Quote from comintel:
It would be impossible to pass through every single transaction in real time over the public internet without risking falling behind occasionally (which is what happens at other brokers who attempt it).
I think a lot of people do not understand this and why it is so.
When I started futures trading I opted for a broker with the Rithmic feed because of all the wonderful things I read about it when doing my research.
Being in Thailand with great domestic Internet speeds that fall off dramatically for international connections is not a great foundation for the better feeds.
First sign all was not right was putting in limit orders 2 ticks away from the market in ES during RTH and getting an instant fill.
Charting appeared perfectly normal, and it did not happen all the time so no big deal. The perfect storm of fast markets and slow Internet connection did not happen every week even, so one did not notice.
Until one fine day the market appeared to be exceptionally slow moving, I hovered the cursor over the bar and checked against the clock and horrors, the bar was about 15 minutes behind the clock.
Days of testing and checking everything with the help of my brokers TS manager found nothing wrong save the Internet connection, and with 10 to 15 international hops AFTER it left Thailand, nothing the ISP could fix on their own either.
Bottom line, this was the explanation given by the TS manager. Data feeds use the packet system, and because of the sensitivity of the data and its use, later packets are not sent until the server gets confirmation that earlier packets have been received correctly and in full. BTW, line quality matters hugely, when I faced up to 20% packet loss was when things were worst. So if there is constant re-sending because of Internet issues, you fall behind.
Someone on the Sierra Charts support board was kind enough to provide me a spreadsheet that constantly monitored the data time stamps against the computer clock. One could choose the threshold for the alarm to trigger and it was a real eye opener how much the feed lagged. Less on CL, more on ES because of the volumes.
I don't have the problem with IB, so I'll happily live with the snapshot data. At least I'm not looking at what happened 15 minutes ago and thinking it is real time.