High Availability

The SMS Aggregation service is made up of several different components. Each of these components is configured for High Availability (HA) using either active/active pairs or active/hot-standby pairs. All client-facing components are made available in an active/active configuration. Any components not configured as active/active have a designated failover policy that is applied in case a component fails. In addition, each component is provisioned with enough capacity to handle the complete system load at any time.

To avoid any interruptions to service during maintenance or outages, follow the recommendations below:

HTTP APIs

The HTTP APIs are available for use at any site where the SMS Aggregation service is located. Details of the URLs are discussed in HTTP API URLs.

The architecture of the SMS Aggregation service is the same for each site and is shown in the following figure.

HTTP API Architecture (click to enlarge)

The flow of a client's request is as follows:

1. The client's request is received at the datacenter firewall.
2. The request is forwarded to load-balancing cluster.
3. The load balancer selects the HTTP node using round-robin balancing.

SMSC APIs

The SMSC component provides HA by supporting multiple client connections to each node within the SMSC cluster. The complete list of available URIs are listed here. To take full advantage of the HA SMSC service, a client should comply with the following recommendations:

1. Create one SMPP connection per SMSC node (as shown in the figure below).
2. If the SMSC node becomes unavailable, attempt to rebind in 30 second intervals.
3. Clients should expect MOs and DRs to be sent to ANY active SMPP connection.

SMSC DR Example Configuration (click to enlarge)