Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 



Forgot your password?
typodupeerror
Get HideMyAss! VPN, PC Mag's Top 10 VPNs of 2016 for 55% off for a Limited Time ×

Comment Re: Tesla's Autopilot is in the "uncanny valley" (Score 1) 440

It seems to be solved quite well on my 2014 Honda:
- qruise control keeps the speed, but lowers it if there is a car in front and keeps a configurale distance by automatically accelerating and decelerating (and braking if needed). It uses radar as opposed to Tesla's camera-only solution which is vulnerable to blinding low sunlight (which seems to be a possible culprit in this case).
- lane guidance steers to keep the car in lane if i switch it on. But if I let go of the wheel, it will soon give an audible and visual warning and soon disable itself.

This for me is unintrusive enough to use on a highway, yet it is strictly "driver assist" not self-driving, so I will not lose focus.

Comment Re: Tesla's Autopilot is in the "uncanny valley" (Score 1) 440

What do you mean "off the map"? There will be no such places.

Also - it is my understanding that most of the selfdriving car development is based on cameras - therefore a selfdriving car can continue on unmapped territories as well as with missing gps just as good as a human driver - using visual cues. If you need to give it instructions how to get to point b, you poke with a finger on a map or an arial photo on the screen and leave the immediate obstacles to the automation.

But all that aside - i mainly meant that the wheel is useless in an emergency situation.

Comment Re: Tesla's Autopilot is in the "uncanny valley" (Score 1) 440

I remember a quote from google selfdriving car team that there is no point in having a steering wheel in a self driving car, because the "driver" would never be able to take over in time to avoid any danger. Which makes sense. We should not switch to automated driving before we are ready to surrender all control.

Comment Re: Printer with public internet ip? why? (Score 1) 390

You should really read up on some networking.

So you're saying forged packets aren't broadcasted in the open? I've seen plenty of firewall rules where the source from = public IP to destination = private IP. That could be internal 192.168.x.x or 10.x.x.x. Obviously, if that private IP subnet doesn't exist, there won't be a route for it.

You are misguided if you think that just setting a private ip as a destination address the packet would reach from the internet to a NATted LAN. The router only routes packets to NATted subnets for which there are NAT entries - either from configuration or dynamic port mappings that usually are generated from NATting outgoing traffic. It would be dropped even without the SPI.

Forged packets? I suppose HP printers will attempt communication via DNS lookup to the outside (because of all those stupid silly feature services for ease of access).

Do you mean that the attacker would forge packets to printer's tcp port 9100 as a forged answer to its outgoing DNS requests? That would again not work, as there is no NAT mapping in the router to printer's address port 9100 as a result of its DNS request unless the DNS request originated from port 9100...

Comment Re: Printer with public internet ip? why? (Score 1) 390

You're welcome.

My point is that SPI is totally irrelevant on accessing NATted private IPs from outside. The router drops the packets from outside not because of SPI, but because it has no instruction what to do with those packets. If you try going around calling your opinions "professional", you should know that.

As far as private ips are concerned - it is quite obvious from the article that the printers had public ips. Universities put them on public ips. That in itself is not a bad practice as they have no shortage of ip addresses like the rest of us. Bad practice is leaving them open to internet.

Also - if you are "of the understanding that you can't pass NAT traffic from outside", how do you propose that they "guessed and "walked" (war dialed) the internal private IP range of 192.168.1.x and attempted port 9100" as you so elegantly put it?

Comment Re: Printer with public internet ip? why? (Score 1) 390

Sysadmins have assigned public ip, because that is how it works in a uni. There is no NAT. NAT is a hack for the puny humans who do not have /16 ip blocks lying around. So assigning an ip from a private ip range would not work at all...

Also it is not at all stupid to assign said public ips with dhcp. On the contrary - it is very smart.

However it was stupid to allow public access from the internet to said ip addresses.

And it is stupid to assume that all networks are like your network at home.

Comment Re: Printer with public internet ip? why? (Score 1) 390

You throw around nice terms like "walk" and "spi" having no clue of network architecture and how ip actually works. Stateful package inspection has no relevance to inability to pass NAT to the internal network. It requires a reverse NAT action from their router to reach inside.

Also you assume that the printers were on private ip range and NATted - which they probably were not.

Comment Re: Printer with public internet ip? why? (Score 1) 390

"Fixed" as opposed to what? I give out all my fixed ip's through dhcp. Its much easier to manage and everything is in a single plaece.

The question here is why the printers are accessible through the internet.

I myself have printed out many hefty manuals on the modern wifi printers that have been left with factory settings with public printing. I never considered that "hacking" though. It is strange what nowadays passes as "hacking".

Comment Re: Probably in the EULA (Score 1) 225

Well. I have tried to fix the error i am having with my business on google maps.

Its an extremely strange one - google maps finds the business at the correct address. The pin is in the right place. But - and here it goes weird - if you click for driving directions, it takes you to a different location (our previous address from 5 years ago). If you zoom out, it is even visible that the route does not end on the pin, but elsewhere. But no one does that and people drive to the wrong address all the time. It is the same on mobile and web google maps.

I have contacted google repeatedly. Google my business support finally bounced it to maps team, and that does not have any support, but only "feedback" (which i have left repeatedly). Meanwhile our customers are directed to a wrong place.

Comment Re: Simple. (Score 1) 265

First of all, it might come as a surprise for you, but not all people live in the US of A. Where i live, "search and frisk" is not something that i know to have been done to anybody that knows anybody i know.

Second of all - you probably get my point despite the maybe-not-so-universal analogy i gave. Maximum you should do, is null-route the portscanning ips automatically. Me - i just ignore them and have done so since the early nineties. If your network security relies on people in the internet NOT portscanning you, you are screwed anyway.

Slashdot Top Deals

"We live, in a very kooky time." -- Herb Blashtfalt

Working...