Active Mode

In active mode, the server initiates the data connection back to the client.

C → S: PORT 192,168,1,5,4,57    (client tells server its IP + port = 4*256+57 = 1081)
S → C: 200 PORT command successful
C → S: RETR file.txt
S → C: 150 Opening data connection
S → C: [opens TCP connection from server's port 20 to client's 192.168.1.5:1081]
S → C: ... file data ...
S: 226 Transfer complete

PORT format: h1,h2,h3,h4,p1,p2 where the IP is h1.h2.h3.h4 and port is p1*256+p2.

Active mode worked well in 1980s LANs. Today, problems:

• Client is usually behind NAT. Server cannot connect TO a client.
• Client firewalls block inbound connections.

Even where it works, it's invasive: server scanning client's listening port.

Server's view:

1. Receive PORT command. Parse client's endpoint.
2. Bind a local port (typically 20 — the data port).
3. On RETR/STOR/LIST: open new TCP connection to the announced endpoint.
4. Transfer data.
5. Close connection (sender side closes after sending; signals end of file).

EPRT (extended PORT, RFC 2428) supports IPv6:

EPRT |2|2001:db8::1|9000|

Family (1=IPv4, 2=IPv6), address, port.

When does active mode still make sense?

• Inside a single trusted network (no NAT).
• Server-initiated jobs that need to push files to clients.
• Almost never on the internet.

For internet usage: passive mode is the answer.