Distributed computing in a process control environment
AssigneeSchneider Electric Systems USA, Inc.
InventorsRaja Ramana Macha, Andrew Lee David Kling, Nestor Jesus Camino, Jr., James Gerard Luth, Richard Linwood Linscott
Life of this patent
Estimating the term…
Abstract
High availability and data migration in a distributed process control computing environment. Allocation algorithms distribute data and applications among available compute nodes, such as controllers in a process control system. In the process control system, an input/output device, such as a fieldbus module, can be used by any controller. Databases store critical execution information for immediate takeover by a backup compute element. The compute nodes are configured to execute algorithms for mitigating dead time in the distributed computing environment.
Claims20
An apparatus for controlling a process, comprising:
- an electronic data communication channel;
- a plurality of compute nodes each connected to the communication channel and storing control datasets associated therewith, wherein a first compute node of the plurality of compute nodes implements a process control application;
- and
- an executive node connected to the communication channel, the executive node including an application manager configured to:
- detect a failure of the first compute node;
- deploy, from the executive node in response to the detected failure, an input/output (I/O) or other application of the failed first compute node to a second compute node of the plurality of compute nodes, wherein the second compute node does not establish a peer-to-peer connection with the failed first compute node;
- receive one or more selected subsets of control datasets from the failed first compute node and transmit the selected subsets to the second compute node;
- and
- instruct the second compute node to execute the selected subsets of the control datasets at the I/O or other application deployed to the second compute node, the selected subsets being associated with the failed first compute node and comprising less than entire control datasets of the plurality of compute nodes.
A method for controlling a process, comprising:
- implementing, by a first compute node of a plurality of compute nodes, a process control application, wherein the compute nodes store control datasets associated therewith and are connected to an electronic data communication channel;
- detecting, by an application manager of an executive node, a failure of the first compute node, wherein the executive node is connected to the communication channel;
- deploying, by the application manager of the executive node in response to the detected failure, an input/output (I/O) or other application of the failed first compute node to a second compute node of the compute nodes, wherein the second compute node does not establish a peer-to-peer connection with the failed first compute node;
- receiving one or more selected subsets of control datasets from the failed first compute node;
- transmitting the selected subsets to the second compute node;
- and
- instructing, by the application manager of the executive node, the second compute node to execute the selected subsets of the control datasets at the I/O or other application deployed to the second compute node, wherein the selected subsets are associated with the failed first compute node and comprise less than entire control datasets.
Dependencies are inferred from the claim text.