Broadband Router Comparisons

Mike mike-nanog at
Mon Dec 28 01:56:02 UTC 2015

On 12/27/15, 4:57 PM, Larry Sheldon wrote:
> On 12/26/2015 23:49, Mike wrote:
> [snip]
>> Firstly, they are all junk. Every last one of them. Period. Broadband
>> routers are designed to be cheap and to appeal to people who don't know
>> any better, and who respond well (eg: make purchasing decisions) based
>> on the shape of the plastic, the color scheme employed, and number of
>> mysterious blinking lights that convey 'something important is
>> happening'. Further, the price point is $45 - $70 thereabouts, putting
>> some definite constraints on the actual quality of the engineering and
>> components that go into them. I feel that we, the service provider,
>> endure a significantly high and undue burden of cost associated with
>> providing ongoing support to customers as a result of the defects
>> contained therein.
> Why don't you offer an acceptable (to you) device at a price 
> acceptable to me as a part of the service.  I'd buy it.
NO SUCH DEVICE EXISTS, because you can't afford it. If I were to take 
you seriously however - and we're talking about eliminating all excuses 
and simply getting down to it and making a marginally qualified showing 
at expecting uninterrupted service - the entire environment is what has 
to be solved. The device would be cisco or juniper branded, internal 
redundancy / failover features to allow hitless upgrades or module 
failures, have dual (preferably, triple) power supplies, would be 
required to be housed in a locked enclosure with air conditioning and 
online double conversion battery with the addition of an external backup 
generator with its own separate backup fuel supply, which is further 
tested weekly and mantained with inspections and oil changes. The router 
would be under service contract with the manufacturer, would be 
monitoring by my noc, and would receive appropriate software upgrades as 
required, and you would pay for this monthly in addition to your 
internet service. Furthermore, you also would be required to have at 
least two distinct connections to me and make a deposit to provide 
credit in the event you falsely claim 'trouble' where no trouble exists. 
A seperate 'test pc', also in it's own enclosure and normally offlimits 
to you, and connected to said router and backup power and such, would be 
agreed upon as the test fixture that we would monitor TO. It would 
display current network statistics including packet loss and latencies 
to various on and off-net locations, with current time and date logging 
on screen. You would agree that you are to blame each and every time you 
'can't get on', while the test pc clearly shows on it's local screen to 
you otherwise. You would be required to forfeit a portion of your 
deposit each time you called for technical support and were determined 
to be at fault and to blame for your own issue.

More information about the NANOG mailing list