A system, method and program product for managing e-mails from a source suspected of sending spam. The e-mails are received at a firewall or router en route to a mail server. A determination is made whether a source has sent an e-mail which exhibits characteristics of spam. In response, subsequent e-mails from the source destined for the mail server are rate-limiting at the firewall or router such that the firewall or router limits a rate at which the subsequent e-mails are forwarded from the firewall or router to the mail server. The rate is predetermined and less than a maximum rate at which the firewall or router can physically forward e-mails to the mail server absent the rate limit. A determination is made whether another source has sent another e-mail which exhibits more characteristics of spam than the first said e-mail. In response, subsequent e-mails from this other source are blocked at the firewall or router. The rate limit can be a limit on a number of e-mails per unit of time from the source that will be forwarded from the firewall or router to the mail server.