4 ms·

The amazing thing is realizing that despite the increasing dangers and actual disasters involved, more and more things are going to be put on the Internet. The

The amazing thing is realizing that despite the increasing dangers and actual disasters involved, more and more things are going to be put on the Internet. The equation everywhere is "the cost of the security is always too high because the failure of security is always an unusual situation and something that usually works and is cheaper will win in the marketplace." Have we gone backwards. Yes, expect more of this.

faeriechangling 6y ago
What's described in the article is not a security problem. It's an availability problem. I would argue consumers DO care about the availability and I see lots of cloud based systems with local fallbacks. When IKEA introduced cloud devices, IKEA hardly a company known for high prices or using expensive stuff in their products, they had local fallbacks. Their product is competing with the reliability of less expensive devices controlled with a light switch. Locks are another case where if you reinvent the wheel and get significantly less reliability people will be mad.
- PeterisP 6y ago
  Availability is 1/3 of what we traditionally define as security (Confidentiality, Integrity, & Availability), so it definitely is a security problem.
  - brendoelfrendo 6y ago
    Yes, business continuation and disaster recovery is often considered a security discipline for this exact reason. Not every disaster is a hurricane or an earthquake; there are plenty of man-made threats to availability.
  - jcims 6y ago
    I don’t know if it’s helping but I’ve started to blend in the term safety to try to bridge that cognitive rut of security = hack avoidance.
siculars 6y ago
This is more or less exactly the scenario that played out in Texas this week, Feb 14-19 2021. All the players (gov, regulators, power generators) decided that it was too expensive to weatherize the network for a 0.1% chance event. Yet... here we are. Entirely predictable.
- kilotaras 6y ago
  > weatherize the network for a 0.1% chance event. The last one such event (also with blackouts) happened in 2011 and was generally called "1 in 10 years cold snap."[0] Not weatherizing for 0.1% chance can be acceptable, not doing for 10% is not. [0] https://www.nytimes.com/2011/02/05/us/05storm.html https://www.nytimes.com/2011/02/05/us/05storm.html
  - Ma8ee 6y ago
    It’s how capitalism works nowadays: I win or you lose. As long as everything is going well, like normal weather or excessive growth of house prices, the capital owners make a lot of money. But when something predictable goes wrong, the bill is left to the consumers and tax payers.

faeriechangling 6y ago

What's described in the article is not a security problem. It's an availability problem. I would argue consumers DO care about the availability and I see lots of cloud based systems with local fallbacks. When IKEA introduced cloud devices, IKEA hardly a company known for high prices or using expensive stuff in their products, they had local fallbacks. Their product is competing with the reliability of less expensive devices controlled with a light switch. Locks are another case where if you reinvent the wheel and get significantly less reliability people will be mad.

PeterisP 6y ago
Availability is 1/3 of what we traditionally define as security (Confidentiality, Integrity, & Availability), so it definitely is a security problem.
- brendoelfrendo 6y ago
  Yes, business continuation and disaster recovery is often considered a security discipline for this exact reason. Not every disaster is a hurricane or an earthquake; there are plenty of man-made threats to availability.
- jcims 6y ago
  I don’t know if it’s helping but I’ve started to blend in the term safety to try to bridge that cognitive rut of security = hack avoidance.

siculars 6y ago

This is more or less exactly the scenario that played out in Texas this week, Feb 14-19 2021. All the players (gov, regulators, power generators) decided that it was too expensive to weatherize the network for a 0.1% chance event. Yet... here we are. Entirely predictable.

kilotaras 6y ago
> weatherize the network for a 0.1% chance event. The last one such event (also with blackouts) happened in 2011 and was generally called "1 in 10 years cold snap."[0] Not weatherizing for 0.1% chance can be acceptable, not doing for 10% is not. [0] https://www.nytimes.com/2011/02/05/us/05storm.html https://www.nytimes.com/2011/02/05/us/05storm.html
- Ma8ee 6y ago
  It’s how capitalism works nowadays: I win or you lose. As long as everything is going well, like normal weather or excessive growth of house prices, the capital owners make a lot of money. But when something predictable goes wrong, the bill is left to the consumers and tax payers.