HealthCare BigData

Model

ColabNotebook

Evaluate Performence

Producer

Setup

$ ./bin/kafka-topics.sh --zookeeper localhost:2181 --create --topic test1 --partitions 1 --replication-factor 1

$ ./bin/kafka-topics.sh --zookeeper localhost:2181 --create --topic test2 --partitions 2 --replication-factor 1

$ ./bin/kafka-topics.sh --zookeeper localhost:2181 --create --topic test3 --partitions 3 --replication-factor 1

1. Single thread, async, no replication

./bin/kafka-run-class.sh org.apache.kafka.tools.ProducerPerformance \
--topic test1 \
--throughput -1 \
--num-records 5000000 \
--record-size 66560 \
--producer-props \
bootstrap.servers=localhost:9092 \
acks=1 \

2. Effect of message size

for i in 10 100 1000 10000 100000;
do
echo ""
echo $i
./bin/kafka-run-class.sh org.apache.kafka.clients.tools.ProducerPerformance \
--topic test \
--throughput -1 \
--num-records $((1000*1024*1024/$i)) \
--record-size $i \
--producer-props \
bootstrap.servers=localhost:9092 \
acks=1 \
buffer.memory=67108864 batch.size=128000
done;

Consumer

1. Consumer throughput

./bin/kafka-consumer-perf-test.sh ----bootstrap-server localhost:9092 \
--messages 50000000 \
--topic test \
--threads 1

2. Producer and consumer

./bin/kafka-run-class.sh org.apache.kafka.clients.tools.ProducerPerformance \
--topic test \
--throughput -1 \
--num-records 50000000 \
--record-size 200 \
--producer-props \
bootstrap.servers=localhost:9092 \
acks=1 \
buffer.memory=67108864 batch.size=8196

./bin/kafka-consumer-perf-test.sh --zookeeper localhost:2181 --messages 50000000 --topic test --threads 1

Ref

1. Deeplearning4jPractical

2. ModelTraining

3. Real-Time Remote Health Monitoring System Driven by 5G MEC-IoT

4. Benchmarking Apache Kafka