In ePOST, email messages received from an email client program are parsed and the MIME components of the message (message body and any attachments) are stored as separate messages in POST. Thus, frequently circulated attachments are stored in the system only once.
The message components are first inserted into POST by the sender's ePOST daemon; then, a notification message is sent to the recipient. Sending a message or attachment to a large number of recipients requires very little additional storage overhead beyond sending to a single recipient. If messages are forwarded or sent by different users, the original message data does not need to be stored again; the original message reference is reused.
Due to the necessary data replication in PAST, the storage overhead per message is higher in POST compared to a conventional server-based email system. However, this effect is partly offset by POST's single-copy store, which eliminates large amounts of duplication due to large, widely circulated email attachments. Moreover, exploiting the typically underutilized disk space on desktop computers should more than compensate for this overhead [1]. Lastly, the storage requirements can be further reduced by using erasure codes [15], but we have not yet explored their use in POST.