Tag Archives: cryptography

Homomorphic encryption and trusting the Clouds

Posted on by

Homomorphic encryption is an old idea but only in 2009 and the work of Gentry started to have some possible practical applications. Since then there have been quite impressive improvements in the research in this field of cryptography, also due to the need to improve the security of data managed by Cloud systems.

In brief, homomorphic algorithms are cryptographic algorithms that allow to do computations, like sums, multiplications, searches etc., on encrypted data giving encrypted results, without knowing the encryption key.

It should be obvious that these algorithms would be very useful for Clouds’ applications since the owner of the data would be able to use the data remotely by keeping at the same time the data, and the result of the computations, always encrypted in the Cloud application.

Unfortunately homomorphic encryption is not ready yet for general use, but it has just appeared an interesting research paper by Microsoft Reasearch announcing the release of a SEAL (Simple Encryption Arithmetic Library), a library for using homomorphic encryption in bioinformatics, genomic and other research areas.

Cryptography is too risky: should we use something else to secure IT systems?

Posted on by

Obviously the title of this post is provocative, but reading some recent news it is evident that us, IT professionals and IT industry, are not good in managing cryptography. The consequence is that we deploy cryptography in IT products and give a false sense of security to the users. This actually can have worse consequences than if we would not use cryptography at all. I will give just a couple of examples.

This research paper shows how a well-known brand of hard disks has implemented disk encryption in totally faulty ways, to the point that for some disk models hardly any security is provided by the built-in disk encryption functionalities. This is just another of many similar cases, where cryptographic protocols and algorithms are incorrectly implemented so to cancel all or most of the security that they should provide.

Another research paper shows how a well-funded agency or corporation can in practice break the encryption of any data encrypted with the Diffie-Hellmann (DH) key exchange algorithm using keys up to 1024 bits included. Should we be shocked by this news? Not really since already 10 years ago it was known that a key of 1024 bits is too short for DH. Indeed, as per RFC 7525, a 1024 bit DH key offers a security less than a conventional bit security of 80 bits, but again RFC 7525 states that the absolute (legacy) minimum required conventional bit security must be 112 bits, and the current minimum required conventional bit security is 128 bits, that would practically correspond to a 2048 bits DH key. Even if we, IT professionals and IT industry, have known for at least 10 years that 1024 bits DH keys are too short to offer security to the data that they should protect, as of today a too large number of HTTPS websites, VPNs and SSH servers use DH keys of 1024 bits or less (see again the research paper mentioned above).

Unfortunately these are not two isolated examples, recent news are full of similar facts. So I start to wonder if we are good enough to manage cryptography or if we should look into something else to protect IT systems.

On Ashley-Madison passwords crack

Posted on by

The Ashley-Madison story just got more interesting with the news that it has been possible to crack the supposedly well encrypted users’ password. As Ars Technica (among others) reports, the account passwords have been managed with bcrypt, and this makes it practically impossible to decrypt.

But the account passwords have also been used to create tokens related to the user’s sessions. In this case the password has been hashed with the broken algorithm MD5. From the token it is then easy to recover a lowercase version of the password, and with just a few tries, in some cases even as few as 256 iterations, it is possible to recover the exact password from the bcrypt encrypted value.

This is again another confirmation that security is not a feature to add somewhere in our IT systems, but a fundamental component of each part of it. Everyone doing IT must at least be security aware . In this case it would have been enough to use the SHA2 hash algorithm instead of MD5 to prevent the cracking of the passwords.

A new Ransomware kind of attack

Posted on by

This describes a new kind of IT ransom which should be much more professional and profitable.

The attacker manages to access some company’s servers, then encrypts the data in the databases but he modifies the DBs access routines to encrypt/decrypt on the fly all data with his own encryption key. In this way for the company all continues to work. He then waits a few months so that all DB backups are encrypted with his keys and at this point deletes the encryption keys from the company’s systems and asks for a ransom to give it back.Notice that backups are unusable because they too are encrypted with the attacker key.

Obviously, strong IT security procedures should prevent and detect this, from off-line testing of backups to intrusion detection.

On Cryptolocker and the like

Posted on by

Cryptolocker and similar malware are getting more and more common. The latest versions that appeared work on also Android (one id called Simplelocker). In general what they do is to encrypt some or most of the files on your PC, tablet or smartphone, in particular text, sound, images and video files, which of course includes all your music video library.

Been a ransom, you are asked to pay some bitcoins (or similar untraceable currency) to get your files decrypted.The only defense, a part from keeping your PC clean, up-to-date, with good anti- … whatever … and being very careful on what you click and the email you open, is to keep very updated backups. Indeed once you get infected and locked / encrypted, there is absolutely nothing that you can do to decrypt the files (unless of course if you pay).

The only precaution is to have good and recent backups, and start all-over again from scratch.

But there is a very important point to remember here, not all backups are equal! Good backups are only those done on off-line media, like dvd, blu-ray disks, external usb disks that are connected only for the time of making the backup, and so on. In technical term it is often called an air-gapped backup, that is a storage that you cannot usually access from your device. This excludes most of the Clod storage and backup systems!

The reason for this is that if the backup is on a continuously or very often connected device, and the backup is done automatically as soon as new data is on your device, when the ransomware encrypts your file, the encrypted version is automatically copied on the backup device substituting the original data, and you can end up having also the backup data encrypted.

Note Added: Simplelocker is more a proof-of-concept than a real malware, in these two posts [1] and [2] Simon Bell describes the malware and how to decrypt the files.

A New Approach to Quantum Random Number Generators and news on Quantum Cryptography

Posted on by

I am still interested in developements in the area of Quantum phenomena which can be used in ICT and in particular in ICT Security. Recently there have been quite a few announcements of interest. Here are a some of them:

  • A scientific paper proposes on a new way of generating Quantum Random Number, that is ‘real random numbers’ (whatever that means) by using every day technology like the camera of our smart phone; this does not mean that the smart phone camera is enough to produce real random numbers (for the moment you still need a computer to process the data produced by it) but it is a sign that the technology is providing us with tools of unprecedented power, and soon our smart phone will be enough for a good many things;
  • New developments in Quantum Cryptography (se here and here for details) would make it easier to implement Quantum Cryptography in practice; this is nice, even if it does not changes dramatically the current status and relevance of Quantum Cryptography;
  • Another article (see here for a comment) leaves me instead quite puzzled: either I don’t understand it or there is something fundamentally flawed in the argument otherwise it will look like it is possible to obtain quantum effects in classical physics, which is just what it is not.

Side Channel Cryptanalysis

Posted on by

In line with the previous post, it is of interest, albeit only at the research level and we should not really worry about it right now, the paper published by Adi Shamir, Daniel Genkin and Eran Tromer (download here and here for a comment) in which they describe how they have been able to extract an RSA private key managed by GnuPG 1.4.x  (current version is 2.x) by listening to the noises of the PC.

Yes, an acoustic attack on cryptographic private keys seems very unlikely, even if the idea has been discussed for long time. It is very interesting that it has been shown possible in practice, and this means that also other side channel attacks, like listening on the power cord, should be considered seriously at least when your security requirements are really high.

ENISA Publishes Guidelines on the Use of Cryptography

Posted on by

ENISA just published a report with recommendations on the use of crypto algorithms, keysizes and parameters.

Crypto elements are classified in primitives, schemes, protocols ad key sizes and for each of them it is stated if it is:

  • Legacy not adequate, to be replaced immediately
  • Legacy adequate but with better existing alternatives
  • Future proof and expected to remain secure for 10 to 50 years.

Following the NSA saga and the state of uncertainty we are living in right now, this is a must read.

Social Engineering, Password Reset and DNS Hijack

Posted on by

The DNS provider Web.com has been subject to a Social Engineering attack which allowed a pro-Palestine hacking gang to successfully reset the password of a few important customers, and use the new password to change the resolution of their domain name to other sites. See for example here for a description of the attack.

Again and again, as of today the technical side does not look to be the weak side of  ICT Security. In particular cryptography is sound and reliable, and many technical ICT security products deliver what the promise.

On the other side, username + password show another time how much inappropriate they are to support our current security needs. But what can we use instead?

The general problem lies mostly in our ability to make a system “secure” by including logical, physical and procedural measures to give a 360 degrees protection. Indeed, the security level of a system is that of its weakest point, which for most systems means that they are really insecure.