modem review and showdown

[System Error Message]

Just like we have with router and wifi why dont we do the same with modems? A teardown showing what chips do they use. A performance test using 2 of the same modems in bridge mode of each other, features such as phone sockets and VOIP for example.

 

[KenZ71]

Good question!

 

[sfx2000]

Just like we have with router and wifi why dont we do the same with modems? A teardown showing what chips do they use. A performance test using 2 of the same modems in bridge mode of each other, features such as phone sockets and VOIP for example.

This was asked and answered a couple of weeks back - it requires a substantial amount of backend support (DSLAM, CMTS, etc) - even then, the configs are controlled by the operator, so there's that aspect as well.

 

[KenZ71]

Is that much the same reason we don't see DIY modems?

 

[sfx2000]

Is that much the same reason we don't see DIY modems?

Pretty much... even when we provide our own - e.g. I own my SB6141, it is managed and configured by my provider, best I can do is review status and logs, and reset the config.

 

[System Error Message]

While you may not have control over your own modem and maybe ISPs hand them out but its not just us viewers that may buy them, ISPs may read the reviews too.

There are various things you can measure.
Sync rates
packet losses
signal quality (you would need an oscilloscope for that)
physical features (phone slots for VOIP, gigabit ethernet ports? load balancing multiple analog?)

and maybe more. You dont need a DSLAM, cant you connect 2 modems together in bridge mode?

Also those who would deploy their own lines for remote areas, they could use the advice too. Im not just talking about DSL, theres cable and fibre optics as well for modems. Things that extend further than ethernet.

 

[sfx2000]

While you may not have control over your own modem and maybe ISPs hand them out but its not just us viewers that may buy them, ISPs may read the reviews too.

They likely don't care, as long as consumer gear doesn't interfere with their side.

In the WAN environment - there is a regulated side - the carrier, and the Unregulated Side - which is your stuff - and there's a clear demarcation between the two...

There are various things you can measure.
Sync rates
packet losses
signal quality (you would need an oscilloscope for that)
physical features (phone slots for VOIP, gigabit ethernet ports? load balancing multiple analog?)

And that's why the operators test their end, and approve premises equipment up to the line of demarc.

You dont need a DSLAM, cant you connect 2 modems together in bridge mode?

Go study DSL, the platform and architecture - come back and we'll talk...

Also those who would deploy their own lines for remote areas, they could use the advice too. Im not just talking about DSL, theres cable and fibre optics as well for modems. Things that extend further than ethernet.

Not a valid point - again, because of the line of demarcation between the regulated side provided by the operator, and the customer owned premises equipment.

This is how things are - whether it's small home networks, or carrier grade networks even - and each operator, even with the same exact gear, will be different in their config - which makes testing in a general context, not worth the effort for a general focused website.

That's my opinion - and it bears weight, as that the environment I work in...

I'll leave it to Tim to say yea or nay...

 

[System Error Message]

Thats not my point, my point is if its possible it would be good to actually do reviews on them. While ISPs normally provide modems and do testing many ISPs allow their users to use their own modem but do not provide support for it. In some places, smaller ISPs would likely be reading SNB reviews so there is an audience for this, the question is what do we need in order to actually test a modem. Is a DSLAM really necessary? Even in the US you still do get many small ISPs that only provide internet in a very small area. What about rural areas where people want to lay their own cabling and lease a line? Or 3rd world countries where there are many small ISPs that cant deploy fibre and have to use DSL?

Although people dont really buy modems they tend to do so now much more otherwise draytek wouldnt be in business. Do you see ISPs distributing draytek modems?

 

[sfx2000]

@System Error Message - perhaps you're missing my point - DSL, as an example, it's not symmetric - a DSLAM is not the same as the CPE (the modem in your house). And there are different implementations of how the DSLAM is employed - in the CO, or if you have FTTN, it might be in the Crossbox as a Virtual DSLAM... each operator is different here, and even within the same operator, they're going to deploy it as they need it.

While it might be interesting to see the WAN side, at the end of the day, it really doesn't matter... as long as the bits come in, I think Tim's approach for the Subscriber side is more than good enough.

 

[thiggins]

You dont need a DSLAM, cant you connect 2 modems together in bridge mode?

No, you can't.

Testing the DSL (or cable) connection requires equipment we don't have and don't plan to get. It's a whole other ball of wax than what we do now.

 

[KenZ71]

I am not nearly as experienced as the posters above so my comments may sound silly.

Conceivably is it possible to hook up Modem A to an incoming cable line and measure for through put and energy use then compare to model B and so on?

I say energy use as well because these run 24 x 7 for years.

 

[System Error Message]

If energy use is a concern than SFP modules are much better but its hard to find SFP modules for cable and DSL.

so what exactly is needed to test a modem other than a DSLAM?

 

[RMerlin]

so what exactly is needed to test a modem other than a DSLAM?

Ain't that enough of a road block already? Because it's a pretty big one.

 

[System Error Message]

DSLAM might be something i could procure but DSL isnt the only modem, theres cable and various fibre optics. I doubt you would need a DSLAM for those.

 

[sfx2000]

Ain't that enough of a road block already? Because it's a pretty big one.

No doubt - there are DSLAM emulators, but they're not cheap... and this would only solve the DSL question, the CM and Fibre issues are other problems to solve.

But Tim said it best - it's not in the shed, there is no plan to bring it into the shed, and I agree with him

It's another level of testing, and that's test plan writing, and staff to do it, along with the capital and operational costs to establish and maintain. And for little benefit, as the carriers manage and control the devices in deployment.

 

[sfx2000]

I'd rather see a review/shootout across the prosumer devices - take MicroTik vs. EdgeRouterPro vs. pfSense (now that pfSense is selling branded HW) vs. a HighEnd AC5300 Consumer Router/AP (RTAC5300/R8500/EA9500/etc) and put them into the Router test plan...

All are close to the same price point @ $500USD or there about...

I think that would be very interesting, and good info for the community at large...

 

[RMerlin]

I'd rather see a review/shootout across the prosumer devices - take MicroTik vs. EdgeRouterPro vs. pfSense (now that pfSense is selling branded HW)

Having a large review/article/shootout about this class of products would certainly be interesting. Personally, I get a feeling that they're partly hype. VLAN stuff for instance is better handled by a good managed switch.

 

[RMerlin]

theres cable and various fibre optics.

Which also requires special, expensive equipment, and engineering expertise that might be outside of Tim's skill set. DOCSIS isn't just about plugging things together, you also have to be familiar with the provisioning aspect of it.

And a lab setup wouldn't take into account the stability of a 100m+ line, with the usual factors involved (cable ageing, noise isolation, etc...)

One final argument: a lot of ISPs will flat out give you no choice at all. You have to use either a very specific model, or can chose among 2 or 3 supported models. So in the end, the customer doesn't have to make any decision, therefore reviews would be mostly meaningless.

 

[sfx2000]

There is a growing movement on the boards to migrate into a more segmented approach - so there's good value, and tossing in a high-end all in one solution (e.g. WRT/RTAC/Nighthack/etc) into the mix would be useful, esp. for some use cases like the small office, hospitality type questions that we see these days.

 

[sfx2000]

Which also requires special, expensive equipment, and engineering expertise that might be outside of Tim's skill set. DOCSIS isn't just about plugging things together, you also have to be familiar with the provisioning aspect of it.

And a lab setup wouldn't take into account the stability of a 100m+ line, with the usual factors involved (cable ageing, noise isolation, etc...)

One final argument: a lot of ISPs will flat out give you no choice at all. You have to use either a very specific model, or can chose among 2 or 3 supported models. So in the end, the customer doesn't have to make any decision, therefore reviews would be mostly meaningless.

the carriers/operators already do this testing - and the OEM's do it as well, even before they submit to the carriers for approval... both industry standard tests as well as carrier specific test cases.

It's not that much different that handsets that have carrier logos - I see a bit where SEM is going with this, but even the industry standard testing requires equipment and skill sets that are out of scope here with SNB, in my honest and humble opinion.