Live data from Hacker News

‘War Dialing’ tool exposes Zoom’s password problems

krebsonsecurity.com

31–40 of 247 posts

Re: ‘War Dialing’ tool exposes Zoom’s password problems

#31

How hard is this for them to fix?

Generating a random 6-digit passcode for each meeting by default? Not hard at all.

Rate-limiting incorrect password attempts could take a bit longer to implement, but still not a particularly difficult problem to solve.

Re: ‘War Dialing’ tool exposes Zoom’s password problems

#32
It's 1985 all over again: I'm in my bedroom running a ProDOS wardialer on my 300/1200 baud AppleModem; I have found zero computers, but it is fun watching the numbers flick past, hoping that I, too, can discover a WOPR and start global thermonuclear war.

Re: ‘War Dialing’ tool exposes Zoom’s password problems

#33
post #10

Earlier quoted context omitted.

Yeah but then it sucks for people calling in to have to punch in a 21+ character long meeting ID

> Yeah but then it sucks for people calling in to have to punch in a 21+ character long meeting ID I may be out of touch with the average biz-guy, but how many people are realistically calling in manually, over traditional phone-lines these days? Is it really a significant percentage?

Yes. If you're in a conference room and it's not a Zoom Room(tm), or has a Cisco system, or whatever, you have to use the conference phone. You might be able to tell the Zoom meeting host to call the conference phone and bring it into the meeting, but it'd be easier just to type the ID in (unless it was really 128 char, but then that'd give people a reason to buy new conference hardware I guess).

Also if you don't want to install the Zoom client, you can just dial in from your cell phone or desk phone.

Re: ‘War Dialing’ tool exposes Zoom’s password problems

#35
post #5

Not a good idea to use 9 to 11 digit long IDs with no password requirement by default; they should have used at least 128-bit random ids, i.e. 21 character long base64-encoded strings.

This is likely to support dial-in over the telephone network. I think "no password" is the bigger issue, because repeated attempts with incorrect passwords can be rate-limited. Zoom should be generating a random 6-digit password for each meeting by default. There may be use cases for not having any password, but that should be explicitly opt-in and have a warning message to every participant that anyone can join and…

i'm sure the reason for that is the UX. the zoom had a reputation of "just works" and part of it was that is so easy to jump in to a meeting. if now i have to manage access and so on, it would not be "it just works" like it was

Re: ‘War Dialing’ tool exposes Zoom’s password problems

#36

It's 1985 all over again: I'm in my bedroom running a ProDOS wardialer on my 300/1200 baud AppleModem; I have found zero computers, but it is fun watching the numbers flick past, hoping that I, too, can discover a WOPR and start global thermonuclear war.

Yeah but this time, it’s an easy as guessing a world leaders zoom meeting, and tricking them into believing something preposterous

Re: ‘War Dialing’ tool exposes Zoom’s password problems

#37
TLDR: With 17 digits meeting password is not needed at all. If meeting will be 17 numbers it will be the same as to protect 11 length digit number with 6 digit password. So basically that's the trade off. One could say that password is not the same as meeting ID, but usually they both sent in one email/message and lifetime and protection for them is equal. Also it's easier to input one number than 2 different.

Re: ‘War Dialing’ tool exposes Zoom’s password problems

#38
post #20
post #5

Not a good idea to use 9 to 11 digit long IDs with no password requirement by default; they should have used at least 128-bit random ids, i.e. 21 character long base64-encoded strings.

Or my personal favorite for anything you show to a user: https://www.crockford.com/base32.html

Good idea for this! But trickier for phone calling into a conversation. They could also just add 3 digits and a slight delay in their connection API, making it much harder to brute force, albeit only by a constant factor.

Re: ‘War Dialing’ tool exposes Zoom’s password problems

#39
post #14
post #5

Not a good idea to use 9 to 11 digit long IDs with no password requirement by default; they should have used at least 128-bit random ids, i.e. 21 character long base64-encoded strings.

It's an incredibly simple thing to screw up. I wonder where else they use low entropy random strings. I wonder if their password reset functionality can be brute forced too. Another problem is where they put rate limiting as it seems probable based on this article there are holes.

Every time I read about a Zoom "screwup" I see a feature that's UX centric. It's pretty cool tbh.

Re: ‘War Dialing’ tool exposes Zoom’s password problems

#40
post #5

Not a good idea to use 9 to 11 digit long IDs with no password requirement by default; they should have used at least 128-bit random ids, i.e. 21 character long base64-encoded strings.

This is likely to support dial-in over the telephone network. I think "no password" is the bigger issue, because repeated attempts with incorrect passwords can be rate-limited. Zoom should be generating a random 6-digit password for each meeting by default. There may be use cases for not having any password, but that should be explicitly opt-in and have a warning message to every participant that anyone can join and…

In this context, a password is the same as an id.

There's never a reason to share the id without the password.

Post reply on HN