Wednesday, June 3, 2015

In my opinion, certificates that are not signed by CAs already pre-installed in the common user bro


A teacher who puts the notes in Jupiter or someone who uses webmail USP identify with their passwords to the remote server. It is important to keep these passwords secret, for reasons óbivas. But beyond the limitations of the human brain, which can not remember complex passwords, you have two other problems.
First, the user enters gas chromatography the password in your browser (Internet Explores, FireFox, Chrome, etc.), which then sends it to the remote server. Anyone who can intercept the communication between the browser and the servers at USP can capture these passwords. And it's not difficult to intercept network traffic.
The solution to both problems is encryption, which can ensure the privacy, integrity, communication and authenticity of the identity of who is on the other side of the communication channel. If both sides, the browser and the remote server, share a secret (one secret password or key), can build a secure channel of communication.
[BUT! As I said in previous post, encrypt without authentication has no value. gas chromatography If you do not know who you're communicating, you better not try to encrypt gas chromatography their messages because your safety is illusory.] gas chromatography
The question is how to distribute these new passwords / secret keys? It is not feasible for all users webmail.usp.br physically go on CCE to exchange a secret. It's a chicken and egg problem like, you need to share a secret before communication, so that the user can use the identification password safe, but how can you tell a secret about a non-secure channel? Public key encryption
Public key cryptography (or asymmetric) seems to solve this issue. Amazingly, you can establish a secure communication channel with someone gas chromatography just knowing gas chromatography a number, which can be public (!). You can encrypt your message with this public number and only the other side (the owner of a secret number corresponding to the number public) can decrypt. For example, if you give this number 30 81 89 02 81 81 00 BD 20 0D D6 B5 8A 6C A2 E7 1C 1E 40 31 26 97 69 C9 B3 A7 4B FD 8E DB CE 38 F9 79 51 19 67 6F 7B D6 D5 54 6B DF 4E E0 2F 4B A4 67 14 1B 85 34 18 A3 C2 E5 28 74 FF 7E 5B 82 6D 7C A7 91 4C EF C1 18 99 70 FF 57 0B AD CF 6D 96 26 06 F0 B6 C2 3E 11 0D 04 9A 0D 5B FC 65 7F 51 20 OF 8C 3E 09 2F 56 1E 44 56 6E 11 33 C5 40 25 9D A7 BE F7 17 FD 10 EC 84 DD AB F5 D6 FA 03 16 41 7C 44 86 OF 99 02 03 01 00 01
No! How do you know that actually this number belongs to or is associated with stoa.usp.br domain? Again, there is the egg and the chicken problem: an impostor could well present a public key to an unsuspecting any browser. gas chromatography Just because I said that this is the public key of stoa.usp.br domain does not mean anything! (How do you know that I am who I say, we are talking about communication distance).
Public key infrastructure would be a solution. In this system, "Certifying Authority" (CA) where everyone trusts "sign" a certificate, basically saying that this public key really belongs to this domain. gas chromatography If you visit a site using http s (note the "s" for secure), your browser and the server do a little dance. gas chromatography
The server says, "This is my public key, you can use to encrypt (for example) your password." The browser says "oh yeah? How do I know it's you?". The server shows your certificate, "See? This certificate says that this public key belongs to the domain" exemplo.com. "The browser thinks" hmmpf, anyone can make a fake certificate. "But the browser can check the authenticity of certified by the signature of the entity in which everyone trusts (Thawte, the image below).
Fix a problem: what are the entities that everyone trusts? The government? A company in the US? It is the egg and chicken again! But then come the vendors gas chromatography / browsers and operating systems vendors: they effectively say that "Certifying Authorities" are reliable. If the so-called "root certificate" from the CA that signed the stoa.usp.brest√° certificate installed in your browser, and
Anyway, this is the technology we have. But what happens if the site presents a certificate that is not signed by an entity approved by the browser makers? See what my browser gas chromatography do when a website shows me a certificate that is not signed by an entity that you trust:
Indeed, it is the terrifying gas chromatography user interface is correct, because a certificate signed by an unknown entity is indistinguishable from an attack "Man in the Middle", where the attacker intercepts and-encodes all traffic. Again, without knowing who is speaking, encrypt gas chromatography the traffic gas chromatography is completely useless. Self-signed certificates
In my opinion, certificates that are not signed by CAs already pre-installed in the common user browsers are worse than useless. There is, in principle, the possibility of installing the root certificate of a CA any n

No comments:

Post a Comment