ICS Facilitating Inaugural ACM Conference on Data and Application Security and Privacy

The Institute for Cyber Security will be facilitating a conference for the SIGSAC in the Association of Computing Machinery in February 2011 in San Antonio, Texas. The conference, called the Conference on Data and Application Security and Privacy (Or CODASPY) will cover topics on cyber security, data privacy, access control, and other fields related to security.

See the announcement below, and at the the CODASPY site.

1st ACM Conference on Data and Application Security and Privacy

February 2011, San Antonio, Texas
Pending ACM Approval

Announcement: ACM SIGSAC announces the creation of a new annual ACM Conference on Data and Applications Security. The inaugural conference will be held in February 2011 in San Antonio, Texas (exact dates and hotel to be announced). With rapid global penetration of the Internet and smart phones and the resulting productivity and social gains, the world is becoming increasingly dependent on its cyber infrastructure. Criminals, spies and predators of all kinds have learnt to exploit this landscape much quicker than defenders have advanced in their technologies. Security and Privacy has become an essential concern of applications and systems throughout their lifecycle. Security concerns have rapidly moved up the software stack as the Internet and web have matured. The security, privacy, functionality, cost and usability tradeoffs necessary in any practical system can only be effectively achieved at the data and application layers. This new conference provides a dedicated venue for high-quality research in this arena, and seeks to foster a community with this focus in cyber security.

Scope: Data and the applications that manipulate data are the crucial assets in today's information age. With the increasing drive towards availability of data and services anytime anywhere, security and privacy risks have increased. Vast amounts of privacy-sensitive data are being collected today by organizations for a variety of reasons. Unauthorized disclosure, modification, usage or denial of access to these data and corresponding services may result in high human and financial costs. New applications such as social networking and social computing provide value by aggregating input from numerous individual users and/or the mobile devices they carry with them and computing new information of value to society and individuals. To achieve efficiency and effectiveness in traditional domains such as healthcare there is a drive to make these records electronic and highly available. The need for organizations and government agencies to share information effectively is underscored by rapid innovations in the business world that require close collaboration across traditional boundaries and the dramatic failure of old-style approaches to information protection in government agencies in keeping information too secret to connect the dots. Security and privacy in these and other arenas can be meaningfully achieved only in context of the application domain. Data and applications security and privacy has rapidly expanded as a research field with many important challenges to be addressed. The goal of the conference is to discuss novel exciting research topics in data and application security and privacy and to lay out directions for further research and development in this area. The conference seeks submissions from diverse communities, including corporate and academic researchers, open source projects, standardization bodies, governments, system and security administrators, software engineers and application domain experts.

Topics: Include but not limited to

• Application layer security policies
• Authorization /Access Control for Applications
• Authorization/Access Control for Databases
• Data dissemination controls
• Data forensics
• Enforcement layer security policies
• Privacy preserving techniques
• Private information retrieval
• Search on protected/encrypted data
• Secure auditing
• Secure collaboration
• Secure data provenance
• Secure electronic commerce
• Secure information sharing
• Secure knowledge management
• Secure multiparty computations
• Secure software development
• Securing data/apps on untrusted platforms
• Securing the semantic web
• Security and Privacy in GIS/Spatial Data
• Security and Privacy in Healthcare
• Security policies for databases
• Social computing security and privacy
• Social networking security and privacy
• Application, data and user trust metrics
• Web application security

General Chair: Ravi Sandhu, Institute for Cyber Security (ICS), University of Texas at San Antonio

Program Chair: Elisa Bertino, Center for Education and Research in Information Assurance and Security (CERIAS), Purdue U.

Local Arrangements Chair: Jeffrey Reich, Institute for Cyber Security (ICS), University of Texas at San Antonio

Program Committee: To be announced

Conference Web Site: www.codaspy.org

Important Dates: To be announced

Interested Volunteers: Contact Ravi Sandhu at ravi.sandhu@utsa.edu or Elisa Bertino at bertino@purdue.edu as appropriate

Erhan J. Kartaltepe, PMP
erhan.kartaltepe-at-utsa.edu

Creating a Rogue Certificate Authority

The 25th Annual Chaos Communication Congress released a paper that shows how to exploit a MD5 collision issue. The authors show that all root certificate authorities that use MD5 hashes are susceptible to the attack. A good technical read that includes the attack details and its impact.

Erhan J. Kartaltepe,
erhan.kartaltepe-at-utsa.edu

Secure Hash Functions

The news of NIST and their SHA-3 algorithm competition reminded me of the Cryptographic lectures I gave at UTSA. One of the hardest concepts my students had grasping was secure cryptographic hash functions, partially because of the number theory, but also in differentiating between the three properties of a secure hash function: preimage resistance, collision resistance, and second preimage resistance.

Preimage resistance: Given a hash value h, it should be hard to find any message m such that h = hash(k, m).

This is the most usual property developers think of when they think of a cryptographic hash function. Unlike an encryption, there should be no "dehashing" function. A good preimage resistant function should be "hard" to invert.  An example of a hash function that is not preimage resistant is h = hash(k, m) = m mod 2^k. For cryptographic hash functions, "hard" is, for a hash function with a range of size 2^k, it should take the adversary 2^k/2 or 2^k-1 attempts before he or she finds the solution. Since this is very easy to invert, for any value h, a message of m can be found (basically, every message of the form h + x * 2^k_, where x is an integer).

Collision resistance: Given two messages m₁ and m₂, it should be "hard" to find a hash such that hash(k, m₁) = hash(k, m₂), where k is the hash key.

What this says is that given complete control over picking any messages you want, it should be "hard" to find two of them such that have the same hash value the same hash. An example of a hash function that is not collision resistant is hash(k, m) = 4, since all hashes result in 4, making it 100% likely that two messages will have the same hash. This is the hardest standard for a hash function to attain; collision resistant hash functions are in fact a superset of second preimage resistant ones (if you've found a second preimage of a hash function, you have necessarily found a collision).

Second preimage resistance: Given a message m₁, it should be hard to find a different message m₂ such that hash(k, m₁) = hash(k, m₂).

This is the toughest of the three for students in cryptography to get their head around. On the surface this seems to be an easier version of collision resistance, because it seems that we have extra information to use. We only need one more message rather than two. In actuality, this is the birthday paradox in action—second preimage resistance is actually a much stronger property than collision resistance.

Of course, it's important to reiterate that hashes are "digests", and not "encryption" algorithms. Encryption is a two-way operation. That is, given a message m, an encryption algorithm enc and its corresponding decryption algorithm dec, dec(enc(m)) = m. Hash functions are different—there is (ideally) no way to "decrypt" a hashed message, per preimage resistance. Because of this, it is common to say that hash functions are one-way operations.

Erhan J. Kartaltepe,
erhan.kartaltepe-at-utsa.edu

NIST Announces SHA-3 Competition Round 1 Candidates

About a year ago, the National Institute of Standards and Technology (NIST) announced an open competition for a new cryptographic hash algorithm called Secure Hash Algorithm 3 (SHA-3), the next in the line of SHA algorithms. Submissions were accepted until October, and the competition timeline suggests that the winner will be announced in 2012.

Since then, 64 entries were submitted, of which 51 were accepted, and one by one, each are being systematially tested to make sure they adhere to secure hashing principles, and some have already been broken. It'll be interesting to see which one makes it, if any do.

Image is an aerial shot of the NIST compound.

Erhan J. Kartaltepe,
erhan.kartaltepe-at-utsa.edu