Our client’s systems consist of a back office machine (Windows) and a number
of QNX devices (1-15).
We have recently experienced a flooding of the network thanks to a virus on
the Windows box that caused the device to multi-cast as much as it could to
the point it was taking up 9.8 of the possible 10M available.
Are there any watchdog programs available that monitor usage? If no, what
can be monitored to detect this issue? I’m not wanting to leave this
detection up to the Windows box and thought that a watchdog on each device
would be the best solution.
Thanks,
Doug
Even if/when you detect the situation, how will you handle it? Your network
is already almost saturated, so network connectivity is 'if’y at best, and
non-existant at worst. You could shutdown the networking on the QNX nodes,
and/or perhaps use x10 to turn the Windows box off but that’s hardly a great
way of doing things.
-Adam
Doug Rixmann <rixmannd@rdsdata.com> wrote in message
news:b3dsd5$sbn$1@inn.qnx.com…
Our client’s systems consist of a back office machine (Windows) and a
number
of QNX devices (1-15).
We have recently experienced a flooding of the network thanks to a virus
on
the Windows box that caused the device to multi-cast as much as it could
to
the point it was taking up 9.8 of the possible 10M available.
Are there any watchdog programs available that monitor usage? If no, what
can be monitored to detect this issue? I’m not wanting to leave this
detection up to the Windows box and thought that a watchdog on each device
would be the best solution.
Thanks,
Doug
I was actually thinking more along the line of notifying (gui message) of
potential problems and then figuring out what to do after.
“Adam Mallory” <amallory@qnx.com> wrote in message
news:b3dumj$4c5$1@nntp.qnx.com…
Even if/when you detect the situation, how will you handle it? Your
network
is already almost saturated, so network connectivity is 'if’y at best, and
non-existant at worst. You could shutdown the networking on the QNX
nodes,
and/or perhaps use x10 to turn the Windows box off but that’s hardly a
great
way of doing things.
-Adam
Doug Rixmann <> rixmannd@rdsdata.com> > wrote in message
news:b3dsd5$sbn$> 1@inn.qnx.com> …
Our client’s systems consist of a back office machine (Windows) and a
number
of QNX devices (1-15).
We have recently experienced a flooding of the network thanks to a virus
on
the Windows box that caused the device to multi-cast as much as it could
to
the point it was taking up 9.8 of the possible 10M available.
Are there any watchdog programs available that monitor usage? If no,
what
can be monitored to detect this issue? I’m not wanting to leave this
detection up to the Windows box and thought that a watchdog on each
device
would be the best solution.
Thanks,
Doug
\
Doug Rixmann <rixmannd@rdsdata.com> wrote in message
news:b3duqm$1l4$1@inn.qnx.com…
I was actually thinking more along the line of notifying (gui message) of
potential problems and then figuring out what to do after.
Right, but notification on what computer/system? Since only the QNX4 nodes
are doing this ‘detection’, then I’d have to assume it’s the QNX4 node which
will display the message - otherwise, how will you cause another computer to
take action if the network isn’t available.
-Adam
“Adam Mallory” <amallory@qnx.com> wrote in message
news:b3dumj$4c5$1@nntp.qnx.com…
Even if/when you detect the situation, how will you handle it? Your
network
is already almost saturated, so network connectivity is 'if’y at best, and
non-existant at worst. You could shutdown the networking on the QNX
nodes,
and/or perhaps use x10 to turn the Windows box off but that’s hardly a
great
way of doing things.
If remote spam machine’s NIC support network power down… There are some
variants 
Just switch it off.
// wbr
I’m not so much concerned about WHAT to do. I want to now if it is possible
to detect the network usage.
“Adam Mallory” <amallory@qnx.com> wrote in message
news:b3ed9j$d8l$1@nntp.qnx.com…
Doug Rixmann <> rixmannd@rdsdata.com> > wrote in message
news:b3duqm$1l4$> 1@inn.qnx.com> …
I was actually thinking more along the line of notifying (gui message)
of
potential problems and then figuring out what to do after.
Right, but notification on what computer/system? Since only the QNX4
nodes
are doing this ‘detection’, then I’d have to assume it’s the QNX4 node
which
will display the message - otherwise, how will you cause another computer
to
take action if the network isn’t available.
-Adam
Use the netraw interface to get raw ethernet frames, but you will have to
come up with the heuristic yourself.
-Adam
Doug Rixmann <rixmannd@rdsdata.com> wrote in message
news:b3o75l$njb$1@inn.qnx.com…
I’m not so much concerned about WHAT to do. I want to now if it is
possible
to detect the network usage.
“Adam Mallory” <> amallory@qnx.com> > wrote in message
news:b3ed9j$d8l$> 1@nntp.qnx.com> …
Doug Rixmann <> rixmannd@rdsdata.com> > wrote in message
news:b3duqm$1l4$> 1@inn.qnx.com> …
I was actually thinking more along the line of notifying (gui message)
of
potential problems and then figuring out what to do after.
Right, but notification on what computer/system? Since only the QNX4
nodes
are doing this ‘detection’, then I’d have to assume it’s the QNX4 node
which
will display the message - otherwise, how will you cause another
computer
to
take action if the network isn’t available.
-Adam
\
Adam Mallory wrote:
Use the netraw interface to get raw ethernet frames, but you will have to
come up with the heuristic yourself.
-Adam
netstat -in can be called with a -w secs arg. The output could be piped
to something that hoots when the counts go wild.
Richard
Doug Rixmann <> rixmannd@rdsdata.com> > wrote in message
news:b3o75l$njb$> 1@inn.qnx.com> …
I’m not so much concerned about WHAT to do. I want to now if it is
possible
to detect the network usage.
“Adam Mallory” <> amallory@qnx.com> > wrote in message
news:b3ed9j$d8l$> 1@nntp.qnx.com> …
Doug Rixmann <> rixmannd@rdsdata.com> > wrote in message
news:b3duqm$1l4$> 1@inn.qnx.com> …
I was actually thinking more along the line of notifying (gui message)
of
potential problems and then figuring out what to do after.
Right, but notification on what computer/system? Since only the QNX4
nodes
are doing this ‘detection’, then I’d have to assume it’s the QNX4 node
which
will display the message - otherwise, how will you cause another
computer
to
take action if the network isn’t available.
-Adam
\