系统架构全景图
图表
一、系统平台优化(CentOS Stream 8)
1. 系统基础配置
bash
# 1. 系统更新与加固
sudo dnf update -y
sudo dnf install epel-release -y
sudo dnf install fail2ban firewalld -y
# 2. 创建专用运维账户
sudo useradd -m -s /bin/bash iotadmin
sudo passwd iotadmin
sudo usermod -aG wheel iotadmin
# 3. SSH安全加固
sudo sed -i 's/^PermitRootLogin yes/PermitRootLogin no/' /etc/ssh/sshd_config
sudo sed -i 's/^PasswordAuthentication yes/PasswordAuthentication no/' /etc/ssh/sshd_config
sudo systemctl restart sshd
# 4. 防火墙配置
sudo systemctl enable --now firewalld
sudo firewall-cmd --permanent --add-port=1883/tcp # MQTT
sudo firewall-cmd --permanent --add-port=8883/tcp # MQTT/SSL
sudo firewall-cmd --permanent --add-port=9092/tcp # Kafka
sudo firewall-cmd --permanent --add-port=3000/tcp # Grafana
sudo firewall-cmd --reload
2. 内核参数优化(/etc/sysctl.conf)
conf
网络性能优化
net.core.somaxconn = 65535
net.core.netdev_max_backlog = 65536
net.ipv4.tcp_max_syn_backlog = 65536
文件句柄限制
fs.file-max = 2097152
fs.nr_open = 2097152
MQTT连接优化
net.ipv4.tcp_keepalive_time = 600
net.ipv4.tcp_keepalive_probes = 3
net.ipv4.tcp_keepalive_intvl = 15
二、MQTT Broker集群部署(EMQX企业版)
1. 集群化部署
bash
# 安装EMQX企业版
curl -s https://assets.emqx.com/scripts/install-emqx-rpm.sh | sudo bash
sudo dnf install emqx-enterprise -y
# 配置集群(3节点示例)
# 节点1(10.0.0.1):
echo "cluster.name = iot_platform" >> /etc/emqx/emqx.conf
echo "node.name = [email protected]" >> /etc/emqx/emqx.conf
# 节点2(10.0.0.2):
emqx_ctl cluster join [email protected]
2. 安全加固配置
bash
# 1. 启用TLS加密
sudo mkdir /etc/emqx/certs
sudo certbot certonly --standalone -d mqtt.example.com
sudo cp /etc/letsencrypt/live/mqtt.example.com/* /etc/emqx/certs/
# 2. 配置EMQX(/etc/emqx/emqx.conf)
listeners.ssl.default {
bind = "0.0.0.0:8883"
max_connections = 100000
ssl_options {
keyfile = "/etc/emqx/certs/privkey.pem"
certfile = "/etc/emqx/certs/fullchain.pem"
}
}
# 3. 设备级认证
emqx_ctl users add device_001 6rounds=10000somesalthashed_password
3. 主题权限控制
conf
/etc/emqx/acl.conf
{allow, {user, "device_001"}, publish, ["sensors/001/#"]}
{allow, {user, "backend"}, subscribe, ["sensors/#"]}
{deny, all}
三、数据处理与存储架构
1. 消息队列缓冲(Kafka)
bash
# 安装Kafka
wget https://downloads.apache.org/kafka/3.4.0/kafka_2.13-3.4.0.tgz
tar -xzf kafka_2.13-3.4.0.tgz
# 配置集群(3节点)
# server.properties
broker.id=1
listeners=PLAINTEXT://:9092
advertised.listeners=PLAINTEXT://node1:9092
zookeeper.connect=node1:2181,node2:2181,node3:2181
2. 时序数据库(TimescaleDB)
bash
# 安装PostgreSQL 15 + TimescaleDB
sudo dnf install https://download.postgresql.org/pub/repos/yum/reporpms/EL-8-x86_64/pgdg-redhat-repo-latest.noarch.rpm
sudo dnf module disable postgresql
sudo dnf install postgresql15-server postgresql15-contrib timescaledb-2-postgresql-15
# 初始化数据库
sudo /usr/pgsql-15/bin/postgresql-15-setup initdb
sudo systemctl enable --now postgresql-15
# 创建超级表
CREATE TABLE sensor_data (
time TIMESTAMPTZ NOT NULL,
device_id TEXT NOT NULL,
value DOUBLE PRECISION NOT NULL
);
SELECT create_hypertable('sensor_data', 'time');
3. 数据清洗服务(Python示例)
python
from kafka import KafkaConsumer
import psycopg2
# Kafka消费者
consumer = KafkaConsumer(
'raw_sensor_data',
bootstrap_servers=['kafka1:9092', 'kafka2:9092'],
security_protocol='SSL',
ssl_cafile='ca.pem'
)
# TimescaleDB连接
conn = psycopg2.connect("dbname=tsdb user=tsdbadmin")
cursor = conn.cursor()
for message in consumer:
data = json.loads(message.value)
# 数据验证
if not validate_sensor_data(data):
continue
# 数据清洗
cleaned = clean_data(data)
# 写入数据库
cursor.execute(
"INSERT INTO sensor_data (time, device_id, value) VALUES (%s, %s, %s)",
(cleaned['timestamp'], cleaned['device_id'], cleaned['value'])
)
conn.commit()
# 更新缓存
redis.set(f"latest:{cleaned['device_id']}", json.dumps(cleaned))
四、安全加固体系
1. 传输层安全
协议 | 端口 | 加密方式 | 证书管理 |
---|---|---|---|
MQTT | 8883 | TLS 1.3 | Let's Encrypt自动更新 |
HTTPS | 443 | TLS 1.3 | 企业级证书 |
Database | 5432 | TLS双向认证 | 自签名CA |
2. 数据加密策略
python
# 设备端数据加密示例
from cryptography.hazmat.primitives.ciphers import Cipher, algorithms, modes
from cryptography.hazmat.backends import default_backend
def encrypt_data(data, key):
iv = os.urandom(12)
cipher = Cipher(
algorithms.AES(key),
modes.GCM(iv),
backend=default_backend()
)
encryptor = cipher.encryptor()
ciphertext = encryptor.update(data) + encryptor.finalize()
return iv + encryptor.tag + ciphertext
3. 访问控制矩阵
角色 | MQTT权限 | DB访问 | API权限 |
---|---|---|---|
设备 | 发布特定主题 | 无 | 无 |
数据服务 | 订阅所有主题 | 只写 | 内部网络访问 |
前端应用 | 无 | 只读 | JWT认证+RBAC |
管理员 | 管理主题 | 读写 | 管理员权限 |
五、运维监控体系
1. 监控组件部署
bash
# Prometheus安装
sudo dnf install prometheus
# Node Exporter
sudo dnf install node_exporter
# Grafana
sudo dnf install grafana
2. 关键监控指标
yaml
# prometheus.yml 片段
scrape_configs:
- job_name: 'emqx'
static_configs:
-
targets: ['emqx1:18083', 'emqx2:18083']
-
job_name: 'postgres'
static_configs:
-
targets: ['db1:9187']
-
job_name: 'kafka'
static_configs:
- targets: ['kafka1:7071']
3. 告警规则示例
yaml
groups:
- name: MQTT服务
rules:
- alert: EMQX节点离线
expr: up{job="emqx"} == 0
for: 5m
labels:
severity: critical
annotations:
summary: "MQTT节点 {{ $labels.instance }} 离线"
- alert: 消息积压
expr: kafka_consumergroup_lag > 10000
for: 10m
labels:
severity: warning
六、设备接入与扩展方案
1. 设备接入流程
图表
2. 设备管理API设计
python
# 设备注册API
@app.route('/api/v1/devices', methods=['POST'])
@jwt_required()
def register_device():
data = request.get_json()
device_id = generate_device_id()
# 创建数据库记录
db.execute("""
INSERT INTO devices (id, name, type, owner)
VALUES (%s, %s, %s, %s)
""", (device_id, data['name'], data['type'], get_jwt_identity()))
# 生成设备凭证
credential = generate_device_credential(device_id)
return jsonify({
'device_id': device_id,
'username': credential['username'],
'password': credential['password'],
'certificate': credential['cert_pem']
}), 201
3. 多协议支持方案
协议 | 转换方式 | 适用场景 |
---|---|---|
HTTP | EMQX Webhook | 传统设备改造 |
CoAP | CoAP-MQTT代理网关 | 低功耗设备 |
Modbus | 边缘计算转换 | 工业设备 |
LoRaWAN | 网络服务器集成 | 长距离物联网 |
七、日常运维手册
1. 备份策略
bash
# 数据库每日备份
pg_dump -U postgres -Fc tsdb > /backup/tsdb-$(date +%F).dump
# 配置文件备份
rsync -av /etc/emqx /backup/configs/emqx
rsync -av /etc/kafka /backup/configs/kafka
# 证书备份
tar -czf /backup/certs-$(date +%F).tar.gz /etc/letsencrypt/{live,archive}
2. 灾难恢复流程
- 恢复最新数据库备份
- 重建EMQX集群
- 恢复Kafka偏移量
- 验证数据完整性
- 逐步恢复设备连接
3. 性能调优命令
bash
# 查看MQTT连接数
emqx_ctl clients list
# 检查Kafka积压
kafka-consumer-groups.sh --describe --group data_consumers
# 时序数据库维护
timescaledb-tune --quiet --yes
八、扩展架构设计
1. 边缘计算集成
图表
2. 数据管道扩展
python
# 添加AI处理管道
from kafka import KafkaProducer
ai_producer = KafkaProducer(bootstrap_servers='kafka:9092')
def process_for_ai(data):
# 特征提取
features = extract_features(data)
# 发送到AI服务队列
ai_producer.send('ai_processing', json.dumps(features).encode())
# 在清洗服务中调用
process_for_ai(cleaned_data)
3. 多区域部署
bash
# 跨区域EMQX集群
emqx_ctl cluster join [email protected]
# 数据库级联复制
# 主库(欧洲)
CREATE PUBLICATION euro_publication FOR TABLE sensor_data;
# 从库(亚洲)
CREATE SUBSCRIPTION asia_subscription
CONNECTION 'host=euro-db port=5432 dbname=tsdb'
PUBLICATION euro_publication;
九、前端展示系统
1. 实时数据大屏
javascript
// 使用MQTT.js直接订阅
const client = mqtt.connect('wss://mqtt.example.com:8084/mqtt', {
username: 'web_user',
password: 'secure_token'
})
client.subscribe('sensors/+/status')
client.on('message', (topic, payload) => {
const data = JSON.parse(payload)
updateDashboard(data)
})
2. 设备管理界面功能
- 设备状态监控(在线/离线)
- 实时数据曲线(Chart.js)
- 历史数据查询(时间范围选择)
- 告警管理(阈值设置)
- 固件OTA升级
3. 移动APP集成
kotlin
// Android数据获取示例
suspend fun fetchSensorData(deviceId: String): List<SensorData> {
return withContext(Dispatchers.IO) {
val response = apiService.getDeviceData(
deviceId = deviceId,
from = Instant.now().minus(1, ChronoUnit.DAYS).toString(),
to = Instant.now().toString()
)
response.data ?: emptyList()
}
}
十、持续演进路线
- 阶段1(基础平台)
- EMQX集群部署
- 核心数据处理流水线
- 基础监控
- 阶段2(安全加固)
- 设备证书管理
- 数据端到端加密
- 审计日志
- 阶段3(智能扩展)
- 边缘计算节点
- AI异常检测
- 预测性维护
- 阶段4(全球化部署)
- 多区域集群
- 数据主权合规
- 跨云架构
本方案基于CentOS Stream 8构建了企业级物联网平台,通过多层次安全加固、全链路监控、弹性扩展架构,支持从数百到数百万设备的平滑扩展,日均处理能力可达亿级数据点,满足工业4.0场景需求。