US 10,853,125 B2Granted

Resource efficient acceleration of datastream analytics processing using an analytics accelerator

AssigneeOracle International Corporation

InventorsKarthik Ganesan, Shrinivas B. Joshi, Yao-Min Chen, Luyang Wang, Ahmed Khawaja

Life of this patent

Estimating the term…

Loading patent…

Abstract

An offload engine may attempt to offload, on behalf of applications, data operations to be performed on a datastream. The offload engine may intercept one or more data operations, such as may be part of a collections API, performed by an application. The data operations and the datastream may be specified and/or provided by the application and, in response, the offload engine may be configured to execute (or attempt to execute) the data operations on the datastream using an analytics accelerating co-processor rather than using a general purpose CPU core. The offload engine may determine whether or not to offload the data operations to the analytics accelerating co-processor. If the offload is unsuccessful or if the offload engine determines that the data operations are not suitable for offloading, the offload engine may then cause the data operations to be performed using general purpose CPU cores on the system.

Claims20

Dependent claims

A system, comprising:

one or more general purpose central processing unit (CPU) cores;
an analytics accelerating co-processor;
and
a memory comprising program instructions that, when executed by the one or more general purpose CPU cores cause the one or more general purpose CPU cores to implement a platform-independent, object-oriented, runtime environment;
wherein the runtime environment comprises an offload engine configured to:
intercept one or more method calls invoked by an application executing within the runtime environment, wherein the one or more method calls are configured to perform one or more data operations on a stream of data using the one or more general purpose CPU cores;
in response to intercepting the one or more method calls, execute the one or more data operations on the stream of data;
wherein to execute the one or more data operations on the stream of data, the offload engine is configured to:
attempt to perform the one or more data operations on the stream of data using the analytics accelerating co-processor;
and
determine whether the attempt to perform the one or more data operations on the stream of data using the analytics accelerating co-processor was unsuccessful based at least in part on determining whether a garbage collection operation is required.

8 more elements

A method, comprising:

performing by one or more computers:
executing, by an offload engine executing of a platform-independent, object-oriented, runtime environment, one or more data operations on a stream of data in response to intercepting one or more method calls invoked by an application executing within the runtime environment, wherein the one or more intercepted method calls are configured to perform the one or more data operations on the stream of data using one or more general purpose CPU cores of the one or more computers;
wherein said executing comprises:
attempting, by the offload engine, to perform the one or more data operations on the stream of data using an analytics accelerating co-processor;
and
determining whether the attempt to perform the one or more data operations on the stream of data using the analytics accelerating co-processor was unsuccessful based at least in part on determining whether a garbage collection operation is required.

3 more elements

A non-transitory, computer-readable storage medium storing program instructions that when executed on one or more computers cause the one or more computers to perform:

executing, by an offload engine executing of a platform-independent, object-oriented, runtime environment, one or more data operations on a stream of data in response to intercepting one or more method calls invoked by an application executing within the runtime environment, wherein the one or more intercepted method calls are configured to perform the one or more data operations on the stream of data using one or more general purpose CPU cores of the one or more computers;
wherein said executing comprises:
attempting, by the offload engine, to perform the one or more data operations on the stream of data using an analytics accelerating co-processor;
and
determining whether the attempt to perform the one or more data operations on the stream of data using the analytics accelerating co-processor was unsuccessful based at least in part on determining whether a garbage collection operation is required.

2 more elements

Dependencies are inferred from the claim text.

Facts

Application: 15/287,695Filed 6 Oct 2016 · us provisional application
Priority: 19 Aug 2016
Published: US 2018/0052708 A122 Feb 2018
Granted: 1 Dec 2020
Term ends: Not estimatedThe source records don’t support an estimate.

Classification7 codes

G06F 9/485
By program, e.g. task dispatcher, supervisor, operating system: task life-cycle, e.g. stopping, restarting, resuming execution
G06F 9/4812
Task transfer initiation or dispatching: by interrupt, e.g. masked
G06F 9/5044
The resource being a machine, e.g. CPUs, Servers, Terminals: considering hardware capabilities
G06F 11/1482
Generic software techniques for error detection or fault masking: using middleware or operating system [OS] functionalities