blob: 6992bd87f96fc20e1aef6bfb06fcb0f0c8276954 [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 org.apache.apex.benchmark;
import org.apache.hadoop.conf.Configuration;
import com.datatorrent.api.Context.PortContext;
import com.datatorrent.api.DAG;
import com.datatorrent.api.DAG.Locality;
import com.datatorrent.api.StreamingApplication;
import com.datatorrent.api.annotation.ApplicationAnnotation;
/**
* Performance Demo Application:
* <p>
* This demo demonstrates the performance of Datatorrent platform.
* Performance is measured by the number of events processed per second and latency.
* Performance varies depending on container memory, CPU and network I/O.
* The demo can be used to check how the performance varies with stream locality.
*
* Stream locality decides how the operators are deployed:
* ThreadLocal - the operators are deployed within the same thread.
* ContainerLocal -the operators are deployed as separate threads within the process.
* NodeLocal- the operators are deployed as separate processes on a machine.
* RackLocal - the operators are deployed on different nodes of the same rack.
* NoLocality - lets the engine decide how to best deploy the operator.
*
* Note: NodeLocal and RackLocal are preferences that can be specified to Hadoop ResourceManager.
* It is not guaranteed that the operators will be deployed as requested.
* ResourceManager makes the call depending on resource availability.
*
* Refer to demos/docs/PerformanceDemo.md for more details.
*
* </p>
*
* @since 0.9.0
*/
@ApplicationAnnotation(name = "PerformanceBenchmarkingApp")
public abstract class Benchmark
{
abstract static class AbstractApplication implements StreamingApplication
{
public static final int QUEUE_CAPACITY = 32 * 1024;
@Override
public void populateDAG(DAG dag, Configuration conf)
{
RandomWordInputModule wordGenerator = dag.addOperator("wordGenerator", RandomWordInputModule.class);
dag.getMeta(wordGenerator).getMeta(wordGenerator.output).getAttributes()
.put(PortContext.QUEUE_CAPACITY, QUEUE_CAPACITY);
WordCountOperator<byte[]> counter = dag.addOperator("counter", new WordCountOperator<byte[]>());
dag.getMeta(counter).getMeta(counter.input).getAttributes().put(PortContext.QUEUE_CAPACITY, QUEUE_CAPACITY);
dag.addStream("Generator2Counter", wordGenerator.output, counter.input).setLocality(getLocality());
}
public abstract Locality getLocality();
}
/**
* Let the engine decide how to best place the 2 operators.
*/
@ApplicationAnnotation(name = "PerformanceBenchmarkNoLocality")
public static class NoLocality extends AbstractApplication
{
@Override
public Locality getLocality()
{
return null;
}
}
/**
* Place the 2 operators so that they are in the same Rack.
* The operators are requested to be deployed on different machines.
*/
@ApplicationAnnotation(name = "PerformanceBenchmarkRackLocal")
public static class RackLocal extends AbstractApplication
{
@Override
public Locality getLocality()
{
return Locality.RACK_LOCAL;
}
}
/**
* Place the 2 operators so that they are in the same node.
* The operators are requested to be deployed as different processes within the same machine.
*/
@ApplicationAnnotation(name = "PerformanceBenchmarkNodeLocal")
public static class NodeLocal extends AbstractApplication
{
@Override
public Locality getLocality()
{
return Locality.NODE_LOCAL;
}
}
/**
* Place the 2 operators so that they are in the same container.
* The operators are deployed as different threads in the same process.
*/
@ApplicationAnnotation(name = "PerformanceBenchmarkContainerLocal")
public static class ContainerLocal extends AbstractApplication
{
@Override
public Locality getLocality()
{
return Locality.CONTAINER_LOCAL;
}
}
/**
* Place the 2 operators so that they are in the same thread.
*/
@ApplicationAnnotation(name = "PerformanceBenchmarkThreadLocal")
public static class ThreadLocal extends AbstractApplication
{
@Override
public Locality getLocality()
{
return Locality.THREAD_LOCAL;
}
}
}