Issue Of Internet Privacy And Open-Public Wifi

As everyone's been saying, the conventional wisdom is that open/public (o/p) WiFi is terribly dangerous and should be avoided at all costs because anyone can just intercept your packets and read them, whereas with WPA2, the packets are encrypted. For this reason, it is often recommended that open/public (o/p) WiFi be used with VPN to encrypt WiFi (FTC, 2014; Kaspersky, 2013; Norton, 2015) . But something's been really bothering me about this reasoning - what happened to HTTPS/TLS??? Doesn't TLS encrypt packets anyway? So who needs a VPN or even WPA2 for that matter?

Altho I hesitate to disagree with the majority, (after spending way too much time on this), I've come to the conclusion that the paranoia surrounding o/p WiFi is no longer justified. It seems to me that the primary concern with o/p WiFi is due to none-HTTPS sites that don't use TLS. But this is 2020, can anyone name a single none-HTTPS site that deals with anything remotely sensitive? (If there are still any sites like that, you shouldn't be using it anyway, even with Ethernet and VPN.) I think this paranoia comes from the past decade before HTTP/TLS was so widespread, and the security community made an effort to create awareness abought the dangers of o/p WiFi. As a result, now even tho it's not much of an issue anymore, there remains this taboo/fear of o/p WiFi.

I'm not saying that o/p WiFi is totally secure, nothing is; and of course, it is less secure than cellular connection or WPA2. But the truth is that WPA2 is less secure than Ethernet, so should we not use WIFI at all anymore? Just because there are more secure options available does not mean they are necessary. In general, security comes at a cost - money, time, convenience, flexibility, etc. each case requires it's own cost-benefit analysis to determine what level of protection is best. However, such analysis is not possible unless we have the facts about what the risk is. So, I'll go through some additional reasons/excuses given for why o/p WiFi is terrible and should only be used with VPN, and comment on each one.

This isn't really an o/p WiFi issue. The most straightforward anecdote to shoulder surfing is situational awareness. Use a password manager so that you don't have to type stuff in. Also, if you think someone might be looking, you can cover the screen & keyboard while putting in your password (on a laptop just close it most of the way, on a desktop, stand over it so that your body provides cover. Just use common sense.

Another issue that's been bothering me is why open WiFi can't be encrypted just like private WiFi? In other words, why should encryption be dependent on the router being password protected (which is for authentication)? The reason seems to be that the WiFi password is used not only for authentication but also serves as the pre-shared key (PSK) to generate the encryption key (so naturally, if there's no password, there can't be encryption). For this reason, the problem with passwordless open WiFi also applies to public WiFi (where many people have the password), since everyone with the password has the PSK. 

However, a significant difference is that with public WiFi (with a known password) the attacker not only needs the WiFi password for the PSK, but also needs to be present for the initial 4-way handshake between client and router (in order to get the nonces which are combined with PSK and MAC to get the session key). (I think this may be (one of) the reasons of 'guest mode,' which enables other users to be authenticated with a different password, thereby enabling the owner to allow others to use the WiFi without having to give up his PSK.) (I should note that all this applies to WPA2-PSK; however, WPA2-Enterprise uses 802.1x.)

But all of these issues beg the question - why make the WiFi password the encryption key? Why not just use asymmetric public-key encryption with a Diffie–Hellman key exchange (the public key can be the SSID)? This would enable both public and open WiFi to be secure (as well as obviate the need for guest modes with multiple passwords). The fact that this is not done may be further evidence that the whole hysteria about open/public WiFi is overblown, since if it was such a big deal, they could just use an asymmetric encryption standard with DHKE.

However, there are a couple of issues with this: even if o/p WiFi isn't that bad, if it can be made more secure, why not do it - what's the downside? Additionally, even I agree that o/p WiFi was a bigger issue before HTTPS became so widespread, so why wasn't this method used then? I think the answer to both questions is that it wasn't & isn't worth the increase in overhead. Even tho overhead is becoming less of an issue (due to advancing technology), this is balanced out by the decreased risk associated with o/p WiFi (due to the spread of HTTPS).

The WiFi alliance does have something called Passpoint, which authenticates without a password (Seltzer, 2015; WiFi Alliance, 2019), and seems to provide unique encryption keys for each user. However, it seems to require PKI for authentication. I'm not sure if it would work without a PKI on an open WiFi that doesn't require user authentication, or an open WiFi using just password authentication.

In conclusion, while o/p WiFi is less secure than WPA2, for the average user (as opposed to those who need to be concerned about targeted attacks, such as organizations and those targeted by governments), o/p WiFi is not that big a deal anymore. We should be focusing more on the major threats of this decade (phishing, ransomware, etc.) rather than those of the past.