秋田预测市场 · Discord 审计

FIX API是否具有延迟和可靠性优势的争议

2026-07-16 · Kalshi · 35 条相关讨论

theed 2026-07-16 02:43:42

Another unrelated question, is it worth the time to migrate to FIX?

Another unrelated question, is it worth the time to migrate to FIX?

Bumbling Bayesian 2026-07-16 02:44:09

I think the only people who might know the answer to that are not giving away any edge.

I think the only people who might know the answer to that are not giving away any edge.

theed 2026-07-16 02:44:11

Is there an improvement in lifecycle latency?

Is there an improvement in lifecycle latency?

theed 2026-07-16 02:45:06

I'll infer that means no, there's no improvement

I'll infer that means no, there's no improvement

theed 2026-07-16 02:45:10

got it

got it

Bumbling Bayesian 2026-07-16 02:45:56

I certainly do not arrive at the same conclusion.

I certainly do not arrive at the same conclusion.

Marcoase 2026-07-16 02:46:12

you have FIX access as non-institutional and low volume?

you have FIX access as non-institutional and low volume?

Bumbling Bayesian 2026-07-16 02:46:21

I am right now wondering if FIX is more reliable for getting events due to the Web Socket feed being flaky lately.

I am right now wondering if FIX is more reliable for getting events due to the Web Socket feed being flaky lately.

Marcoase 2026-07-16 02:46:26

some have claimed lower latency via FIX

some have claimed lower latency via FIX

Apex 2026-07-16 02:46:30

high volume non institutional have FIX access

high volume non institutional have FIX access

Apex 2026-07-16 02:46:38

premier api tier

premier api tier

Marcoase 2026-07-16 02:46:46

yes but low volume part is why it's confusng

yes but low volume part is why it's confusng

Apex 2026-07-16 02:51:42

rest and fix are processed on different paths on the exchange's side but from what I could measure FIX had lower latency

rest and fix are processed on different paths on the exchange's side but from what I could measure FIX had lower latency

theed 2026-07-16 02:53:34

<@351839065919979530> do you mind if I DM you?

<@351839065919979530> do you mind if I DM you?

Apex 2026-07-16 02:53:38

sure

sure

betaclone 2026-07-16 03:17:17

I'd imagine a big part of it is simply that there's less traffic on FIX. Hence the gatekeeping

I'd imagine a big part of it is simply that there's less traffic on FIX. Hence the gatekeeping

betaclone 2026-07-16 03:17:56

If they gave everyone access to FIX you'd probably start to see similar latencies between that and REST. Just a theory, no way to test it

If they gave everyone access to FIX you'd probably start to see similar latencies between that and REST. Just a theory, no way to test it

betaclone 2026-07-16 03:18:39

They're clearly trying to keep a VIP lane open for the big boys providing the bulk of the liquidity

They're clearly trying to keep a VIP lane open for the big boys providing the bulk of the liquidity

SteveB 2026-07-16 03:24:50

tmk FIX is on a private link which is going to have a more consistent connection pathway. So less p99 spikes

tmk FIX is on a private link which is going to have a more consistent connection pathway. So less p99 spikes

benstu 2026-07-16 03:38:54

private link is a separate, harder-to-get, benefit than just FIX

private link is a separate, harder-to-get, benefit than just FIX

lesik 2026-07-16 04:24:08

Did kalshi add the orderbooks stream functional to fix api? I've heard fix had only place order advantage and everyone used default ws.

Did kalshi add the orderbooks stream functional to fix api? I've heard fix had only place order advantage and everyone used default ws.

lesik 2026-07-16 04:26:48

https://docs.kalshi.com/fix/market-data this is basically a websocket replacement, am i correct ?

https://docs.kalshi.com/fix/market-data this is basically a websocket replacement, am i correct ?

betaclone 2026-07-16 04:27:32

seems like it but i don't think anyone here has FIX access

seems like it but i don't think anyone here has FIX access

maverick23 2026-07-16 08:30:35

are you able to quote rfqs through the rest api or it has to be through fix?

are you able to quote rfqs through the rest api or it has to be through fix?

Apex 2026-07-16 08:31:13

fix does not give any special access

fix does not give any special access

Apex 2026-07-16 08:31:18

you can do everything through rest

you can do everything through rest

maverick23 2026-07-16 08:32:35

awesome thank you. gladd to know i don't need fix access to submit bids for rfqs and can do it thourgh the rest api was stressing about it.

awesome thank you. gladd to know i don't need fix access to submit bids for rfqs and can do it thourgh the rest api was stressing about it.

betaclone 2026-07-16 09:12:48

I mean it's well documented in the REST docs, not sure why you'd be stressing it

I mean it's well documented in the REST docs, not sure why you'd be stressing it

benstu 2026-07-16 09:18:50

i've seen no latency benefit for FIX compared to rest but i might be missing something

i've seen no latency benefit for FIX compared to rest but i might be missing something

Apex 2026-07-16 10:04:58

I remeasured and I dont have a definitive conclusion that one is faster than the other

I remeasured and I dont have a definitive conclusion that one is faster than the other

Noah 2026-07-16 10:33:31

At the network level itself, I see that first hop time is shorter for REST (`0.25 ms`) vs FIX (`0.80 ms`). But REST first hop is to ELB, so we don't know how long it has to go from there to Kalshi hosts. And FIX might get you there in one hop. Either way, still sub millisecond at the network itself

At the network level itself, I see that first hop time is shorter for REST (`0.25 ms`) vs FIX (`0.80 ms`). But REST first hop is to ELB, so we don't know how long it has to go from there to Kalshi hosts. And FIX might get you there in one hop. Either way, still sub millisecond at the network itself

Noah 2026-07-16 10:36:07

I'd guess the Kalshi server performance ends up as the bottleneck if you got this latency sensitive. (speculation)

I'd guess the Kalshi server performance ends up as the bottleneck if you got this latency sensitive. (speculation)

lactobacillus 2026-07-16 13:00:17

fix has features not present in rest/websocket, such as `CancelOrdersOnDisconnect` (https://docs.kalshi.com/fix/authentication#optional-fields)

fix has features not present in rest/websocket, such as `CancelOrdersOnDisconnect` (https://docs.kalshi.com/fix/authentication#optional-fields)

Apex 2026-07-16 13:00:47

Yeah I was mostly referring to actual trading operations

Yeah I was mostly referring to actual trading operations

Apex 2026-07-16 13:00:49

thats true though

thats true though