[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[taler-anastasis] 03/03: related work
From: |
gnunet |
Subject: |
[taler-anastasis] 03/03: related work |
Date: |
Sun, 31 May 2020 16:10:18 +0200 |
This is an automated email from the git hooks/post-receive script.
dennis-neufeld pushed a commit to branch master
in repository anastasis.
commit 77277113fbc4c329b609518a2ee8bee52080ee5b
Author: Dennis Neufeld <dennis.neufeld@students.bfh.ch>
AuthorDate: Sun May 31 14:09:18 2020 +0000
related work
---
doc/thesis/bibliothek.bib | 20 +++++++++++++++++
doc/thesis/related_work.tex | 54 ++++++++++++++++++++++++++++++++++++---------
2 files changed, 64 insertions(+), 10 deletions(-)
diff --git a/doc/thesis/bibliothek.bib b/doc/thesis/bibliothek.bib
index 8fe49d2..f4a3200 100644
--- a/doc/thesis/bibliothek.bib
+++ b/doc/thesis/bibliothek.bib
@@ -179,4 +179,24 @@
year={2016},
organization={IEEE}
}
+@book{trimberger2012,
+ title={Field-programmable gate array technology},
+ author={Trimberger, Stephen M},
+ year={2012},
+ publisher={Springer Science \& Business Media}
+}
+@misc{madurawe2006,
+ title={Alterable application specific integrated circuit (ASIC)},
+ author={Madurawe, Raminda Udaya},
+ year={2006},
+ month=jun # "~20",
+ publisher={Google Patents},
+ note={US Patent 7,064,579}
+}
+@article{stamp2003,
+ title={Once upon a time-memory tradeoff},
+ author={Stamp, Mark},
+ journal={San Jose State University, Department of Computer Science},
+ year={2003}
+}
diff --git a/doc/thesis/related_work.tex b/doc/thesis/related_work.tex
index 71b5aae..ab9ece5 100644
--- a/doc/thesis/related_work.tex
+++ b/doc/thesis/related_work.tex
@@ -19,28 +19,62 @@ In Anastasis we use SHA-512 for hashing data.
\subsubsection{HMAC}
When it comes to integrity of messages during communication of two parties
over an insecure channel Keyed-Hash Message Authentication Codes (HMAC) are
used as check values. An HMAC function is based on a hash function and takes
two arguments, a key K and a message M:
-HMAC\textsubscript{K}(M) = H(K $\oplus$ opad,H(K $\oplus$ ipad, M)) with
"ipad" and "opad" being constants which fill up the key K to the blocksize of
the hash function \cite{BCK1996}. The blocksize of a modern hash function like
SHA-512 is 64 Byte.
+HMAC\textsubscript{K}(M) = H(K $\oplus$ opad,H(K $\oplus$ ipad, M)) with
"ipad" and "opad" being constants which fill up the key K to the blocksize of
the hash function \cite{BCK1996}. The blocksize of a modern hash function like
SHA-512 is 64 Byte.\\
+In Anastasis we use HMACs to achieve verifiability.
\subsubsection{HKDF}
-While HMAC functions are at least not used directly in Anastasis, we use
so-called HKDFs. A HKDF is a key derivation function (KDF) based on a HMAC. A
KDF "is a basic and essential component of crypto-
-graphic systems: Its goal is to take a source of initial keying material,
usually containing some good amount of randomness, but not distributed
uniformly or for which an attacker has some partial knowledge, and derive from
it one or more cryptographically strong secret keys" \cite{krawczyk2010}.
+A HKDF is a key derivation function (KDF) based on a HMAC. A KDF "is a basic
and essential component of crypto-
+graphic systems: Its goal is to take a source of initial keying material,
usually containing some good amount of randomness, but not distributed
uniformly or for which an attacker has some partial knowledge, and derive from
it one or more cryptographically strong secret keys" \cite{krawczyk2010}.\\
+Anastasis uses HKDFs to derive symmetric keys for encryption purposes.
\subsubsection{Argon2}
-Hash functions like SHA-512 are very fast to compute. Therefor passwords
stored in a hashed form are vulnerable to dictionary attacks with new hardware
architectures like FPGAs and dedicated ASIC modules. But those architectures
"experience difficulties when operating on large amount of memory"
\cite{BDK2016}.
+Hash functions like SHA-512 are very fast to compute. Therefor passwords
stored in a hashed form are vulnerable to dictionary attacks with new hardware
architectures like FPGAs \cite{trimberger2012} and dedicated ASIC
\cite{madurawe2006} modules. But those architectures "experience difficulties
when operating on large amount of memory" \cite{BDK2016}.\\
+Argon2 is a memory-hard function that won the Password Hashing Competition in
2015. It minimizes time-memory tradeoff \cite{stamp2003} and thus maximizes the
costs to implement an ASIC for given CPU computing time \cite{BDK2016}. Aside
from the fact that Argon2 makes dictionary attacks much harder, we use Argon2
for another feature too: Memory-hard schemes like Argon2 are very useful for
key derivation from low-entropy sources \cite{BDK2016}.\\
+Argon2 is used in Anastasis to derive an identifier for the user from some
low-entropy material.
+
\subsection{Secret sharing}
Secret splitting, also known as secret sharing, is a technique for
distributing a secret amongst multiple recipients. This is achieved by
assigning a share of the secret to each recipient. By combining a sufficient
number of those shares, it is possible to reconstruct the secret.
-In a secret sharing theme the recipients of a share often are called
\textit{players}. The figure who gives a share of the secret to the players is
called \textit{dealer}.\\
+In a secret sharing theme the recipients of a share often are called
\textit{players}. The figure who gives a share of the secret to the players is
called \textit{dealer}.
-Regarding secret sharing there are several interesting approaches. For
example, the algorithm "Shamir's Secret Sharing" „divide[s] data D into n
pieces in such a way that D is easily reconstruct able from any k pieces, but
even complete knowledge of k - 1 pieces reveals absolutely no information about
D“ \cite{shamir_sharing}.
+\subsubsection{Shamir's Secret Sharing}
+The algorithm "Shamir's Secret Sharing" is one of the most well known secret
sharing scheme. It „divide[s] data D into n pieces in such a way that D is
easily reconstructible from any k pieces, but even complete knowledge of k - 1
pieces reveals absolutely no information about D“ \cite{shamir_sharing}.\\
Shamir’s simple secret sharing scheme has two key limitations. First, it
requires a trusted dealer who initially generates the secret to be distributed,
and second the shares are not verifiable during reconstruction. Therefore,
malicious shareholders could submit corrupt shares to prevent the system from
reconstructing the secret -- without these corrupt shareholders being
detectable as malicious. Furthermore, the dealer distributing the shares could
be corrupt and distribute some incons [...]
+
+\subsubsection{Verifiable Secret Sharing}
Verifiability can be achieved by using so called commitment schemes like the
Pederson commitment. It allows „to distribute a secret to n persons such that
each person can verify that he has received correct information about the
secret without talking with other persons“ \cite{pedersen_sharing_0}. In his
paper „A Practical Scheme for Non-interactive Verifiable Secret Sharing“, Paul
Feldman combines the two algorithms above. His algorithm for verifiable secret
sharing, short VSS, allows [...]
-Distributed key generation algorithms, short DKG, solve the problem of needing
a trustworthy dealer by relying on a threshold of honest persons. Contrary to
the above-mentioned schemes, in distributed key generation algorithms every
participant is involved in key generation.
-The Pederson DKG is such „a secret sharing scheme without a mutually trusted
authority“ \cite{pedersen_sharing_5.2}. Basically, this DKG works as follows:
First, each involved party generates a pre-secret and distributes it to all
parties using the verifiable secret sharing scheme of Feldman. Afterwards, each
party recombines the received shares, including its own pre-secret, to a share
of the main secret. The main secret can be reconstructed by summing up each
recombination of the share [...]
+
+\subsubsection{Distributed Key Generation}
+Distributed key generation algorithms, short DKG, solve the problem of needing
a trustworthy dealer by relying on a threshold of honest persons. Contrary to
the above-mentioned schemes, in distributed key generation algorithms every
participant is involved in key generation.\\
+The Pederson DKG is such „a secret sharing scheme without a mutually trusted
authority“ \cite{pedersen_sharing_5.2}. Basically, this DKG works as follows:
First, each involved party generates a pre-secret and distributes it to all
parties using the verifiable secret sharing scheme of Feldman. Afterwards, each
party recombines the received shares, including its own pre-secret, to a share
of the main secret. The main secret can be reconstructed by summing up each
recombination of the share [...]
+
+\subsubsection{MIDATA}
+MIDATA is a project that aims to give patients back control over their medical
data and to enable them to share their data only with those they trust. In case
the patient lost his device running the MIDATA application and his
MIDATA-password, MIDATA build in a key recovery system using the Shamir Secret
Sharing Scheme mentioned above. In their case a few "persons working at MIDATA
have generated a public-private key pair (Recovery key) on their own computer.
They keep the private recover [...]
+In our opinion the security of MIDATA is broken in two ways:
+\begin{enumerate}
+ \item The password is constructed at the server, not at the patients
device. An administrator of the server can read the recovered password.
+ \item It is not clear which channel the persons working for MIDATA use
for their decisions and activities regarding the key recovery. The channel
could be vulnerable. For example, an attacker could illegitimately trigger a
recovery process via e-mail if it is the chosen channel.
+\end{enumerate}
+
+
+\subsubsection{Key sharing in Anastasis}
For Anastasis we do not need a DKG because the dealer is the user himself and
therefore, he is fully trustworthy. But we need verifiability. In our case we
achieve verifiability by using HMACs. Furthermore, for our purposes the
above-mentioned algorithms are inadequate because we are dealing with a
manageable number of sharing parties and we need a more flexible solution.
+
\subsection{Authentication}
-Anastasis is using standard authentication procedures to authorize its users.
There are several authentication methods available, a short overview of the
methods is presented here. Password authentication is the most widely used
authentication procedure. But as studies show the procedure has its problems
\cite{authentication_methods_review}. The handling of the passwords is done
poorly, like storage or transmission. Additionally, the user must remember his
password, therefore the passwor [...]
-To build a secure authentication procedure, today multifactor authentication
is the standard \cite{multifactor_authentication}. Multifactor authentication
combines multiple authentication procedures, to enhance the security of the
system. For Anastasis we are building a multifactor authentication system,
which combines a wide range of authentication methods to provide authenticity.
+Anastasis is using standard authentication procedures to authorize its users.
There are several authentication methods available, a short overview of the
methods is presented here.
+
+\subsubsection{Password authentication}
+Password authentication is the most widely used authentication procedure. But
as studies show the procedure has its problems
\cite{authentication_methods_review}. The handling of the passwords is done
poorly, like storage or transmission. Additionally, the user must remember his
password, therefore the password is limited to the capabilities of the user.
+
+\subsubsection{SMS authentication}
+Another way to authenticate is SMS authentication. The most popular use case
is the mobile TAN used to authorize online banking transactions. But SMS is no
longer considered secure. The SMS authentication relies on the security of the
mobile network, which has different possible attacks \cite{rieck_detection}.
There are also specialized mobile Trojans which are used to eavesdrop these
messages.\\
+Instead of using SMS one can also use other forms of messages such as email or
physical mail. They all face the threat of interception.
+
+\subsubsection{Biometric authentication}
+Another way of authenticating is the biometric approach \cite{biometric_auth}.
Biometric authentication is based on "something you are", like your iris or
your fingerprint. There are also threats against biometric authentication.
There are documented attacks against fingerprint and iris scan authentication.
For example, a member of the German CCC e.V. was able to generate replicas from
Angela Merkel's iris and Ursula von der Leyen's fingerprint \cite{ccc_merkel}.
+
+\subsubsection{Multi-factor authentication}
+To build a secure authentication procedure, today multi-factor authentication
is the standard \cite{multifactor_authentication}. Multi-factor authentication
combines multiple authentication procedures, to enhance the security of the
system. For Anastasis we are building a multi-factor authentication system,
which combines a wide range of authentication methods to provide authenticity.
--
To stop receiving notification emails like this one, please contact
gnunet@gnunet.org.