Re: TCP SYNSENT Timeout [7:66178]
I don't know any Java but standard UNIX sockets allow a non-blocking connect. Thus you don't care what the underlying stack is doing, you just time-out at the application layer. rgds Marc John Neiberger wrote: > > One of our programmers is asking me about this and I really don't have an > answer. I've checked RFC 793 and haven't spotted the answer yet. > > Is there a default time specified in TCP to remain in the SYN SENT state? > If a device sends a SYN and doesn't receive a response, is the timeout a > built-in TCP parameter or is that a function of the application or operating > system? > > I'm starting to think that this is specific to the operating system, but we > have a need to make it specific to a certain connection without affecting > all TCP connections. To be specific, they're writing something in Java > 1.3.1 (I think) and it doesn't have the capability to tweak TCP parameters. > For a particular set of connections they'd like the timeout to be 10 > seconds, but it seems to be defaulting to 45. > > They tell me that if we were using Java 1.4 they'd be able to adjust these > parameters, which makes me think this is an application or OS-specific > parameter and is only relevant to a particular TCP implementation and could > vary from platform to platform. > > Any thoughts on this? > > Many thanks, > John Message Posted at: http://www.groupstudy.com/form/read.php?f=7&i=66286&t=66178 -- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
TCP SYNSENT Timeout [7:66178]
One of our programmers is asking me about this and I really don't have an answer. I've checked RFC 793 and haven't spotted the answer yet. Is there a default time specified in TCP to remain in the SYN SENT state? If a device sends a SYN and doesn't receive a response, is the timeout a built-in TCP parameter or is that a function of the application or operating system? I'm starting to think that this is specific to the operating system, but we have a need to make it specific to a certain connection without affecting all TCP connections. To be specific, they're writing something in Java 1.3.1 (I think) and it doesn't have the capability to tweak TCP parameters. For a particular set of connections they'd like the timeout to be 10 seconds, but it seems to be defaulting to 45. They tell me that if we were using Java 1.4 they'd be able to adjust these parameters, which makes me think this is an application or OS-specific parameter and is only relevant to a particular TCP implementation and could vary from platform to platform. Any thoughts on this? Many thanks, John Message Posted at: http://www.groupstudy.com/form/read.php?f=7&i=66178&t=66178 -- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
RE: TCP SYNSENT Timeout [7:66178]
Did you check the host requirements RFC, RFC 1122? I think it's OS-dependent though and not specified Priscilla John Neiberger wrote: > > One of our programmers is asking me about this and I really > don't have an answer. I've checked RFC 793 and haven't spotted > the answer yet. > > Is there a default time specified in TCP to remain in the SYN > SENT state? If a device sends a SYN and doesn't receive a > response, is the timeout a built-in TCP parameter or is that a > function of the application or operating system? > > I'm starting to think that this is specific to the operating > system, but we have a need to make it specific to a certain > connection without affecting all TCP connections. To be > specific, they're writing something in Java 1.3.1 (I think) and > it doesn't have the capability to tweak TCP parameters. For a > particular set of connections they'd like the timeout to be 10 > seconds, but it seems to be defaulting to 45. > > They tell me that if we were using Java 1.4 they'd be able to > adjust these parameters, which makes me think this is an > application or OS-specific parameter and is only relevant to a > particular TCP implementation and could vary from platform to > platform. > > Any thoughts on this? > > Many thanks, > John > > Message Posted at: http://www.groupstudy.com/form/read.php?f=7&i=66185&t=66178 -- FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
