Is there a server-imposed time limit for the client to send a Logon once physical connection is established

, ,
1

Imported from previous forum

2

[ original email was from David Hu - davidhu@netvigator.com ]
Hi, just wondering if anyone would know if FIX 4.0 imposes a time limit as to how soon the client needs to send a Logon message once it establishes a physical connection with the server ?

Thanks.

David Hu
CSFB HK

3

The Logon exchange should be the first messages sent and received. If no data is received within a reasonable period of time (typically tied to the HeartBtInterval or a mulitple of it), then you can consider the connection unresponsive/dead and should terminate it (don’t send TestRequests if not properly initiated with Logons). I’m not sure if your question is regarding the “initiator” (the one who opens the socket) sending the first Logon or the “acceptor” (the one who accepts the socket) sending the Logon response or both.

> Hi, just wondering if anyone would know if FIX 4.0 imposes a time limit as to how soon the client needs to send a Logon message once it establishes a physical connection with the server ?
>
> Thanks.
>
> David Hu
> CSFB HK
>

4

[ original email was from David Hu - wof39@hotmail.com ]
Hi Scott,
I am referring to whether the “acceptor” will impose a time limit on the “initiator” sending the first logon message. The “acceptor” knows when the physical connection is established. However since the “initiator” hasn’t sent the heartbeat interval to the “acceptor” yet in this situation, the “acceptor” does not know what timer value he needs to use to consider the link unresponsive.
Thanks for any comments suggestions.

> The Logon exchange should be the first messages sent and received. If no data is received within a reasonable period of time (typically tied to the HeartBtInterval or a mulitple of it), then you can consider the connection unresponsive/dead and should terminate it (don’t send TestRequests if not properly initiated with Logons). I’m not sure if your question is regarding the “initiator” (the one who opens the socket) sending the first Logon or the “acceptor” (the one who accepts the socket) sending the Logon response or both.
>
> > Hi, just wondering if anyone would know if FIX 4.0 imposes a time limit as to how soon the client needs to send a Logon message once it establishes a physical connection with the server ?
> >
> > Thanks.
> >
> > David Hu
> > CSFB HK
> >
>

5

It’s not specified, so I would use the “reasonable” rule. Note that I’ve seen PGP take a few seconds. If you need a specific number, I would recommend considering the HeartBtInterval. I, personally, wait a maximum of 100 seconds. Note also that in practice sometimes you can successfully open a socket but not really connect to the other side if “proxy firewalls” are involved.

> Hi Scott,
> I am referring to whether the “acceptor” will impose a time limit on the “initiator” sending the first logon message. The “acceptor” knows when the physical connection is established. However since the “initiator” hasn’t sent the heartbeat interval to the “acceptor” yet in this situation, the “acceptor” does not know what timer value he needs to use to consider the link unresponsive.
> Thanks for any comments suggestions.
>
> > The Logon exchange should be the first messages sent and received. If no data is received within a reasonable period of time (typically tied to the HeartBtInterval or a mulitple of it), then you can consider the connection unresponsive/dead and should terminate it (don’t send TestRequests if not properly initiated with Logons). I’m not sure if your question is regarding the “initiator” (the one who opens the socket) sending the first Logon or the “acceptor” (the one who accepts the socket) sending the Logon response or both.
> >
> > > Hi, just wondering if anyone would know if FIX 4.0 imposes a time limit as to how soon the client needs to send a Logon message once it establishes a physical connection with the server ?
> > >
> > > Thanks.
> > >
> > > David Hu
> > > CSFB HK
> > >
> >
>