blob: 53189de44d22cff9319e999fd6dcc6d7d2e527e8 [file] [log] [blame]
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
/**
* <!-- Package description. -->
* TensorFlow integration that allows to start and maintain TensorFlow cluster on top of Apache Ignite cluster
* infrastructure. The TensorFlow cluster is built for the specified cache the following way:
* <ol>
* <li>The TensorFlow cluster maintainer is created to maintain the cluster associated with the specified cache so
* that this service works reliable even the node will fail. It's achieved using Ignite Service Grid.</li>
* <li>TensorFlow cluster maintainer finds out that cluster is not started and begins starting procedure.</li>
* <li>TensorFlow cluster resolver builds cluster specification based on the specified cache so that every
* TensorFlow task is associated with partition of the cache and assumed to be started on the node where the
* partition is kept.</li>
* <li>Based on the built cluster specification the set of tasks is sent to the nodes on purpose to start TensorFlow
* servers on the nodes defined in the specification.</li>
* <li>When this set of tasks is completed successfully the started process identifiers are returned and saved
* for future using.</li>
* <li>The starting procedure is completed. In case a server fails the cluster will be turn down and then started
* again by maintainer.</li>
* </ol>
*/
package org.apache.ignite.tensorflow;