Merge branch 'master' of https://github.com/DSC-SPIDAL/twister2

Author: pulasthi

docker/standalone/Dockerfile

@@ -3,11 +3,15 @@ FROM ubuntu:18.04
 
 # install dependencies
 RUN apt-get update && apt-get install -y --no-install-recommends openjdk-8-jdk sudo  \
-	python-dev python-pip g++ git build-essential  \
-	automake cmake libtool-bin zip libunwind-setjmp0-dev \
-    zlib1g-dev unzip pkg-config python-setuptools \
-    maven wget  ssh && \
-    rm -rf /var/lib/apt/lists/* && apt autoclean && apt-get clean && apt autoremove
+    python3-dev python3-pip g++ git build-essential  \
+    automake cmake libtool-bin zip libunwind-setjmp0-dev \
+    zlib1g-dev unzip pkg-config python3-setuptools \
+    maven wget  ssh \
+    && cd /usr/local/bin \
+    && ln -s /usr/bin/python3 python \
+    && pip3 install --upgrade pip \
+    && sudo pip3 install wheel \
+    && rm -rf /var/lib/apt/lists/* && apt autoclean && apt-get clean && apt autoremove
 
 
 # add user
@@ -20,32 +24,35 @@ WORKDIR /home/twister2
 
 #install openmpi
 RUN cd && \
-	wget https://www.open-mpi.org/software/ompi/v3.1/downloads/openmpi-3.1.2.tar.gz && \
-	tar -zxvf openmpi-3.1.2.tar.gz && \
-	OMPI_BUILD=~/openmpi && \
-	OMPI_312=~/openmpi-3.1.2 && \
-	PATH=${OMPI_BUILD}/bin:${PATH} && \
-	LD_LIBRARY_PATH=${OMPI_BUILD}/lib:${LD_LIBRARY_PATH} && \
-	cd $OMPI_312 && \
-	./configure --prefix=$OMPI_BUILD --enable-mpi-java --enable-mpirun-prefix-by-default --enable-orterun-prefix-by-default && make;make install && \
-	echo "export OMPI_BUILD=$OMPI_BUILD" >> ~/.bashrc && \
-	echo "export PATH=\${OMPI_BUILD}/bin:\${PATH}" >> ~/.bashrc && \
-	echo "export LD_LIBRARY_PATH=\${OMPI_BUILD}/lib:\${LD_LIBRARY_PATH}" >> ~/.bashrc && \
-	rm -rf ~/openmpi-3.1.2.tar.gz $OMPI_312
+    wget https://download.open-mpi.org/release/open-mpi/v4.0/openmpi-4.0.1.tar.gz && \
+    tar -zxvf openmpi-4.0.1.tar.gz && \
+    OMPI_BUILD=~/openmpi && \
+    OMPI_401=~/openmpi-4.0.1 && \
+    PATH=${OMPI_BUILD}/bin:${PATH} && \
+    LD_LIBRARY_PATH=${OMPI_BUILD}/lib:${LD_LIBRARY_PATH} && \
+    cd $OMPI_401 && \
+    ./configure --prefix=$OMPI_BUILD --enable-mpi-java --enable-mpirun-prefix-by-default --enable-orterun-prefix-by-default && make;make install && \
+    echo "export OMPI_BUILD=$OMPI_BUILD" >> ~/.bashrc && \
+    echo "export PATH=\${OMPI_BUILD}/bin:\${PATH}" >> ~/.bashrc && \
+    echo "export LD_LIBRARY_PATH=\${OMPI_BUILD}/lib:\${LD_LIBRARY_PATH}" >> ~/.bashrc && \
+    rm -rf ~/openmpi-4.0.1.tar.gz $OMPI_401
 
 
 #install bazel
-RUN wget https://github.com/bazelbuild/bazel/releases/download/0.18.1/bazel-0.18.1-installer-linux-x86_64.sh && \
-	chmod +x bazel-0.18.1-installer-linux-x86_64.sh && \
-	./bazel-0.18.1-installer-linux-x86_64.sh --user && \
-	echo "export PATH=~/.bazel/bin:\$PATH" >> ~/.bashrc && \
-	rm bazel-0.18.1-installer-linux-x86_64.sh
+RUN wget https://github.com/bazelbuild/bazel/releases/download/0.28.1/bazel-0.28.1-installer-linux-x86_64.sh && \
+    chmod +x bazel-0.28.1-installer-linux-x86_64.sh && \
+    ./bazel-0.28.1-installer-linux-x86_64.sh --user && \
+    echo "export PATH=~/.bazel/bin:\$PATH" >> ~/.bashrc && \
+    rm bazel-0.28.1-installer-linux-x86_64.sh
+
 
 
 # clone & build twister2
 RUN git clone https://github.com/DSC-SPIDAL/twister2.git && \
-	cd twister2 && ~/.bazel/bin/bazel build --config=ubuntu scripts/package:tarpkgs && cd && \
-	tar -xzvf twister2/bazel-bin/scripts/package/twister2-0.3.0.tar.gz && \
-	echo "export JAVA_HOME=/usr/lib/jvm/java-1.8.0-openjdk-amd64" >> ~/.bashrc && \
-	sed -i '/twister2.resource.scheduler.mpi.mpirun.file/c\twister2.resource.scheduler.mpi.mpirun.file: "mpirun"' /home/twister2/twister2-0.3.0/conf/standalone/resource.yaml && \
-	rm -rf bin ~/.cache/bazel
+    cd twister2 && ~/.bazel/bin/bazel build --config=ubuntu scripts/package:tarpkgs && cd && \
+    tar -xzvf twister2/bazel-bin/scripts/package/twister2-0.3.0.tar.gz && \
+    echo "export JAVA_HOME=/usr/lib/jvm/java-1.8.0-openjdk-amd64" >> ~/.bashrc && \
+    sed -i '/twister2.resource.scheduler.mpi.mpirun.file/c\twister2.resource.scheduler.mpi.mpirun.file: "mpirun"' /home/twister2/twister2-0.3.0/conf/standalone/resource.yaml && \
+    rm -rf bin ~/.cache/bazel
+
+

docs/README.md

@@ -1,4 +1,4 @@
-# We have moved the documentation to [https://dsc-spidal.github.io/twister2](https://dsc-spidal.github.io/twister2)
+# We have moved the documentation to [https://twister2.org](https://twister2.org)
 
 This documentation is written using mark down files. These markdown files are converted into HTML and deployed on to github pages.
 
@@ -79,4 +79,4 @@ id: img_proc
 title: Image Processing
 ```
 
-Inside website folder there is a file called, sidebars.json. You can modify that file to include the documentation in the sidebar.
+Inside website folder there is a file called, sidebars.json. You can modify that file to include the documentation in the sidebar.

docs/docs/concepts/operator-api.md

@@ -36,8 +36,14 @@ to a single value at the target. How to reduce multiple values to a single value
 a user defined function.
 
 ```java
+LogicalPlanBuilder logicalPlanBuilder = LogicalPlanBuilder.plan(
+  3, // 3 sources
+  3, // 3 targets
+  workerEnv
+).withFairDistribution();
+
 // setup the batch operation
-BReduce reduce = new BReduce(communicator, logicalPlan, sources, target,
+BReduce reduce = new BReduce(communicator, logicalPlanBuilder,
     reduceFunction, reduceReceiver, datatype);
 // send the data
 reduce.reduce(task, data, flag);

docs/docs/concepts/task-system/task-scheduler/batch.md

@@ -103,3 +103,21 @@ Finally, the algorithm pack the task instance plan and the container plan into t
 and return the same.
 
 [Data Locality Batch Task Scheduler Source Code](https://github.com/DSC-SPIDAL/twister2/blob/master/twister2/taskscheduler/src/java/edu/iu/dsc/tws/tsched/batch/datalocalityaware/DataLocalityBatchTaskScheduler.java)
+
+
+## Batch Task Scheduler
+
+Batch Task Scheduler is capable of scheduling of single task graph as well as multiple graphs which
+depends on the input from other task graphs. If the batch task scheduler receives only single task
+graph, first it will check whether it has any receptor and collector tasks in the graph. If the receptor
+and collector task doesn't match the parallelism it throws the runtime exception to the user to provide
+the same parallelism for the dependent tasks in the graph. If the batch task scheduler receives multiple
+task graph, first it will store the collectible name set and receivable name set in the appropriate 
+set values. If the collectible name set (input key) matches with the receivable name set (input key) 
+it validate the parallelism of the dependent tasks in the task graph, if it doesn't match it will 
+guide the user to specify the same parallelism. If it matches, it proceeds with the scheduling of the 
+task graphs to the same workers. For example, if the map task, 0th task has the data in worker 0, it 
+will schedule the reduce task 0th task to the worker 1. Batch Task Scheduler considers both the locality 
+of the data and scheduling the tasks in a round robin fashion. 
+
+[Batch Task Scheduler Source Code](https://github.com/DSC-SPIDAL/twister2/blob/master/twister2/taskscheduler/src/java/edu/iu/dsc/tws/tsched/batch/batchscheduler/BatchTaskScheduler.java)

docs/docs/concepts/task-system/task-scheduler/task-scheduler.md

@@ -26,7 +26,7 @@ that consists of two main methods namely initialize and schedule methods to init
 configuration and schedule the taskgraph based on the workers information available in the resource plan.
 
 ```text
-edu.iu.dsc.tws.tsched.spi.taskschedule.ITaskScheduler
+edu.iu.dsc.tws.api.compute.schedule.ITaskScheduler
 ```
 
 has the following methods namely
@@ -71,9 +71,13 @@ container id, task instance plan, required and scheduled resource of the contain
 plan holds the jobid or the taskgraph id and the container plan. The task schedule plan list is mainly 
 responsible for holding the taskschedule of the batch tasks.
 
-\`\`bash message Resource { double availableCPU = 1; double availableMemory = 2; double availableDisk = 3; }
-
 ```text
+message Resource { 
+   double availableCPU = 1; 
+   double availableMemory = 2; 
+   double availableDisk = 3; 
+}
+
 message TaskInstancePlan {
    int32 taskid = 1;
    string taskname = 2;
@@ -99,8 +103,6 @@ message TaskSchedulePlanList {
 }
 ```
 
-\`\`
-
 ## YAML file
 
 The task scheduler has task.yaml in the config directory. The task scheduler mode represents either 
@@ -109,49 +111,84 @@ memory, disk, and cpu values assigned to the task instances. The default contain
 represents the percentage of values to be added to each container. The default container instance 
 values represents the default size of memory, disk, and cpu of the container. The task parallelism 
 represents the default parallelism value assigned to each task instance. The task type represents 
-the streaming or batch task.The task scheduler dynamically loads the respective streaming and batch 
+the streaming or batch task. The task scheduler dynamically loads the respective streaming and batch 
 task schedulers based on the configuration values specified in the task.yaml.
 
-\`\`yaml
-
 ```text
-#Streaming Task Scheduler Mode "roundrobin" or  "firstfit" or "datalocalityaware"
-twister2.streaming.taskscheduler: "roundrobin"
+# Task scheduling mode for the streaming jobs "roundrobin" or "firstfit" or "datalocalityaware" or "userdefined"
+twister2.taskscheduler.streaming: "roundrobin"
 
-#Streaming Task Scheduler Class
-twister2.streaming.taskscheduler.class: "edu.iu.dsc.tws.tsched.streaming.roundrobin.RoundRobinTaskScheduler"
+# Task Scheduler class for the round robin streaming task scheduler
+twister2.taskscheduler.streaming.class: "edu.iu.dsc.tws.tsched.streaming.roundrobin.RoundRobinTaskScheduler"
 
-#Batch Task Scheduler Mode "roundrobin" or  "datalocalityaware"
-twister2.batch.taskscheduler: "roundrobin"
+# Task scheduling mode for the batch jobs "roundrobin" or "datalocalityaware" or "userdefined" or "batchscheduler"
+twister2.taskscheduler.batch: "roundrobin"
 
-#Batch Task Scheduler Class
-twister2.batch.taskscheduler.class: "edu.iu.dsc.tws.tsched.batch.roundrobin.RoundRobinBatchTaskScheduler"
+# Task Scheduler class for the round robin batch task scheduler
+twister2.taskscheduler.batch.class: "edu.iu.dsc.tws.tsched.batch.roundrobin.RoundRobinBatchTaskScheduler"
 
-#Default Task Instance Values
-twister2.task.instances: 2
-twister2.task.instance.ram: 512.0
-twister2.task.instance.disk: 500.0
-twister2.task.instance.cpu: 2.0
+# Number of task instances to be allocated to each worker/container
+twister2.taskscheduler.task.instances: 2
 
-#Default Container Padding Values
-twister2.ram.padding.container: 2.0
-twister2.disk.padding.container: 12.0
-twister2.cpu.padding.container: 1.0
-twister2.container.padding.percentage: 2
+# Ram value to be allocated to each task instance
+twister2.taskscheduler.task.instance.ram: 512.0
 
-#Default Container Instance Values
-twister2.container.instance.ram: 2048.0
-twister2.container.instance.disk: 2000.0
-twister2.container.instance.cpu: 4.0
+# Disk value to be allocated to each task instance
+twister2.taskscheduler.task.instance.disk: 500.0
 
-#Default Task Parallelism Value
-twister2.task.parallelism: 2
+# CPU value to be allocated to each task instancetwister2.task.parallelism
+twister2.taskscheduler.instance.cpu: 2.0
 
-#Default Task Type "streaming" or "batch"
-twister2.task.type: "streaming"
-```
+# Default Container Instance Values
+# Ram value to be allocated to each container
+twister2.taskscheduler.container.instance.ram: 4096.0
 
-\`\`
+# Disk value to be allocated to each container
+twister2.taskscheduler.container.instance.disk: 8000.0
+
+twister2.taskscheduler.container.instance.cpu: 16.0
+
+# Default Container Padding Values
+# Default padding value of the ram to be allocated to each container
+twister2.taskscheduler.ram.padding.container: 2.0
+
+# Default padding value of the disk to be allocated to each container
+twister2.taskscheduler.disk.padding.container: 12.0
+
+# CPU padding value to be allocated to each container
+twister2.taskscheduler.cpu.padding.container: 1.0
+
+# Percentage value to be allocated to each container
+twister2.taskscheduler.container.padding.percentage: 2
+
+# Static Default Network parameters
+# Bandwidth value to be allocated to each container instance for datalocality scheduling
+twister2.taskscheduler.container.instance.bandwidth: 100 #Mbps
+
+# Latency value to be allocated to each container instance for datalocality scheduling
+twister2.taskscheduler.container.instance.latency: 0.002 #Milliseconds
+
+# Bandwidth to be allocated to each datanode instance for datalocality scheduling
+twister2.taskscheduler.datanode.instance.bandwidth: 200 #Mbps
+
+# Latency value to be allocated to each datanode instance for datalocality scheduling
+twister2.taskscheduler.datanode.instance.latency: 0.01 #Milliseconds
+
+# Prallelism value to each task instance
+twister2.taskscheduler.task.parallelism: 2
+
+# Task type to each submitted job by default it is "streaming" job.
+twister2.taskscheduler.task.type: "streaming"
+
+# number of threads per worker
+twister2.exector.worker.threads: 1
+
+# name of the batch executor
+twister2.executor.batch.name: "edu.iu.dsc.tws.executor.threading.BatchSharingExecutor2"
+
+# number of tuples executed at a single pass
+twister2.exector.instance.queue.low.watermark: 10000
+```
 
 ## User-Defined Task Scheduler
 
@@ -165,8 +202,31 @@ as "user-defined" with the corresponding "user-defined" task scheduler class nam
 #User-defined Streaming Task Scheduler
 twister2.streaming.taskscheduler: "user-defined"
 
-#User-defined Streaming Task Scheduler Class
-twister2.streaming.taskscheduler.class: "edu.iu.dsc.tws.tsched.userdefined.UserDefinedTaskScheduler"
+# Task Scheduler for the userDefined Streaming Task Scheduler
+#twister2.taskscheduler.streaming.class: "edu.iu.dsc.tws.tsched.userdefined.UserDefinedTaskScheduler"
+
+# Task Scheduler for the userDefined Batch Task Scheduler
+#twister2.taskscheduler.batch.class: "edu.iu.dsc.tws.tsched.userdefined.UserDefinedTaskScheduler"
+```
+
+\`\`
+
+## Batch Task Scheduler 
+
+Batch Task Scheduler is able to handle and schedule both single task graph as well as multiple 
+dependent task graphs. The main constraint considered in the batch task scheduler is specify the same 
+parallelism value for the dependent tasks in the task graphs. It schedule the tasks in a round
+robin fashion but, while scheduling the child or the dependent tasks it considers the data locality 
+of the input data from the parent tasks and schedule the tasks in a round robin fashion to the workers. 
+
+\`\`yaml
+
+```text
+#Batch Task Scheduler
+twister2.taskscheduler.batch: "batchscheduler"
+
+#Task Scheduler class for the batch task scheduler
+twister2.taskscheduler.batch.class: "edu.iu.dsc.tws.tsched.batch.batchscheduler.BatchTaskScheduler"
 ```
 
 \`\`
@@ -176,25 +236,18 @@ twister2.streaming.taskscheduler.class: "edu.iu.dsc.tws.tsched.userdefined.UserD
 The other task schedulers and their respective class names are given below. The user have to specify
 the respective scheduler mode and their corresponding class names.
 
+
 \`\`yaml
 
 ```text
-#Streaming Task Scheduler Mode "roundrobin" or  "firstfit" or "datalocalityaware"
-twister2.streaming.taskscheduler: "roundrobin"
-
-#Streaming Task Scheduler Class
-twister2.streaming.taskscheduler.class: "edu.iu.dsc.tws.tsched.streaming.roundrobin.RoundRobinTaskScheduler"
-
-#twister2.streaming.taskscheduler.class: "edu.iu.dsc.tws.tsched.streaming.datalocalityaware.DataLocalityStreamingTaskScheduler"
-#twister2.streaming.taskscheduler.class: "edu.iu.dsc.tws.tsched.streaming.firstfit.FirstFitStreamingTaskScheduler"
+# Task Scheduler for the Data Locality Aware Streaming Task Scheduler
+#twister2.taskscheduler.streaming.class: "edu.iu.dsc.tws.tsched.streaming.datalocalityaware.DataLocalityStreamingTaskScheduler"
 
-#Batch Task Scheduler Mode "roundrobin" or  "datalocalityaware"
-#twister2.batch.taskscheduler: "roundrobin"
-twister2.batch.taskscheduler: "datalocalityaware"
+# Task Scheduler for the FirstFit Streaming Task Scheduler
+#twister2.taskscheduler.streaming.class: "edu.iu.dsc.tws.tsched.streaming.firstfit.FirstFitStreamingTaskScheduler"
 
-#Batch Task Scheduler Class
-twister2.batch.taskscheduler.class: "edu.iu.dsc.tws.tsched.batch.datalocalityaware.DataLocalityBatchTaskScheduler"
-#twister2.batch.taskscheduler.class: "edu.iu.dsc.tws.tsched.batch.roundrobin.RoundRobinBatchTaskScheduler"
+# Task Scheduler for the Data Locality Aware Batch Task Scheduler
+#twister2.taskscheduler.batch.class: "edu.iu.dsc.tws.tsched.batch.datalocalityaware.DataLocalityBatchTaskScheduler"
 ```
 
 \`\`