Support Board
Date/Time: Fri, 22 Nov 2024 21:56:16 +0000
Post From: New TT trading
| [2019-04-13 23:08:55] |
| |
|
We apologize for the delay. Immediately after the issue, that evening we did move to a different server with a different network for one of the servers (Server 1). We changed Server 1 to a different Chicago server which uses different network connectivity. Server 2 was still with Barchart but they did not have any further incidents. And we have now the CME Aurora server set up with Server 2 pointing to it. We will provide a table showing the current assignments for each of the servers. Another issue that occurred was with not being able to reconnect to TT. When there was an abnormal loss of connection to the TT FIX server, when our order routing software reconnected, TT denied the connection for two minutes because it was unsure of the prior session state. We fixed this by adding a backup session that we switch to under this condition. And we also bypass the TT FIX recovery process under this condition as well so we do not take another 5 to 10 seconds to get through that. The time for that is on the TT side. Not our side. Everything within Sierra Chart can be handled within microseconds. In regards to the above issue, we have not experienced this before with TT FIX and we have 10 years of experience with TT FIX and have completed conformance with TT and have done extensive testing. So this really caught us off guard. But we have implemented a durable solution. And then there was also the translation issue between the numeric identifier and the actual symbol for positions. We could not identify what went wrong on that, but in general it must have been an issue on the TT side with the instrument data response not being complete or correct. And there is a lot of complexity on the TT side with this. We did make one refinement with it related to caching so if there is any problem, it will be corrected on a reconnection. We would not expect further issues because it did center around the connectivity issue. So we have definitively addressed every issue here. We are also looking into a direct cross connect in the Aurora data center to TT bypassing the Internet altogether and instead just a direct patch cable/link between systems. Although currently the switching is still within the data center itself. So it is fast and reliable. This is really just contingent on the cost (which we are checking on). At this point we just have a handful of users using this new routing service. This is the current table of servers (Step 11): https://www.sierrachart.com/index.php?page=doc/SierraChartOrderRoutingServiceWithData.php#SetupInstructions The next step is for us to complete the setup of our new server in the Cermak Equinix data center. And there is also the issue of the interruption with the Sierra Chart Exchange Data Feed itself due to the Barchart network. Barchart replaced this weekend one of their core switches to resolve that problem. But that to us is not good enough because clearly they do not have sufficient redundancy. We are working on alternative plans but we cannot publicly discuss that. Sierra Chart Support - Engineering Level Your definitive source for support. Other responses are from users. Try to keep your questions brief and to the point. Be aware of support policy: https://www.sierrachart.com/index.php?l=PostingInformation.php#GeneralInformation For the most reliable, advanced, and zero cost futures order routing, *change* to the Teton service: Sierra Chart Teton Futures Order Routing Date Time Of Last Edit: 2019-04-14 09:02:55
|