forensics student project?
Wondered whether this is a good digital forensics student project?
There are a number of statements made at this site.
2] TORFone is fully portable (it can be run from a flash carrier or virtual TrueCrypt-disk and leaves no residue in the system) and works with all versions of Win32 from Windows 98, has very low system requirements (above PI 233 MHz 32M RAM). It saves bandwidth (required from only 2 kbit/s each side using low-bitrates codecs).
Would this still be relevant today given the number of evolved Windows OSs? When it says "leaves no residue in the system" was the registry considered? What about Win98 user.dat hidden file?
4] TORFone provides full confidentiality (by using DH-4096 to match the session key, AES-256-OCB to encrypt the voice traffic and PKDF2+HMAC for autentification). The attacker who mirrors the traffic is unable to listen to the conversation and is not able to decode it later even if he gains access to the computers of participants.
Even if the attacker is listening in at the PC/laptop attack level and not intercepting?
5] Caller and callee are completely anonymous to each other and to outside observers (a call is made on Tor hidden service user).
So the caller could be talking to some in law enforcement and not know it?
I found the mention of my project in this forum. Thank you for interest in my work. But I wanted to clarify some details that are not quite correctly described in the post.
>Would this still be relevant today
This statement is not quite true for modern Windows. The page you link is the old page of the first project created in 2013. After it was an OnionPhone (2015), and then a new Torfone (2019) for which the page is still under construction. Therefore all information on the page http//torfone.org is not relevant and will be updated soon.
>Even if the attacker is listening in at the PC/laptop
The new project takes into account the possibility of such attacks as well as new vulnerabilities at the processor cache level. Therefore Torfone can be use hardware storage of private key and address book like USB/BT token based on the low-cost microcontroller Cortex M1. But the most reliable option involves run Tor and transport software module on a Linux-based minicomputer (for example, RPi) - partially untrusted, while all the cryptography and audio processing is moved to an isolated module on the Сortex M4 in siggle thread. This trusted module completely controls the serial connection with the Linux module. All math (sponge Keccack, curve X25519, etc.) realized on ASM not optimized by compiler and reduced possibility of some EM and PWR side channel attacks.
Thus, it is possible to choose between reliability and user-friendliness of implementation for the user's requirements.
>So the caller could be talking to some in law enforcement
Yes, this may be normal for certain situations when external methods are used to isolate one’s own identity from its pseudonym. But usually anonymity will be only at the first contact and immediately turns into a pseudonymity (the parties exchange their long-term public keys keeping them in address books). If pseudonymity is still needed at the start then it is assumed that the parties are already familiar (they have something common - a common secret), or they can meet later (can check the short fingerprint of their first session key when they exchange contacts). Both authentication options (similar the Socialist-Millionaire protocol in OTR and SAS comparation in ZRTP) are provided. Also, pseudonymity at the start can be provided by mathematical binding the contact (received public key) to the onion-address under the assumption that it is confirmed by some PKI (similar to authentication is TorChat using the Martin Abadi protocol).
New Torfone uses modern cryptography. So, the key exchange is based on the triple-DH protocol as in Signal. But it is improved by adding a SPEKE zero knowledge protocol which allow to protect the identifiers of both participants with the PFS. It also provides deniability and forward denialness in advance. The full (online) deniability (for example like in OTR V4 using ring signatures) is ambiguous so we chose the KCI resistance instead.
You can test pre-alpha Android app for talking over Tor
(app in active development now so not subscribed by my PGP yet)
This is another my project that has nothing to do with Tofone. This is digital encryptor of analogue voice for reliable p2p protection between two HF/VHF radios or mobile phones (the latter requires certain phone settings and not all models are suitable).
The goal of this project is a specially designed modem with carrier passes AMR codecs without significant distortions. And also the well-known actual NATO standard - MELPE1200 codec is partially adapted to work in real time on low-cost Cortex M4 processors using their DSP capabilities.
Project page http//torfone.org/jackpair/
(all executable files and source codes for Windows and embedded platforms are available for download in the menu at the top of the page and are signed).
Very similar China commercial project
Thanks once more for your interest in my projects. I specifically registered on this forum because the opinion of experts in this field is important to me. I will be glad to hear criticism and discuss in detail my implementation.