Reputation
Next tag badge:
97/100 score
25/20 answers
Badges
14 226 353
Newest
 Nice Answer
Impact
~4.3m people reached

Apr
14
comment Can a SSL certificate have longer validity period than its parent in X.509 chain?
@cwyang: each pair may be valid only at dates that fall within the validity ranges of all involved certificates, so yes, some overlapping is needed. Considerable overlapping is common upon renewals, to give time for smooth transitions.
Apr
14
comment Can a SSL certificate have longer validity period than its parent in X.509 chain?
@MikeOunsworth: yes -- that's what Peter Gutmann calls "pointwise checking".
Apr
13
comment How can you spoof a TCP connection to a device that uses the same initial sequence number for each run of a TCP handshake?
If the server is willing to accept a TCP connection on a given port, then it will always send a SYN+ACK in response to a SYN, because that's exactly what it means to "accept a TCP connection". If the server does not listen to that port, then it won't respond with a SYN+ACK, but then the question is meaningless because without a connection, then there is no notion of "sequence number".
Mar
22
comment Why do PGP master keys only have a single subkey, and tie certification with signing by default?
Normally, emails signed with PGP include some stuff that helps the recipient find out about the sender's key, either a copy of the key itself, or at least a reference to a key server. Recipient should still validate the key (with the WoT, or a phone-call-and-dictate-hash-value) but at least in some cases the signature public key itself is copied in the email.
Mar
16
comment Why does Google prefer ECDHE-RSA-AES128-GCM-SHA256?
@MichaelKjörling: note also that there is a difference between "there is a working quantum computer" and "there is a working quantum computer that works at the gigahertz-type of frequency that we now expect from classical computers". There are three successive challenges for QC: assembling enough qubits, preventing decoherence for sufficient time to run the attack, performing enough quantum operations per second for the attack to succeed. Equating quantum-AES-256 with classic-AES-128 means assuming that the 2nd and 3rd challenges are trivial, which seems a bit optimistic to me.
Mar
16
comment How, if at all, are some programming languages inherently less safe?
Proponents of static (respectively dynamic) types always present their favourite language feature as the best way to achieve "safety" but it is not really substantiated in actual practice -- developer's discipline is much more important. But note that not all developers are equivalent; some will be more at ease with static type-checking, while others will be more efficient with dynamic type-checking. In general, the best language a developer may choose is the one he masters, and that won't be the same language for every developer.
Mar
11
comment Secure Configuration of Ciphers/MACs/Kex available in SSH
@Mecki: I would contend that speed matters only up to what the underlying network can absorb. Even with CBC, a normal PC can do AES encryption faster than what can be sent over gigabit ethernet -- especially since all modern x86 systems have the AES-NI opcodes. What CTR parallelism gains is better performance for constant-time implementations (not table-based implementations) on platforms that have large integer registers but do not have dedicated AES hardware -- there we are talking about smallish ARM systems, not PC.
Mar
10
comment Is there a PRNG in Java with a period of at least 256 bits?
@TrevorBernard: Cryptographic security relies on unpredictability. It does not matter whether all shuffles are mathematically reachable, only that outsiders (attackers) cannot differentiate between a "true" random shuffle and the one you make with your PRNG.
Mar
10
comment Is C a good choice for security-related software any longer?
@Adjit: it depends on the context, but, as a starting point, C# is not bad. Thanks to Mono, code written in C# can run on Windows, OS X and Linux. It has automatic memory management, checks array accesses, and no undefined behaviour on integer computations.
Mar
3
comment How does zeroing out/wiping your disk before encrypting it increase op-sec?
If you don't "wipe" the disk then the attacker can make out which sectors you actually rewrote in the course of machine usage, which might (just might) leak a tiny bit of information about your usage pattern (average size of created files, or something like that). It is not a very serious leak.
Mar
3
comment How does an attack on a digital signature work?
Public and private key are mathematically linked together; this is what the expression "key pair" means. The owner of the public key is who knows the private key. From the outside, you see only the public key; you do not care how Bob really manages to produce signature, only that he is the only one who can make signatures that match his public key.
Mar
2
comment Is it safe to allow www-data to execute privileged commands?
If you remove the NOPASSWD: then it is somewhat better because sudo will require that you type the password for the www-data account before actually granting the extra privileges. But this raises the question of how the invoking code obtains the password in the first place -- if it is hard-coded in the PHP source, then things are just worse, not better.
Mar
2
comment Password based Asymmetric Key generation
The primes are "just there", mathematically speaking. See this answer.
Mar
1
comment TinyCA missing CRL information in root and leaf certificates
"Netscape CA revocation URL" is reminiscent of the old Netscape-specific extensions that are not the standard extension (such as CDP), and are simply ignored by most modern software. Maybe the URL was stored in the wrong place in the certificate.
Feb
29
comment What does this say? It's in MD5
Known collision attacks on MD5 use differential paths that work mostly on the high-order bits of input words (because these bits are linear in 32-bit additions) so we do not have cheap methods to compute ASCII-compatible MD5 collisions. Or if there is, I want to see it !
Feb
29
comment What does this say? It's in MD5
Not infinite in that the MD5 input is formally limited to 2^61 bytes or so, but yeah, a pretty high number. However, finding another one that matches that one could be quite challenging. Even computing a full-ASCII collision (collisions between two messages that consist only of ASCII characters) would be article-worthy; two meaningful colliding ASCII messages would be even better.
Feb
25
comment Is my understanding of buffer overflows correct?
Ah, well, while the OS traps attempts at writing data over code, it traditionally did not stop using data as code, so the classical exploit would include the attacker's chosen alternate code right into the data it sends for the overflow. Modern desktop/server OS mark the stack as "non-executable" so this kind of thing is harder.
Feb
25
comment What is the difference between a “Thumbprint Algorithm” “Signature Algorithm” and “Signature Hash Algorithm” for a certificates?
So the thumbprint on certificates is for validation jobs that are not covered by the generic certificate validation mechanism. In particular, installation of root CA, because root CA certificates cannot be validated otherwise (because that's what it means to be a root).
Feb
25
comment What is the difference between a “Thumbprint Algorithm” “Signature Algorithm” and “Signature Hash Algorithm” for a certificates?
Thumbprint is for humans. All the certificate validation is about verifying signatures from CA, themselves verified with signatures from other CA. It must start somewhere, with a "root CA" that is trusted because it is already there. When you install a new root CA in your Windows, the interface will display the root CA certificate thumbprint (in hexadecimal); you are supposed to check that it is the correct one (presumably, you compare against a reference printed document that contains the expected root CA thumbprint, or you phone the sysadmin and speak the hex digits).
Feb
25
comment Is this use of PBKDF2 for generating passwords secure?
Also, concatenation would be substantially better than XOR here.