English man spends 11 hours trying to make cup of tea with Wi-Fi kettle

[Phillip Hallam-Baker <[e-mail]>]

This might sound amusing. If people hope to make a living off selling the
public IoT devices, it isn't.

https://www.theguardian.com/technology/2016/oct/12/english-man-spends-11-hours-trying-to-make-cup-of-tea-with-wi-fi-kettle

The problem here is one of systems integration. There are too many moving
parts and when the system goes wrong, there is insufficient information to
diagnose the cause or even which device is at fault.

For the past few months I have had a pattern of Internet outages that would
only end when the router was rebooted. This was eventually solved when the
hardware was replaced. But there is no way for me to know if the reason the
device was going out was a hardware failure or the device had been
compromised and was participating in a botnet - as reports claim many
devices supplied by my ISP were.

Replacing it, I seem to still have a bufferbloat issue but I have no way to
solve that till the Turris Omnia arrives and even then I will probably have
to do some fiddling.

If this type of issue doesn't get fixed we are not going to be able to sell
consumers the stuff we want to sell them. The lack of diagnostic and
debugging support is increasing costs for everyone. When a device goes out
there are typically three vendors who might be the cause. I count myself
lucky if I get the right one the second time.

Yes we have SNMP but that is firmly established as being an 'enterprise'
capability that consumers won't be told even exists.

Right now we seem to have more industry IoT interest groups than ever. But
what we don't seem to have is a group whose brand is going to mean 'it will
work, it is secure and there are no hidden vendor lockins'.