So,
I’m looking for a means to uniquely identify a QNX4 box as opposed to
another box. I’ve looked into several different things, but nothing has
seemed totally satisfactory. The context: I’m writing something that I would
like to control the deployment of, so I would like a way to bind it to
specific machines.
I have looked into trying to find hardware-specific identifiers, but the
only thing I seem to be able to use is the MAC address. This isn’t a
completely satisfactory solution as a MAC address can be fairly easily
changed by something as simple as an argument passed to a net driver.
Reading an EEPROM of a particular piece of hardware seems like a possibility
but that would depend on that same piece of hardware being in every
deployment.
I’ve also looked into maybe using the inode of a particular file, but this
introduces issues of making things inconvenient for people – suppose
someone wanted to compress their drive… also, this is completely
satisfactory in that someone could just clone an image of a drive and copy
it to another drive.
I’ve looked into maybe using something like the clock drift of particular
machines (between the CPU clock, RTC clock, or BIOS clock), but this doesn’t
seem to be something I could consistently depend on. There seems to be too
much noise in the clock skews of a single machine to be able to dependably
differentiate it from other machines.
Anyone have any good suggestions or areas to explore?
Thanks,
-Albert
Bill Caroselli (Q-TPS) wrote:
“William A. Flowers” <> wflowers_NOSPAM@insightcontrol.com> > wrote in message
news:amqeh7$g67$> 1@inn.qnx.com> …
“Bill Caroselli (Q-TPS)” <> QTPS@EarthLink.net> > wrote in message
news:amq6n6$aso$> 1@inn.qnx.com> …
DOn’t many advanced CPUs have an embedded serial number?
Intel did that briefly with the Pentium III (or was it P2?) but then
stopped. Rumor has it too many people complained about invasion of
privacy
issues.
Not sure about AMD parts.
Gee, I’m have to tell GM that the next time I go to buy a car. I don’t want
no serial number.
And how, exactly, does GM putting a VIN on a car that you purchase
create a potential wide range exploit of your personal information ?
DOn’t many advanced CPUs have an embedded serial number?
“Albert Chu” <albert.chu@attbi.com> wrote in message
news:amosdj$amu$1@inn.qnx.com…
So,
I’m looking for a means to uniquely identify a QNX4 box as opposed to
another box. I’ve looked into several different things, but nothing has
seemed totally satisfactory. The context: I’m writing something that I
would
like to control the deployment of, so I would like a way to bind it to
specific machines.
I have looked into trying to find hardware-specific identifiers, but the
only thing I seem to be able to use is the MAC address. This isn’t a
completely satisfactory solution as a MAC address can be fairly easily
changed by something as simple as an argument passed to a net driver.
Reading an EEPROM of a particular piece of hardware seems like a
possibility
but that would depend on that same piece of hardware being in every
deployment.
I’ve also looked into maybe using the inode of a particular file, but this
introduces issues of making things inconvenient for people – suppose
someone wanted to compress their drive… also, this is completely
satisfactory in that someone could just clone an image of a drive and copy
it to another drive.
I’ve looked into maybe using something like the clock drift of particular
machines (between the CPU clock, RTC clock, or BIOS clock), but this
doesn’t
seem to be something I could consistently depend on. There seems to be too
much noise in the clock skews of a single machine to be able to dependably
differentiate it from other machines.
Anyone have any good suggestions or areas to explore?
Thanks,
-Albert
“Bill Caroselli (Q-TPS)” <QTPS@EarthLink.net> wrote in message
news:amq6n6$aso$1@inn.qnx.com…
DOn’t many advanced CPUs have an embedded serial number?
Intel did that briefly with the Pentium III (or was it P2?) but then
stopped. Rumor has it too many people complained about invasion of privacy
issues.
Not sure about AMD parts.
Bill Flowers
Clearwater, FL
“William A. Flowers” <wflowers_NOSPAM@insightcontrol.com> wrote in message
news:amqeh7$g67$1@inn.qnx.com…
“Bill Caroselli (Q-TPS)” <> QTPS@EarthLink.net> > wrote in message
news:amq6n6$aso$> 1@inn.qnx.com> …
DOn’t many advanced CPUs have an embedded serial number?
Intel did that briefly with the Pentium III (or was it P2?) but then
stopped. Rumor has it too many people complained about invasion of
privacy
issues.
Not sure about AMD parts.
Gee, I’m have to tell GM that the next time I go to buy a car. I don’t want
no serial number.
Albert,
How do you intend to resolve the ID of the QNX box? If you can use a
network, I have a solution.
Write a simple UDP application that receives BROADCAST packets (QNX side).
When received, have the QNX box respond. I have used this in numberous
applications when “looking” for QNX servers on a Windows NT network.
Of course another application must send out the resolution request on the
network – maybe a Windows app, etc.
Larry Fly
“Albert Chu” <albert.chu@attbi.com> wrote in message
news:amosdj$amu$1@inn.qnx.com…
So,
I’m looking for a means to uniquely identify a QNX4 box as opposed to
another box. I’ve looked into several different things, but nothing has
seemed totally satisfactory. The context: I’m writing something that I
would
like to control the deployment of, so I would like a way to bind it to
specific machines.
I have looked into trying to find hardware-specific identifiers, but the
only thing I seem to be able to use is the MAC address. This isn’t a
completely satisfactory solution as a MAC address can be fairly easily
changed by something as simple as an argument passed to a net driver.
Reading an EEPROM of a particular piece of hardware seems like a
possibility
but that would depend on that same piece of hardware being in every
deployment.
I’ve also looked into maybe using the inode of a particular file, but this
introduces issues of making things inconvenient for people – suppose
someone wanted to compress their drive… also, this is completely
satisfactory in that someone could just clone an image of a drive and copy
it to another drive.
I’ve looked into maybe using something like the clock drift of particular
machines (between the CPU clock, RTC clock, or BIOS clock), but this
doesn’t
seem to be something I could consistently depend on. There seems to be too
much noise in the clock skews of a single machine to be able to dependably
differentiate it from other machines.
Anyone have any good suggestions or areas to explore?
Thanks,
-Albert