一、引言
隨著物聯網技術的迅猛發展,大量的裝置和感測器產生了海量的資料。本文利用了 MQTT、Kafka 和 MongoDB 各自的優點,滿足實時資料處理和大規模資料儲存的需求。
如圖:
二、總結
優點:
1. 可靠和解耦:
Kafka的複製機制和持久化儲存確保了資料在傳輸過程中的可靠性,即使某個節點發生故障,也不會導致資料丟失,將資料生產者和消費者解耦,各模組可以獨立擴充套件和最佳化,減少了相互影響。
2. 高可用和靈活性:
MongoDB的複製集和分片機制提供了資料的高可用性和容錯能力,保證了資料儲存的可靠性和靈活性。
缺點:
1. 複雜度高:
包含多個元件(MQTT、Kafka、MongoDB)配置、部署和維護、各元件之間的協調和整合也增加了實現的複雜性。
2. 延遲:
資料從裝置上傳到最終儲存在MongoDB之間經過多個處理環節,每個環節都可能增加一些延遲。
3. 一致性:
資料在Kafka和MongoDB之間傳遞時可能需要額外的處理機制來確保一致性。
三、實現
準備工作
使用docker-compose.yml建立Kafka服務和MongoDB,簡易程式碼如下:
version: '3.8'
networks:
app-tier:
driver: bridge
services:
kafka:
image: 'bitnami/kafka:latest'
networks:
- app-tier
ports:
- "9092:9092"
environment:
- KAFKA_CFG_NODE_ID=0
- KAFKA_CFG_PROCESS_ROLES=controller,broker
- KAFKA_CFG_LISTENERS=PLAINTEXT://0.0.0.0:9092,CONTROLLER://0.0.0.0:9093
- KAFKA_CFG_ADVERTISED_LISTENERS=PLAINTEXT://127.0.0.1:9092
- KAFKA_CFG_LISTENER_SECURITY_PROTOCOL_MAP=CONTROLLER:PLAINTEXT,PLAINTEXT:PLAINTEXT
- KAFKA_CFG_CONTROLLER_QUORUM_VOTERS=0@kafka:9093
- KAFKA_CFG_CONTROLLER_LISTENER_NAMES=CONTROLLER
volumes:
- kafka-data:/bitnami/kafka
mongodb:
image: 'mongo:latest'
networks:
- app-tier
container_name: mongodb
ports:
- "27017:27017"
volumes:
- mongo-data:/data/db
volumes:
kafka-data:
driver: local
mongo-data:
driver: local
實現步驟
1. 裝置資料上傳:
服務端程式碼
var mqttFactory = new MqttFactory();
var mqttServerOptions = new MqttServerOptionsBuilder()
.WithDefaultEndpointPort(1883)//監聽的埠
.WithDefaultEndpoint()
.WithoutEncryptedEndpoint()// 不啟用tls
.WithDefaultCommunicationTimeout(TimeSpan.FromSeconds(10 * 1000))//10秒超時
.WithPersistentSessions(true)//啟用session
.WithConnectionBacklog(1000)//積累的最大連線請求數
.Build();
using (var mqttServer = mqttFactory.CreateMqttServer(mqttServerOptions))
{
AddMqttEvents(mqttServer);
await mqttServer.StartAsync();
Console.WriteLine("Press Enter Ctrl+C to exit.");
Console.ReadLine();
Console.CancelKeyPress += async (sender, e) =>
{
e.Cancel = true; // 防止程序直接終止
await mqttServer.StopAsync();
Environment.Exit(0);
};
}
private static void AddMqttEvents(MqttServer mqttServer)
{
MqttServerEvents mqttEvents = new MqttServerEvents();
mqttServer.ClientConnectedAsync += mqttEvents.Server_ClientConnectedAsync;
mqttServer.StartedAsync += mqttEvents.Server_StartedAsync;
mqttServer.StoppedAsync += mqttEvents.Server_StoppedAsync;
mqttServer.ClientSubscribedTopicAsync += mqttEvents.Server_ClientSubscribedTopicAsync;
mqttServer.ClientUnsubscribedTopicAsync += mqttEvents.Server_ClientUnsubscribedTopicAsync;
mqttServer.ValidatingConnectionAsync += mqttEvents.Server_ValidatingConnectionAsync;
mqttServer.ClientDisconnectedAsync += mqttEvents.Server_ClientDisconnectedAsync;
mqttServer.InterceptingInboundPacketAsync += mqttEvents.Server_InterceptingInboundPacketAsync;
mqttServer.InterceptingOutboundPacketAsync += mqttEvents.Server_InterceptingOutboundPacketAsync;
mqttServer.InterceptingPublishAsync += mqttEvents.Server_InterceptingPublishAsync;
mqttServer.ApplicationMessageNotConsumedAsync += mqttEvents.Server_ApplicationMessageNotConsumedAsync;
mqttServer.ClientAcknowledgedPublishPacketAsync += mqttEvents.Server_ClientAcknowledgedPublishPacketAsync;
}
客戶端程式碼
var mqttFactory = new MqttFactory();
var mqttClient = mqttFactory.CreateMqttClient();
var mqttOptions = new MqttClientOptionsBuilder()
.WithClientId("MqttServiceClient")
.WithTcpServer("127.0.0.1", 1883)
.Build();
mqttClient.ConnectedAsync+=(e =>
{
Console.WriteLine("MQTT連線成功");
return Task.CompletedTask;
});
mqttClient.DisconnectedAsync+=(e =>
{
Console.WriteLine("MQTT連線斷開");
return Task.CompletedTask;
});
await mqttClient.ConnectAsync(mqttOptions, CancellationToken.None);
//傳送訊息
MqttApplicationMessage applicationMessage = new MqttApplicationMessage
{
Topic = "mqtttest",
PayloadSegment = new ArraySegment<byte>(System.Text.Encoding.UTF8.GetBytes(input))
};
var res = await mqttClient.PublishAsync(applicationMessage);
2. Kafka訊息處理:
生產者程式碼
var config = new ProducerConfig
{
BootstrapServers = "localhost:9092"
};
using var producer = new ProducerBuilder<string, string>(config).Build();
try
{
var message = new Message<string, string>
{
Key = e.ClientId,
Value = JsonConvert.SerializeObject(e.Packet)
};
var deliveryResult = await producer.ProduceAsync("mqttMsg-topic", message);
Console.WriteLine($"Delivered '{deliveryResult.Value}' to '{deliveryResult.TopicPartitionOffset}'");
}
catch (ProduceException<string, string> ke)
{
Console.WriteLine($"Delivery failed: {ke.Error.Reason}");
}
消費者程式碼
var config = new ConsumerConfig
{
GroupId = "my-consumer-group",
BootstrapServers = "127.0.0.1:9092",
AutoOffsetReset = AutoOffsetReset.Earliest
};
using var consumer = new ConsumerBuilder<string, string>(config).Build();
consumer.Subscribe("mqttMsg-topic");
//消費訊息並儲存到mongodb
var client = new MongoClient("mongodb://127.0.0.1:27017");
var collection = client.GetDatabase("mqtttest").GetCollection<BsonDocument>($"history_{DateTime.UtcNow.Year}_{DateTime.UtcNow.Month}");
while (true)
{
try
{
var consumeResult = consumer.Consume(cancellationToken.Token);
Console.WriteLine($"收到Kafka訊息 '{consumeResult.Message.Value}' at: '{consumeResult.TopicPartitionOffset}'.");
var document = new BsonDocument
{
{ "clientId", consumeResult.Message.Key },
{ "JsonData", MongoDB.Bson.Serialization.BsonSerializer.Deserialize<BsonDocument>(consumeResult.Message.Value) },//不同裝置上報資料格式不一定一樣
{ "created", DateTime.UtcNow }
};
await collection.InsertOneAsync(document);
}
catch (ConsumeException e)
{
Console.WriteLine($"處理Kafka訊息異常: {e.Error.Reason}");
}
}
原始碼地址:https://github.com/jclown/MqttPersistence