Network question

[techguy85]

This is very strange.
The only time I would think randomizing your MAC address would be at all helpful is to keep that from identifying you directly on a non-trusted public WiFi network. There is no reason that a fixed piece of hardware should be doing that, each time the MAC changes, the device would get a new DHCP address, which could be the reason behind you loosing access to it. Can you configure the DVR to use a statically assigned address?

[techmike]

Strange? Yes, yes it is. The DVR is configured to use the same IP. I have tried using reserved IP's in the router, but since that is associated with the MAC, I get the same result. I currently have the DHCP pool restricted to 4 clients, but again when the MAC changes, the DVR is a new "unknown device" with the same IP as when it was a known device. I thought it might be related to the DHCP lease renewing, but the timing does not match.

[AlanM]

Is it possible that the Chinese aren't registering (purchasing) MAC addresses and just winging it?

[rimfireOH]

Is it possible that the Chinese aren't registering (purchasing) MAC addresses and just winging it?

This is my theory. But it doesn't explain why it's not sticking to a single MAC.

[JediSkipdogg]

I think he has a defective unit. Zosi is pretty big in the security camera world and I couldn't find anyone else having this issue online.

Of course, it still doesn't answer why his unit is doing it though.

[techmike]

Zosi has shipped my new DVR. There are Amazon reviews complaining of losing remote viewing connection. Wonder if mine is not the only one with this issue?

[JediSkipdogg]

Zosi has shipped my new DVR. There are Amazon reviews complaining of losing remote viewing connection. Wonder if mine is not the only one with this issue?

Could be and you are just smart enough to know to notice the MAC address changing. Glad they are sending you a new one. I'm looking at adding a network DVR and cameras to my house and barn and Zosi is one of my options.

[djthomas]

Here's what I suspect - crappy programming. There's no legitimate reason for this behavior, but at the same time I can't think of anything nefarious either.

Hardware programming is often done by different people at different times. It's possible that at the time the programmer wrote the firmware code the function to access the device's MAC wasn't available. So he told it to just pick one at random, allowing him to get on with his work. He couldn't hardcode one because then it would be difficult to test multiple devices, there would be DHCP caching issues, etc. As shown here in the field using a random MAC does work, so the programmer likely 1) forgot to go back and change the function responsible for getting the MAC, 2) quit, moved on, died, etc, 3) just didn't give a crap.

That leads to the second part of my theory. Not being the best and brightest programmer out there his code is probably pretty buggy. So buggy in fact that the NIC is crashing every now and then. It reboots itself, and in the process comes up with another random MAC.

If you really cared enough you could run a packet capture to see what's going on ... not that you can fix it or anything. I'll bet each time the MAC changes it's broadcasting traffic typically seen when a device boots up and asks who's out there. Years ago I used to do network performance management work. We'd get called in when there were performance complaints. A lot of times the problem would be the result of someone bringing their own network device like a switch and plugging it in to a conference room. That wasn't a problem per se but consumer-grade network hardware was (is) notoriously chatty and oftentimes buggy. I remember one time someone had a Netgear 4 port switch and the uplink port was literally restarting itself every half second, and in doing so was generating 2-3 Mb/s all by itself on a 10 M/b line.

[deanimator]

Is it possible that the Chinese aren't registering (purchasing) MAC addresses and just winging it?

That sounds very plausible to me.

[deanimator]

Is it possible that the Chinese aren't registering (purchasing) MAC addresses and just winging it?

This is my theory. But it doesn't explain why it's not sticking to a single MAC.

It's just a question of how the programmer(s) chose to address the problem.

It's entirely possible that they wrote their "error handling" to resolve a MAC address conflict by randomly generating another MAC address and assigning it to the unit. As long as no other device had that MAC address (or was visible), it'd work like "normal"... until another conflict.

[djthomas]

Is it possible that the Chinese aren't registering (purchasing) MAC addresses and just winging it?

This is my theory. But it doesn't explain why it's not sticking to a single MAC.

It's just a question of how the programmer(s) chose to address the problem.

It's entirely possible that they wrote their "error handling" to resolve a MAC address conflict by randomly generating another MAC address and assigning it to the unit. As long as no other device had that MAC address (or was visible), it'd work like "normal"... until another conflict.

The MAC address space is approximately 281 trillion. Based on the screen shots shown above it seems to be picking them from the entire pool as opposed to say, the last two digits. You've got better odds winning the lottery than seeing a conflict between two randomly chosen addresses, let alone the issue OP and other Amazon reviewers are seeing on a regular basis.

[sodbuster95]

The MAC address space is approximately 281 trillion. Based on the screen shots shown above it seems to be picking them from the entire pool as opposed to say, the last two digits. You've got better odds winning the lottery than seeing a conflict between two randomly chosen addresses, let alone the issue OP and other Amazon reviewers are seeing on a regular basis.

Of course, if they got lazy with the code, it might be generating a new MAC every time it boots or every time the DHCP lease expires or any other of a number of events. So, I'd concur with your earlier assessment - crappy programming.

[techguy85]

Here's what I suspect - crappy programming. There's no legitimate reason for this behavior, but at the same time I can't think of anything nefarious either.

Hardware programming is often done by different people at different times. It's possible that at the time the programmer wrote the firmware code the function to access the device's MAC wasn't available. So he told it to just pick one at random, allowing him to get on with his work. He couldn't hardcode one because then it would be difficult to test multiple devices, there would be DHCP caching issues, etc. As shown here in the field using a random MAC does work, so the programmer likely 1) forgot to go back and change the function responsible for getting the MAC, 2) quit, moved on, died, etc, 3) just didn't give a crap.

That leads to the second part of my theory. Not being the best and brightest programmer out there his code is probably pretty buggy. So buggy in fact that the NIC is crashing every now and then. It reboots itself, and in the process comes up with another random MAC.

If you really cared enough you could run a packet capture to see what's going on ... not that you can fix it or anything. I'll bet each time the MAC changes it's broadcasting traffic typically seen when a device boots up and asks who's out there. Years ago I used to do network performance management work. We'd get called in when there were performance complaints. A lot of times the problem would be the result of someone bringing their own network device like a switch and plugging it in to a conference room. That wasn't a problem per se but consumer-grade network hardware was (is) notoriously chatty and oftentimes buggy. I remember one time someone had a Netgear 4 port switch and the uplink port was literally restarting itself every half second, and in doing so was generating 2-3 Mb/s all by itself on a 10 M/b line.

When we bought our house last summer I went ahead and did real wired ethernet and put in some real networking gear e.g. Ubiquiti stuff. It was amazing how much better everything works with that stuff rather than whatever I happened to grab at Microcenter. Sure, I had to put a controller up on a VM if I wanted to control stuff from wherever but even at $1,700 worth of hardware I will say it was worth every penny.

[techmike]

Thanks for all of the comments, info and suggestions. New DVR arrived and is installed. working for 3 days with no issues so far.

[techmike]

Follow-up post - new DVR worked about 5 days, then tablet lost connection with DVR. Odd, in that the tablet could see and ping the DVR, but the ZOSI app would not connect to the ZOSI DVR. The IP and MAC address had not changed. Only way I found to recover was to perform a shutdown, then power cycle the DVR. At this point the MAC address would change, but then everything would connect and work. Until the next time. Time between connection failures appears to be random, as long as 5 days, short as 24 hours. ZOSI customer service has just stated they don't know what the problem is, but to contact Amazon for a refund. Now I am DVR shopping again.