Radius Accounting Accuracy

Hi,

I use PPPoE with Radius extensively, and I’ve notice a substantial difference in accounting using PPP on 3.23… Anyone else noticed this, or can shed some light on it perhaps?

Stats on my PPPoE Client:

[admin@Client] > /interface print stats where name="pppoe-out1"
Flags: D - dynamic, X - disabled, R - running, S - slave
 #     NAME                                                                                                                                                                BYTES                     PACKETS             DROPS     ERRORS
 0  R  pppoe-out1                                                                                                                                                          17727147826/454002        23340464/7501       0/0       0/0

Radius Accounting Update:

rad_recv: Accounting-Request packet from host 192.168.1.1:53785, id=252, length=286
        Service-Type = Framed-User
        Framed-Protocol = PPP
        NAS-Port = 604
        NAS-Port-Type = Ethernet
        User-Name = "cknipe"
        Calling-Station-Id = "00:0C:29:A0:3F:04"
        Called-Station-Id = "ip-pppoe"
        NAS-Port-Id = "PPPoE Bridge"
        Class = 0x5050506f45457374617465
        Acct-Session-Id = "81300050"
        Framed-IP-Address = 10.0.10.108
        Acct-Authentic = RADIUS
        Event-Timestamp = "May 18 2009 02:13:06 SAST"
        Acct-Session-Time = 3600
        Acct-Input-Octets = 449738
        Acct-Input-Gigawords = 0
        Acct-Input-Packets = 7431
        Acct-Output-Octets = 2351451390
        Acct-Output-Gigawords = 4
        Acct-Output-Packets = 25706175
        Acct-Status-Type = Interim-Update
        NAS-IP-Address = 192.168.1.1
        Acct-Delay-Time = 0
        Mikrotik-Realm = "opticnetworks.net"

Based on http://www.mikrotik.com/testdocs/ros/2.9/guide/aaa_radius_content.php, the size of 1 Gigaword = 2^32, or 4, 294, 967, 296. Looking at the output value, 4 x 4, 294, 967, 296 = 17, 179, 869, 184 + 2, 351, 451, 390 = 19, 531, 320, 574.

There is a substantial difference here… :confused: