This topic describes how to add a UDP listener to an SLB instance. UDP is applicable to services that prioritize real-time content delivery over reliability, such as video chats and real-time quotes. You can add a UDP listener to forward UDP requests.

Prerequisites

An SLB instance is created. For more information, see Create an SLB instance.

Background information

Before you configure a UDP listener, take note of the following items:
  • Ports 250, 4789, and 4790 of a UDP listener are reserved and therefore are unavailable for your configuration.
  • Fragmented packets are not supported.
  • The UDP listeners of an SLB instance in the classic network do not allow you to view source IP addresses.
  • The following operations take five minutes to take effect if they are performed for a UDP listener:
    • Remove backend servers
    • Set the weight of a backend server to 0 after it is detected unhealthy
  • IPv6 addresses have longer IP headers than IPv4 addresses. When you create a UDP listener for an IPv6 SLB instance, make sure that the maximum transmission unit (MTU) size of the network interface controller (NIC) on each backend server (ECS instance in most cases) that communicates with the SLB instance is not greater than 1,200 bytes. Otherwise, oversized packets may be discarded. The configuration files of some applications need to be modified based on the MTU size.

    If you use a TCP, HTTP, or HTTPS listener, you do not need to perform additional configuration steps because TCP supports automatic maximum segment size (MSS) adjustment.

Step 1: Configure the UDP listener

To configure the UDP listener, perform the following operations:

  1. In the Protocol and Listener step, configure the following parameters.
    Parameter Description
    Select Listener Protocol Select the protocol of the listener.

    In this example, select UDP.

    Listening Port Set the listening port that is used to receive requests and forward them to backend servers.

    Valid values: 1 to 65535.
    Advanced
    Scheduling Algorithm SLB supports four scheduling algorithms: RR, WRR, and CH.
    • Weighted Round-Robin (WRR): Backend servers with a higher weight receive more requests than those with a lower weight.
    • Round-Robin (RR): Requests are sequentially distributed to backend servers.
    • Consistent Hash (CH):
      • Source IP: the source IP hash. Requests from the same source IP address are scheduled to the same backend server.
      • Tuple: a quadruple hash that consists of the source IP address, destination IP address, source port number, and destination port number. Requests with the same quadruple are scheduled to the same backend server.
      • QUIC ID: the hash that is based on Quick UDP Internet Connections (QUIC) IDs. Requests that contain the same QUIC ID are scheduled to the same backend server.
        Notice The QUIC protocol is implemented based on draft-ietf-quic-transport-10 and is rapidly evolving. Therefore, compatibility is not guaranteed for all QUIC versions. We recommend that you perform tests before you apply the protocol to the production environment.
    Enable Access Control Specify whether to configure the listener to restrict access.
    Access Control Method

    Select an access control method after you enable access control.

    • Whitelist: Only the requests from the IP addresses or CIDR blocks in the specified ACL are forwarded. You can use the whitelist feature when you want to allow access from specified IP addresses.

      Risks may arise if you specify an ACL as a whitelist. After a whitelist is enabled, only IP addresses contained in the whitelist can access the SLB listener. If a whitelist is enabled without any IP addresses specified, the SLB listener forwards all requests.

    • Blacklist: Requests from the IP addresses or CIDR blocks in the specified ACL are not forwarded. You can use the blacklist feature when you want to deny access from specified IP addresses.

      If a blacklist is enabled without any IP addresses specified, the SLB listener forwards all requests.

    Access Control List Select an ACL that is used as the whitelist or blacklist of the listener.
    Note IPv6 instances can be associated only with IPv6 ACLs, and IPv4 instances can be associated only with IPv4 ACLs. For more information, see Create an access control list.
    Enable Peak Bandwidth Limit

    You can switch on this option and then set a bandwidth limit for the listener.

    If an SLB instance incurs fees based on the bandwidth, you can set different peak bandwidth values for different listeners to limit the amount of traffic that flows in each listener. The sum of the peak bandwidth values of all listeners added to an SLB instance cannot exceed the bandwidth of this SLB instance.

    By default, this feature is disabled and all listeners share the bandwidth of the SLB instance.

    Note If an SLB instance incurs fees based on the amount of transmitted data, no peak bandwidth limit is applied by default.
    Obtain Client Source IP Address Backend servers of a UDP listener can directly obtain the actual IP addresses of clients.
    Note UDP listeners of an SLB instance in the classic network do not allow you to view source IP addresses.
    Automatically Enable Listener After Creation Specify whether to start the listener after the listener is configured. By default, the listener is started after configuration.
  2. Click Next.

Step 2: Add backend servers

After you configure the listener, you must add backend servers to process client requests. You can add backend servers to the default server group, or create VServer groups or primary/secondary server groups and then add servers to them. For more information, see Backend server overview.

Backend servers are added to the default server group in this example.

  1. Select Default Server Group and click Add More.
    Add backend servers to the default server group
  2. Select ECS instances (backend servers) that you want to add, and then click Next.
    Configure weights
  3. Configure weights for the added backend servers.
    A backend server with a higher weight receives more requests.
    Note If the weight of a backend server is set to 0, the backend server does not receive new requests.
  4. Click Add. On the Default Server Group tab, configure ports for the backend servers.
    Set a port for each backend server to receive requests. Valid values: 1 to 65535. You can specify the same port for multiple backend servers of an SLB instance.
  5. Click Next.

Step 3: Configure health checks

SLB checks the availability of Elastic Compute Service (ECS) instances that serve as backend servers by performing health checks. The health check feature improves the overall availability of your frontend business and mitigates the impacts of exceptions that occur on backend ECS instances. Click Modify to modify the configurations of health checks. For more information, see Health check overview.

Step 4: Confirm the configurations

Perform the following steps to confirm the configurations.

  1. In the Confirm step, check the configurations. You can click Modify to modify the configurations.
  2. After you confirm the configurations, click Submit.
  3. In the Configuration Successful message, click OK.

    You can check the created listener on the Listener tab.