>> ....
>> > Im using RMI to connect java applet with a remote server. The
[quoted text clipped - 42 lines]
>
> I replaced the real IP with 222.222.222.222
> This sounds like a network problem rather than a Java one. Have you
> tried the basic tests, e.g. manually pinging the server?
no oliver the web server is running correctly. In fact i can see the
other
content of page like the text but the applet can't contact the remote
object which is declared in the registry in the same web server
through the Internet.
> - Oliver
Oliver Wong - 25 Sep 2006 19:51 GMT
>> This sounds like a network problem rather than a Java one. Have you
>> tried the basic tests, e.g. manually pinging the server?
>>
> no oliver the web server is running correctly.
I'm very confused by your reply. "No", you didn't test, and yet you're
sure everything is configured correctly?
> In fact i can see the
> other
> content of page like the text
Yes, but can you see whatever server your applet is trying to connect to
from the machine on which the applet is failing?
> but the applet can't contact the remote
> object which is declared in the registry in the same web server
> through the Internet.
What registry are you talking about?
- Oliver
Nigel Wade - 26 Sep 2006 10:54 GMT
>> This sounds like a network problem rather than a Java one. Have you
>> tried the basic tests, e.g. manually pinging the server?
[quoted text clipped - 6 lines]
>
>> - Oliver
Does the web server have a firewall, or are there other firewalls in the path?
Do they allow you to contact the RMI registry on the default port of 1099?
The "connection timed out" shows that your attempt to connect to the RMI server
has failed. The fact that the server is running and the connection timed out
indicates that the packets were dropped. The normal response would be that the
connection was reset immediately if nothing was listening on the port and no
firewall was present. This would tend to indicate the actions of a poorly
configured firewall (i.e. dropping rather than rejecting packets). There are
other reasons why it might happen, but a firewall is the most likely.

Signature
Nigel Wade, System Administrator, Space Plasma Physics Group,
University of Leicester, Leicester, LE1 7RH, UK
E-mail : nmw@ion.le.ac.uk
Phone : +44 (0)116 2523548, Fax : +44 (0)116 2523555
EJP - 28 Sep 2006 01:08 GMT
> This would tend to indicate the actions of a poorly
> configured firewall (i.e. dropping rather than rejecting packets).
Not necessarily. There's an argument that says that a firewall that's
really trying to protect an inner resource should behave as though the
resource doesn't even exist, as in this case.
> There are
> other reasons why it might happen, but a firewall is the most likely.
Agreed
Nigel Wade - 28 Sep 2006 09:59 GMT
>> This would tend to indicate the actions of a poorly
>> configured firewall (i.e. dropping rather than rejecting packets).
>
> Not necessarily. There's an argument that says that a firewall that's
> really trying to protect an inner resource should behave as though the
> resource doesn't even exist, as in this case.
It's a bad argument.
If a firewall on a web server drops packets the firewall is *not* behaving as if
the web server did not exist. The idea that "stealth" mode is somehow magically
proving extra security for your system is wrong and is most often propagated by
misinformed web sites, the most infamous being GRCs website.
If I attempt to verify the existence of a system and the packets are dropped I
can be fairly certain that something is hiding the system. If the system simply
didn't exist I would almost certainly receive an ICMP host unreachable from the
upstream router. Doing basic route tracing to find where the packets are
dropped may tell me more. If the packets are dropped at a border firewall I
can't discover much. However, if they are dropped at the host by a software
firewall I know the system exists, that it's up and running and is "protected"
by a firewall which was most likely setup by someone who doesn't really
understand what they are doing - i.e. a target well worth further
investigation.
Configuring the host firewall to return ICMP host unreachable, or RST, would
provide exactly the same level of protection, would be much more polite and
would not break standard networking protocols.

Signature
Nigel Wade, System Administrator, Space Plasma Physics Group,
University of Leicester, Leicester, LE1 7RH, UK
E-mail : nmw@ion.le.ac.uk
Phone : +44 (0)116 2523548, Fax : +44 (0)116 2523555
EJP - 29 Sep 2006 03:44 GMT
> It's a bad argument.
Quite possibly, but there are existing products that do it, so you have
to cope.