Why is IB's time 5 seconds behind the official time?

Quote from lindq:
As much as 2-3 points difference on ES. Other futures contracts also way off.
More...
Which system seemed to be lagging?

During a lull one should have caught up with the other.

Jack
 
clock1.jpg


Quote from Tums:

Care to post a screen shot?
Who is faster?
More...
 
Quote from Port1385:

IB's timer is still like 5 seconds off. The market closes and then I still see trades for at least another 5-10 seconds.
More...

others reported that their paper trading accounts have similar problems.
 
Quote from shortie:

IB's timer is still like 5 seconds off. The market closes and then I still see trades for at least another 5-10 seconds.

others reported that their paper trading accounts have similar problems.
More...
I reported this problem 2 years ago.
The key to the solution is NTP.
http://en.wikipedia.org/wiki/Network_Time_Protocol
What's amazing in my experience: it's difficult to find an accurate NTP server !
Could that be IB's problem also ?
 
Quote from Port1385:

IB's timer is still like 5 seconds off. The market closes and then I still see trades for at least another 5-10 seconds.
More...

Why does that have to be ib's timer?
Couldn't it be the market or market reporting?

On the other hand I'm looking at TWS and my clock (accurate to within 1/100th of a second currently) and TWS is a couple of seconds behind.

Of course it doesn't matter to me because Sierra Chart timestamps the incoming data based on system time not the time on the bottom of the TWS window.
 
You can actually see the exact offset TWS is using in the log..txt files.

Code: 
MR 09:25:04:001 AWT-EventQueue-0: New CCP Offset: -1121 ms, Old:-1120 CCP:1235658304000 Local:1235658305121
[b]MR 09:35:03:992 JTS-CCPDispatcher-15: New CCP Offset: -113 ms, Old:-1121 CCP:1235658905000 Local:1235658905113[/b]
MR 10:05:04:992 JTS-CCPDispatcher-15: New CCP Offset: -90 ms, Old:-98 CCP:1235660705000 Local:1235660705090


KD 03:56:30:993 JTS-CCPDispatcher-14: New CCP Offset: -842 ms, Old:-849 CCP:1235379391000 Local:1235379391842
KD 06:46:30:993 JTS-CCPDispatcher-14: New CCP Offset: -1108 ms, Old:-1115 CCP:1235389591000 Local:1235389592108

[b]KD 09:26:22:013 JTS-CCPDispatcher-24484: New CCP Offset: -107 ms, Old:-1094 CCP:1235485583000 Local:1235485583107[/b]

MR 14:25:04:991 JTS-CCPDispatcher-15: New CCP Offset: -282 ms, Old:-291 CCP:1235676305000 Local:1235676305282

KD 21:36:23:001 AWT-EventQueue-0: New CCP Offset: -762 ms, Old:-761 CCP:1235529383000 Local:1235529383762

VY 01:52:34:011 JTS-CCPDispatcher-15: New CCP Offset: -949 ms, Old:-938 CCP:1235631154000 Local:1235631154949
VY 02:52:34:201 AWT-EventQueue-0: New CCP Offset: -1025 ms, Old:-824 CCP:1235634754000 Local:1235634755025

MR 06:25:04:001 AWT-EventQueue-0: New CCP Offset: -857 ms, Old:-856 CCP:1235647504000 Local:1235647504857

I dont think it should vary so much, my server's clock is synchronized every 5 minutes and is always withing 10ms, and it's only 45ms round trip between me. I was thinking maybe it had to do with load on the servers

It seems like the IB time usually becomes in sync at around 9:25 am est and then drifts off from their for the rest of the day. Some days it syncs at different times, usually when the offset becomes over 1000ms
 
Actually I think the servers synchronize whenever the offset is over 1000ms, which just usually happens around 9:30. this is odd because it would mean they are checking the time more often than once a day, but they deliberately don't fix it until the offset becomes greater than a second. they should keep them constantly synced, then again maybe their servers would break if they are coded like TWS

and make the reqCurrentTime API call return milliseconds. that would be helpful.
 
You must log in or register to reply here.
Back
Top