Cyrus IMAP server
|Developer(s)||Carnegie Mellon University|
|Stable release||2.4.17 / December 1, 2012|
|Type||Mail Delivery Agent|
|License||original BSD license|
The Cyrus IMAP server differs from other IMAP server implementations in that it is generally intended to be run on sealed servers, where normal users cannot log in. The mail spool uses a filesystem layout and format similar to the Maildir format used by other popular email servers such as qmail, Courier, Dovecot, etc. Cyrus IMAP's spool format, said[by whom?] to have improved performance and scalability to Maildir, resides in parts of the filesystem that are private to the Cyrus IMAP system. Users can access mail through the IMAP/IMAP-S, POP3/POP3-S or KPOP protocols.
The Cyrus IMAP server supports server-side mail filtering through the implementation of a mail filtering language called Sieve.
The private mailbox database design gives the server considerable advantages in efficiency, scalability, and administratability. Multiple concurrent read/write connections to the same mailbox are permitted. The server supports access control lists on mailboxes and storage quotas on mailbox hierarchies.
Prior to 1994, Carnegie Mellon University exclusively used the locally-developed and non-standard Andrew Messaging System (AMS) for its email communication needs. Originally written in the early 1980s as part of the Andrew Project, it was very advanced for its day. However, it had major scalability issues. Carnegie Mellon wanted to move to a standards-compliant mail system that met or exceeded the feature set of AMS, and with an emphasis on disconnected operation and scalability (scalability both in simultaneous online accesses and in large mailboxes).
In 1994 the Computing Services Division at Carnegie Mellon addressed these goals by starting the Cyrus Project. In 1998 (class of 2002), Carnegie Mellon placed all of its incoming freshmen on the Cyrus server for the first time. In December 2001, bboard access (which had been mirrored from AMS to Cyrus), cut over to Cyrus completely. AMS was finally phased out in May 2002.
As development on the project progressed, it became clear that users required further scalability and additional reliability beyond that which a single-machine configuration could provide. In response the Computing Services Division developed Cyrus "Murder" clustering, and after several revisions deployed it within Carnegie Mellon in the summer of 2002.
- Rob Siemborski (2003-07-24). "Cyrus History". Retrieved 2007-01-13. - Copied with permission.