Iv noticed an issue with my route to IB about 6 months ago. Up until mid april this year I pinged a nice 12-14ms to the gateway at IB from Canada. Then there was a sudden change where my ping increased to 25ms.
A lot of you are probably thinking that 11-13ms increase isn't worth talking about, when it comes to automated systems competing against other automated systems it makes a difference. Enough so that I'll spend a day coding new data structures to reduce my application's reaction time by 0.5ms or less.
I emailed my ISP and they said the node which caused the increase in ping time belongs to Verizon who they not affiliated with. They suggested I email IB because they are Verizon clients and will have a better chance of causing action (interactivebrokers-gw.customer.alter.net). Then have them lodge a complaint about the node in question directly with Verizon.
Four months have passed and I haven't heard anything back from IB. Maybe Verizon 'upgraded' hardware and it isn't actually a solvable problem, maybe IB thought the ping change wasn't significant enough to look into.
So here I am posting at ET hoping someone here has more knowledge in this area than me and might be able to suggest some path to a solution which I havnt thought of. Or if anyone knows someone who works at Verizon in this area.
tracert from kingston ON.
6 26 ms 25 ms 27 ms GigabitEthernet3-1.GW2.CHI13.ALTER.NET [63.65.18.153]
7 25 ms 25 ms 25 ms 0.so-2-1-0.XL2.CHI13.ALTER.NET [152.63.71.218]
8 57 ms 57 ms 57 ms 0.so-7-0-0.XL2.BOS4.ALTER.NET [152.63.0.221]
9 63 ms 56 ms 57 ms POS7-0-0.GW12.BOS4.ALTER.NET [152.63.22.181]
10 61 ms 61 ms 62 ms interactivebrokers-gw.customer.alter.net [208.192.181.62]
tracert from ottawa ON.
6 0.so-0-0-0.XT1.MTL1.ALTER.NET (152.63.133.45) 3.666 ms 3.841 ms 3.508 ms
7 0.so-6-0-0.XL1.BOS4.ALTER.NET (152.63.16.129) 20.968 ms 20.927 ms 20.794 ms
8 POS6-0-0.GW12.BOS4.ALTER.NET (152.63.22.177) 20.649 ms 20.697 ms 20.629 ms
9 interactivebrokers-gw.customer.alter.net (208.192.181.62) 25.431 ms 25.679 ms 25.076 ms
As you can see from the two trace routes the hop coming into the BOS4.ALTER.NET adds significant latency regardless if its coming from Montreal or Chicago. I also had someone from California do a tracert and they found the same problem at the BOS4.ALTER.NET node. This could be due to a hardware change at the site, a hardware issue at the site, or simply a byproduct of the geographic location of Boston and is unsolvable. I am hoping that since I did have a latency time of 12ms before, it is solvable.
It seems that most internet traffic going to IB's gateway gets routed through Boston so I'd imagine that this ping increase has effected most of their north american clients.
Id be interested in seeing if anyone else using IB sees this problem also. You can see your path to IB by running the command 'tracert gw1.ibllc.com' in command prompt. Feel free to post snippets of the route that might shed some light on the situation.
A lot of you are probably thinking that 11-13ms increase isn't worth talking about, when it comes to automated systems competing against other automated systems it makes a difference. Enough so that I'll spend a day coding new data structures to reduce my application's reaction time by 0.5ms or less.
I emailed my ISP and they said the node which caused the increase in ping time belongs to Verizon who they not affiliated with. They suggested I email IB because they are Verizon clients and will have a better chance of causing action (interactivebrokers-gw.customer.alter.net). Then have them lodge a complaint about the node in question directly with Verizon.
Four months have passed and I haven't heard anything back from IB. Maybe Verizon 'upgraded' hardware and it isn't actually a solvable problem, maybe IB thought the ping change wasn't significant enough to look into.
So here I am posting at ET hoping someone here has more knowledge in this area than me and might be able to suggest some path to a solution which I havnt thought of. Or if anyone knows someone who works at Verizon in this area.
tracert from kingston ON.
6 26 ms 25 ms 27 ms GigabitEthernet3-1.GW2.CHI13.ALTER.NET [63.65.18.153]
7 25 ms 25 ms 25 ms 0.so-2-1-0.XL2.CHI13.ALTER.NET [152.63.71.218]
8 57 ms 57 ms 57 ms 0.so-7-0-0.XL2.BOS4.ALTER.NET [152.63.0.221]
9 63 ms 56 ms 57 ms POS7-0-0.GW12.BOS4.ALTER.NET [152.63.22.181]
10 61 ms 61 ms 62 ms interactivebrokers-gw.customer.alter.net [208.192.181.62]
tracert from ottawa ON.
6 0.so-0-0-0.XT1.MTL1.ALTER.NET (152.63.133.45) 3.666 ms 3.841 ms 3.508 ms
7 0.so-6-0-0.XL1.BOS4.ALTER.NET (152.63.16.129) 20.968 ms 20.927 ms 20.794 ms
8 POS6-0-0.GW12.BOS4.ALTER.NET (152.63.22.177) 20.649 ms 20.697 ms 20.629 ms
9 interactivebrokers-gw.customer.alter.net (208.192.181.62) 25.431 ms 25.679 ms 25.076 ms
As you can see from the two trace routes the hop coming into the BOS4.ALTER.NET adds significant latency regardless if its coming from Montreal or Chicago. I also had someone from California do a tracert and they found the same problem at the BOS4.ALTER.NET node. This could be due to a hardware change at the site, a hardware issue at the site, or simply a byproduct of the geographic location of Boston and is unsolvable. I am hoping that since I did have a latency time of 12ms before, it is solvable.
It seems that most internet traffic going to IB's gateway gets routed through Boston so I'd imagine that this ping increase has effected most of their north american clients.
Id be interested in seeing if anyone else using IB sees this problem also. You can see your path to IB by running the command 'tracert gw1.ibllc.com' in command prompt. Feel free to post snippets of the route that might shed some light on the situation.
