| How it’s built | a single process, queues on disk, a text configuration | the MTA, a RabbitMQ broker and a cluster database |
| Where settings live | the configuration file and the files it includes | four layers: the process file, the control database, REST, message properties |
| Changing settings | edit the file and reload the configuration | in the database and via REST — on the fly; the process file — with a restart |
| Rules per mailbox provider | domain sections and macros: connections, rate, TLS, retries | a rate for each “IP address → domain” pair: messages, connections, bytes per minute |
| Reaction to bounces | the queue switches to a gentle (backoff) mode based on response patterns | patterns slow the flow down or change the IP address, tags and what happens to the message |
| How messages get into the MTA | SMTP and a pickup directory | SMTP, REST, straight into a RabbitMQ queue, a pickup directory |
| Delivery results | accounting files for every message | result queues: success and failure |
| Scale | high sending volume on a single server | horizontal: a cluster on shared queues and a shared database |
| With email platforms | modules in Protosender, Interspire and Mumara | through the platform’s SMTP servers |
| In MTA-Hub | installed in about 6 minutes, config from a template in Protosender format, live snapshot, accounting stats | installed part by part, database changes without a restart, snapshot from the broker queues |
| Warm-up in the Hub | the plan holds the daily volume; PowerMTA gets gentle mode for IP addresses being warmed up | the plan holds the daily volume; MailerQ’s own schedules are not used |
| License | sold separately — MailCore System | sold separately — MailCore System |