rreddy-22 commented on code in PR #15717: URL: https://github.com/apache/kafka/pull/15717#discussion_r1578083329
########## jmh-benchmarks/src/main/java/org/apache/kafka/jmh/assignor/ServerSideAssignorBenchmark.java: ########## @@ -0,0 +1,281 @@ +/* + * 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.kafka.jmh.assignor; + +import org.apache.kafka.common.Uuid; +import org.apache.kafka.coordinator.group.assignor.AssignmentMemberSpec; +import org.apache.kafka.coordinator.group.assignor.AssignmentSpec; +import org.apache.kafka.coordinator.group.assignor.GroupAssignment; +import org.apache.kafka.coordinator.group.assignor.MemberAssignment; +import org.apache.kafka.coordinator.group.assignor.PartitionAssignor; +import org.apache.kafka.coordinator.group.assignor.RangeAssignor; +import org.apache.kafka.coordinator.group.assignor.SubscribedTopicDescriber; +import org.apache.kafka.coordinator.group.assignor.UniformAssignor; +import org.apache.kafka.coordinator.group.consumer.SubscribedTopicMetadata; +import org.apache.kafka.coordinator.group.consumer.TopicMetadata; +import org.openjdk.jmh.annotations.Benchmark; +import org.openjdk.jmh.annotations.BenchmarkMode; +import org.openjdk.jmh.annotations.Fork; +import org.openjdk.jmh.annotations.Level; +import org.openjdk.jmh.annotations.Measurement; +import org.openjdk.jmh.annotations.Mode; +import org.openjdk.jmh.annotations.OutputTimeUnit; +import org.openjdk.jmh.annotations.Param; +import org.openjdk.jmh.annotations.Scope; +import org.openjdk.jmh.annotations.Setup; +import org.openjdk.jmh.annotations.State; +import org.openjdk.jmh.annotations.Threads; +import org.openjdk.jmh.annotations.Warmup; + +import java.util.ArrayList; +import java.util.Arrays; +import java.util.Collection; +import java.util.Collections; +import java.util.HashMap; +import java.util.HashSet; +import java.util.List; +import java.util.Map; +import java.util.Optional; +import java.util.Set; +import java.util.concurrent.TimeUnit; + +@State(Scope.Benchmark) +@Fork(value = 1) +@Warmup(iterations = 5) +@Measurement(iterations = 5) +@BenchmarkMode(Mode.AverageTime) +@OutputTimeUnit(TimeUnit.MILLISECONDS) +public class ServerSideAssignorBenchmark { + + public enum AssignorType { + RANGE(new RangeAssignor()), + UNIFORM(new UniformAssignor()); + + private final PartitionAssignor assignor; + + AssignorType(PartitionAssignor assignor) { + this.assignor = assignor; + } + + public PartitionAssignor assignor() { + return assignor; + } + } + + /** + * The subscription pattern followed by the members of the group. + * + * A subscription model is considered homogenous if all the members of the group + * are subscribed to the same set of topics, it is heterogeneous otherwise. + */ + public enum SubscriptionModel { + HOMOGENEOUS, HETEROGENEOUS + } + + /** + * The assignment type is decided based on whether all the members are assigned partitions + * for the first time (full), or incrementally when a rebalance is triggered. + */ + public enum AssignmentType { + FULL, INCREMENTAL + } + + @Param({"100", "500", "1000", "5000", "10000"}) + private int memberCount; + + @Param({"5", "10", "50"}) + private int partitionsToMemberRatio; + + @Param({"10", "100", "1000"}) + private int topicCount; + + @Param({"true", "false"}) + private boolean isRackAware; + + @Param({"HOMOGENEOUS", "HETEROGENEOUS"}) + private SubscriptionModel subscriptionModel; + + @Param({"RANGE", "UNIFORM"}) + private AssignorType assignorType; + + @Param({"FULL", "INCREMENTAL"}) + private AssignmentType assignmentType; + + private PartitionAssignor partitionAssignor; + + private static final int NUMBER_OF_RACKS = 3; + + private static final int MAX_BUCKET_COUNT = 5; + + private AssignmentSpec assignmentSpec; + + private SubscribedTopicDescriber subscribedTopicDescriber; + + private final List<Uuid> allTopicIds = new ArrayList<>(topicCount); + + @Setup(Level.Trial) + public void setup() { + Map<Uuid, TopicMetadata> topicMetadata = createTopicMetadata(); + subscribedTopicDescriber = new SubscribedTopicMetadata(topicMetadata); + + createAssignmentSpec(); + + partitionAssignor = assignorType.assignor(); + + if (assignmentType == AssignmentType.INCREMENTAL) { + simulateIncrementalRebalance(); + } + } + + private Map<Uuid, TopicMetadata> createTopicMetadata() { + Map<Uuid, TopicMetadata> topicMetadata = new HashMap<>(); + int partitionsPerTopicCount = (memberCount * partitionsToMemberRatio) / topicCount; + + Map<Integer, Set<String>> partitionRacks = isRackAware ? + mkMapOfPartitionRacks(partitionsPerTopicCount) : + Collections.emptyMap(); + + for (int i = 0; i < topicCount; i++) { + Uuid topicUuid = Uuid.randomUuid(); + String topicName = "topic" + i; + allTopicIds.add(topicUuid); + topicMetadata.put(topicUuid, new TopicMetadata( + topicUuid, + topicName, + partitionsPerTopicCount, + partitionRacks + )); + } + + return topicMetadata; + } + + private void createAssignmentSpec() { + Map<String, AssignmentMemberSpec> members = new HashMap<>(); + + // In the rebalance case, we will add the last member as a trigger. + // This is done to keep the total members count consistent with the input. + int numberOfMembers = assignmentType.equals(AssignmentType.INCREMENTAL) ? memberCount - 1 : memberCount; + + if (subscriptionModel.equals(SubscriptionModel.HOMOGENEOUS)) { + for (int i = 0; i < numberOfMembers; i++) { + addMemberSpec(members, i, new HashSet<>(allTopicIds)); + } + } else { + // Adjust bucket count based on member count when member count < max bucket count. + int bucketCount = Math.min(MAX_BUCKET_COUNT, numberOfMembers); + + // Check minimum topics requirement + if (topicCount < bucketCount) { + throw new IllegalArgumentException("At least " + bucketCount + " topics are recommended for effective bucketing."); + } + + int bucketSizeTopics = (int) Math.ceil((double) topicCount / bucketCount); + int bucketSizeMembers = (int) Math.ceil((double) numberOfMembers / bucketCount); + + // Define buckets for each member and assign topics from the same bucket + for (int bucket = 0; bucket < bucketCount; bucket++) { + int memberStartIndex = bucket * bucketSizeMembers; + int memberEndIndex = Math.min((bucket + 1) * bucketSizeMembers, numberOfMembers); + + int topicStartIndex = bucket * bucketSizeTopics; + int topicEndIndex = Math.min((bucket + 1) * bucketSizeTopics, topicCount); + + Set<Uuid> bucketTopics = new HashSet<>(allTopicIds.subList(topicStartIndex, topicEndIndex)); + + // Assign topics to each member in the current bucket + for (int i = memberStartIndex; i < memberEndIndex; i++) { + addMemberSpec(members, i, new HashSet<>(bucketTopics)); + } + } + } + + this.assignmentSpec = new AssignmentSpec(members); + } + + private Optional<String> rackId(int memberIndex) { + return isRackAware ? Optional.of("rack" + memberIndex % NUMBER_OF_RACKS) : Optional.empty(); + } + + private void addMemberSpec( + Map<String, AssignmentMemberSpec> members, + int memberIndex, + Set<Uuid> subscribedTopicIds + ) { + String memberId = "member" + memberIndex; + Optional<String> rackId = rackId(memberIndex); + + members.put(memberId, new AssignmentMemberSpec( + Optional.empty(), + rackId, + subscribedTopicIds, + Collections.emptyMap() + )); + } + + private static Map<Integer, Set<String>> mkMapOfPartitionRacks(int numPartitions) { + Map<Integer, Set<String>> partitionRacks = new HashMap<>(numPartitions); + for (int i = 0; i < numPartitions; i++) { + partitionRacks.put(i, new HashSet<>(Arrays.asList( + "rack" + i % NUMBER_OF_RACKS, + "rack" + (i + 1) % NUMBER_OF_RACKS, + "rack" + (i + 2) % NUMBER_OF_RACKS + ))); + } + return partitionRacks; + } + + private void simulateIncrementalRebalance() { + GroupAssignment initialAssignment = partitionAssignor.assign(assignmentSpec, subscribedTopicDescriber); + Map<String, MemberAssignment> members = initialAssignment.members(); + + Map<String, AssignmentMemberSpec> updatedMembers = new HashMap<>(); + members.forEach((memberId, memberAssignment) -> { + AssignmentMemberSpec memberSpec = assignmentSpec.members().get(memberId); + updatedMembers.put(memberId, new AssignmentMemberSpec( + memberSpec.instanceId(), + memberSpec.rackId(), + memberSpec.subscribedTopicIds(), + memberAssignment.targetPartitions() + )); + }); + + Collection<Uuid> subscribedTopicIdsForNewMember; + if (subscriptionModel == SubscriptionModel.HETEROGENEOUS) { + subscribedTopicIdsForNewMember = updatedMembers.values().iterator().next().subscribedTopicIds(); Review Comment: We can just assign to the last member either way for the regular cases that we put, I guess it doesn't matter if it's the first or last one when all the buckets are full, at least when the last one is not we are filling in the bucket you're right -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: jira-unsubscr...@kafka.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
