on the in-band information.

15. The method of claim 11 , comprising computing the at least one host weight based on an aggregation of a plurality of fields of the event digest.

16. The method of claim 11 , comprising computing the at least one weight to indicate a resource availability of the plurality of compute hosts.

17. The method of claim 11 , comprising computing the at least one weight to indicate a compatibility of the at least one compute hosts to execute the workload.

18. The method of claim 11 , comprising accessing the event digest via an out-of-band communication channel comprising one of an intelligent platform management interface (IPMI) communication channel or a platform environment control interface (PECI).

19. The method of claim 11 , comprising:
monitoring for an error condition based on the metrics,
detecting the error condition based on the metrics, and
obtaining in-band metrics based on the detection of the error condition to determine a cause of the error condition.

20. The method of claim 11 , comprising:
monitoring for an error condition based on the metrics,
detecting the error condition based on the metrics, and
obtaining in-band metrics based on the detection of the error condition to determine a remedial action for the error condition.

21. A non-transitory computer-readable storage medium that stores instructions for execution by processing circuitry of a system manager, the instructions to cause the system manager to:
determine a workload to schedule;
access an event digest associated with a plurality of compute hosts, the event digest comprising event digest values determined using out-of-band information, at least a portion of the event digest values determined using out-of-band information during a provisioning state of a first compute host of the plurality of compute hosts;
determine metrics from the event digest;
compute at least one host weight for at least a portion of the plurality of compute hosts based on the metrics;
identify at least one candidate host from the portion of the plurality of compute hosts based on the at least one host weight; and
schedule the workload on the at least one candidate host.

22. The computer-readable storage medium of claim 21 , the metrics comprising at least one of processor power, C0 residency, efficient performance indicator (EPI), uncore clocks, memory bandwidth, or core frequency, the provisioning state to comprise a state prior to booting of an operating system on the first compute host.

23. The computer-readable storage medium of claim 21 , the instructions to cause the system manager to:
obtain in-band information responsive to identifying a plurality of candidate hosts, and
determine a single candidate host based on the in-band information.

24. The computer-readable storage medium of claim 21 , the instructions to cause the system manager to:
monitor for an error condition based on the metrics,
detect the error con

Assignees

• Intel Corp

Inventors

• Narayan Ananth S
• Grzymkowski Lukasz
• Ganguli Mrittika
• Oriol Mariusz

Classifications

• H03K19/0175 Primary

  Coupling arrangements; Interface arrangements (interface arrangements for digital computers G06F3/00, G06F13/00) · CPC title

• G06F9/3844

  using dynamic branch prediction, e.g. using branch history tables · CPC title

• G06F9/505

  considering the load · CPC title

• H04Q9/00

  Arrangements in telecontrol or telemetry systems for selectively calling a substation from a main station, in which substation desired apparatus is selected for applying a control signal thereto or for obtaining measured values therefrom · CPC title

• G06F11/3476 Primary

  Data logging (G06F11/14, G06F11/2205 take precedence) · CPC title