Georgia Tech Research Corporation

Telcordia Technologies, Inc.

Richard J Lipton

Wenke Lee

Anirudh Ramachandran

David Dagon

Nick Feamster

Robert Edmonds

Dan Boneh

Giovanni Di Crescenzo

Giovanni DiCrescenzo

Sheldon Walfish

Richard A DeMillo

A system and method for detecting a first network of compromised computers in a second network of computers, comprising: collecting Domain Name System (DNS) data for the second network; examining the collected data relative to DNS data from known comprised and/or uncompromised computers in the second network; and determining the existence of the first network and/or the identity of compromised computers in the second network based on the examination.

Methods and systems for detecting compromised computers

Collect Domain Name System (DNS) data, the DNS data generated by a DNS server and/or similar device, wherein the DNS data comprises DNS queries, wherein the collected DNS data comprises DNS query rate information. Examine the collected DNS data relative to DNS data from known compromised and/or uncompromised computers. Determine an existence of the collection of compromised networks and/or computers, and/or an identity of compromised networks and/or computers, based on the examination.

Method and systems for detecting compromised networks and/or computers

Method and system for detecting and responding to attacking networks

Efficient secure password protocols are constructed that remain secure against offline dictionary attacks even when a large, but bounded, part of the storage of a server responsible for password verification is retrieved by an adversary through a remote or local connection. A registration algorithm and a verification algorithm accomplish the goal of defeating a dictionary attack. A password protocol where a server, on input of a login and a password, carefully selects several locations from the password files, properly combines their content according to some special function, and stores the result of this function as a tag that can be associated with this password and used in a verification phase to verify access by users. Two main instantiations of our method are given; in one, a combination of mathematical tools, called dispersers and pairwise-independent hash functions is used to achieve security against adaptive intrusions (dispersers make sure that the password of each user depends on randomly chosen locations in a large password file, and pairwise-independent hash functions help in making this dependency sufficiently random); in a second one, a combination of mathematical tools, called k-wise independent hash functions and locally-computable and strong extractors (k-wise independent hash functions make sure that the locations chosen in the large password file from each password are sufficiently random, and locally-computable and strong extractors are used to combine the contents of these locations to generate a single long random value, which makes verification harder for the adversary to foil).

Method and system for password protocols in the bounded retrieval mode with security dictionary attacks and intrusions

A method and software product for limit privacy loss due to data shared in a social network, where the basic underlying assumptions are that users are interested in sharing data and cannot be assumed to constantly follow appropriate privacy policies. Social networks deploy an additional layer of server-assisted access control which, even under no action from a user, automatically evolves over time, by restricting access to the user's data. The evolving access control mechanism provides non-trivial quantifiable guarantees for formally specified requirements of utility (i.e., users share as much data as possible to all other users) and privacy (i.e., users expose combinations of sensitive data only with low probability and over a long time).

Social network privacy by means of evolving access control

A useful method of verifying the integrity of a cryptosystem involves using erroneous outputs to obtain secret information. In certain signature schemes which use the Chinese Remainder Theorem, a correct signature of a message and an erroneous signature of the same message permit the modulus to be easily obtained. If the content of the message is known, such cryptosystems may be cracked with only an erroneous signature of the message. Certain other authorization schemes may be cracked by analyzing a number of erroneous outputs caused by a particular type of error called a 'register fault.' A security expert or cryptosystem designer may intentionally induce a tamper proof device generate a faulty computation by subjecting the device, such as a smart card, to physical stress, such as certain types of radiation, atypical voltage levels, or a higher clock rate than the device was designed to accommodate. Cryptosystems should be impervious to the attacks described herein. If not, the system should be modified or discarded.

Method of using transient faults to verify the security of a cryptosystem

A system enables a user to remove a file from a file system and from all backup tapes without ever mounting a single tape. When a file is backed up, it is first encrypted using a randomly generated key and then backed up. When the user wishes to remove the file from the backup tape, the key used to encrypt the file is deleted. Deleting the encryption key renders the data on the tape inaccessible. No one, including the file owner, can ever access the file again. The encrypt before-backup method is completely transparent to the user. Only the computer system's operating system knows which key was used to encrypt the file during backup. This ensures that when the operating system is instructed to delete the encryption key, the key is 'lost' forever and thus the backed up files become unreadable. Unlike prior cryptographic file systems, the revoked backup files are inaccessible to both the user and others.

Growing community of inventors

Lawrenceville, NJ, United States of America

Richard J Lipton