Deep learning takes advantage of large datasets and computationally efficient training algorithms to outperform other approaches at various machine learning tasks. However, imperfections in the training phase of deep neural networks make them vulnerable to adversarial samples: inputs crafted by adversaries with the intent of causing deep neural networks to misclassify. In this work, we formalize the space of adversaries against deep neural networks (DNNs) and introduce a novel class of algorithms to craft adversarial samples based on a precise understanding of the mapping between inputs and outputs of DNNs. In an application to computer vision, we show that our algorithms can reliably produce samples correctly classified by human subjects but misclassified in specific targets by a DNN with a 97% adversarial success rate while only modifying on average 4.02% of the input features per sample. We then evaluate the vulnerability of different sample classes to adversarial perturbations by defining a hardness measure. Finally, we describe preliminary work outlining defenses against adversarial samples by defining a predictive measure of distance between a benign input and a target classification.

The Limitations of Deep Learning in Adversarial Settings

Gigantically comprehensive and carefully researched, Security Engineering makes it clear just how difficult it is to protect information systems from corruption, eavesdropping, unauthorized use, and general malice. Better, Ross Anderson offers a lot of thoughts on how information can be made more secure (though probably not absolutely secure, at least not forever) with the help of both technologies and management strategies. His work makes fascinating reading and will no doubt inspire considerable doubt--fear is probably a better choice of words--in anyone with information to gather, protect, or make decisions about. Be aware: This is absolutely not a book solely about computers, with yet another explanation of Alice and Bob and how they exchange public keys in order to exchange messages in secret. Anderson explores, for example, the ingenious ways in which European truck drivers defeat their vehicles' speed-logging equipment. In another section, he shows how the end of the cold war brought on a decline in defenses against radio-frequency monitoring (radio frequencies can be used to determine, at a distance, what's going on in systems--bank teller machines, say), and how similar technology can be used to reverse-engineer the calculations that go on inside smart cards. In almost 600 pages of riveting detail, Anderson warns us not to be seduced by the latest defensive technologies, never to underestimate human ingenuity, and always use common sense in defending valuables. A terrific read for security professionals and general readers alike. --David Wall Topics covered: How some people go about protecting valuable things (particularly, but not exclusively, information) and how other people go about getting it anyway. Mostly, this takes the form of essays (about, for example, how the U.S. Air Force keeps its nukes out of the wrong hands) and stories (one of which tells of an art thief who defeated the latest technology by hiding in a closet). Sections deal with technologies, policies, psychology, and legal matters.

Security Engineering: A Guide to Building Dependable Distributed Systems

In 1998, Blaze, Bleumer, and Strauss (BBS) proposed an application called atomic proxy re-encryption, in which a semitrusted proxy converts a ciphertext for Alice into a ciphertext for Bob without seeing the underlying plaintext. We predict that fast and secure re-encryption will become increasingly popular as a method for managing encrypted file systems. Although efficiently computable, the wide-spread adoption of BBS re-encryption has been hindered by considerable security risks. Following recent work of Dodis and Ivan, we present new re-encryption schemes that realize a stronger notion of security and demonstrate the usefulness of proxy re-encryption as a method of adding access control to a secure file system. Performance measurements of our experimental file system demonstrate that proxy re-encryption can work effectively in practice.

/pdf/improved-proxy-re-encryption-schemes-with-applications-to-1ugoecwezg.pdf

Improved proxy re-encryption schemes with applications to secure distributed storage

The invention includes a markup language according to the SGML standard in which document type definitions are created under which electronic documents are divided into blocks that are associated with logical fields that are specific to the type of block. Each of many different types of electronic documents can have a record mapping to a particular environment, such as a legacy environment of a banking network, a hospital's computer environment for electronic record keeping, a lending institution's computer environment for processing loan applications, or a court or arbitrator's computer system. Semantic document type definitions for various electronic document types (including, for example, electronic checks, mortgage applications, medical records, prescriptions, contracts, and the like) can be formed using mapping techniques between the logical content of the document and the block that is defined to include such content. Also, the various document types are preferably defined to satisfy existing customs, protocols and legal rules.

Method and system for processing electronic documents

Acknowledgments. 1. Introduction. Roadmap to the Book. What Type of Book Is This? Terminology. Notation. Primer on Networking. Active vs. Passive Attackc. Layers and Cryptography. Authorization. Tempest. Key Escrow for Law Enforcement. Key Escrow for Careless Users. Viruses, Worms, Trojan Horses. The Multi-level Model of Security. Legal Issues. I. CRYPTOGRAPHY. 2. Introduction to Cryptography. What Is Cryptography? Breaking an Encryption Scheme. Types of Cryptographic Functions. Secret Key Cryptography. Public Key Cryptography. Hash Algorithms. Homework. 3. Secret Key Cryptography. Introduction. Generic Block Encryption. Data Encryption Standard (DES). International Data Encryption Algorithm (IDEA). Advanced Encryption Standard (AES). RC4. Homework. 4. Modes of Operation. Introduction. Encrypting a Large Message. Generating MACs. Multiple Encryption DES. CBC Outside vs. Inside. Homework. 5. Hashes and Message Digests. Introduction. Nifty Things to Do with a Hash. MD2. MD4. MD5. SHA-1. HMAC. Homework. 6. Public Key Algorithms. Introduction. Modular Arithmetic. RSA. Diffie-Hellman. Digital Signature Standard (DSS). How Secure Are RSA and Diffie-Hellman? Elliptic Curve Cryptography (ECC). Zero Knowledge Proof Systems. Homework Problems. 7. Number Theory. Introduction. Modular Arithmetic. Primes. Euclid's Algorithm. Chinese Remainder Theorem. Zn. Euler's Totient Function. Euler's Theorem. Homework Problems. 8. Math with AES and Elliptic Curves. Introduction. Notation. Groups. Fields. Mathematics of Rijndael. Elliptic Curve Cryptography. Homework. II. AUTHENTICATION. 9. Overview of Authentication Systems. Password-Based Authentication. Address-Based Authentication. Cryptographic Authentication Protocols. Who Is Being Authenticated? Passwords as Cryptographic Keys. Eavesdropping and Server Database Reading. Trusted Intermediaries. Session Key Establishment. Delegation. Homework. 10. Authentication of People. Passwords. On-Line Password Guessing. Off-Line Password Guessing. How Big Should a Secret Be? Eavesdropping. Passwords and Careless Users. Initial Password Distribution. Authentication Tokens. Physical Access. Biometrics. Homework. 11. Security Handshake Pitfalls. Login Only. Mutual Authentication. Integrity/Encryption for Data. Mediated Authentication (with KDC). Nonce Types. Picking Random Numbers. Performance Considerations. Authentication Protocol Checklist. Homework. 12. Strong Password Protocols. Introduction. Lamport's Hash. Strong Password Protocols. Strong Password Credentials. Strong Password Credentials Download Protocols. Homework. III. STANDARDS. 13. Kerberos V4. Introduction. Tickets and Ticket-Granting Tickets. Configuration. Logging Into the Network. Replicated KDC's. Realms. Interrealm Authentication. Key Version Numbers. Encryption for Privacy and Integrity. Encryption for Integrity Only. Network Layer Addresses in Tickets. Message Formats. Homework. 14. Kerberos V5. ASN.1. Names. Delegation of Rights. Ticket Lifetimes. Key Versions. Making Master Keys in Different Realms Different. Optimizations. Cryptographic Algorithms. Hierarchy of Realms. Evading Password-Guessing Attacks. Key Inside Authenticator. Double TGT Authentication. PKINIT-Public Keys for Users. KDC Database. Kerberos V5 Messages. Homework. 15. PKI (Public Key Infrastructure). Introduction. Some Terminology. PKI Trust Models. Revocation. Directories and PKI. PKIX and X.509. X.509 and PKIX Certificates. Authorization Futures. Homework. 16. Real-time Communication Security. What Layer? Session Key Establishment. Perfect Forward Secrecy. PFS-Foilage. Denial-of-Service/Clogging Protection. Endpoint Identifier Hiding. Live Partner Reassurance. Arranging for Parallel Computation. Session Resumption. Plausible Deniability. Data Stream Protection. Negotiating Crypto Parameters. Easy Homework. Homework. 17. IPsec: AH and ESP. Overview of Ipsec. IP and Ipv6. AH (Authentication Header). ESP (Encapsulating Security Payload). So, Do We Need AH? Comparison of Encodings. Easy Homework. Homework. 18. IPsec: IKE. Photuris. SKIP. History of IKE. IKE Phases. Phase 1 IKE. Phase - 2 IKE: Setting up Ipsec Sas. ISAKMP/IKE Encoding. Homework. 19. SSL/TLS. Introduction. Using TCP. Quick TCP. Quick History. SSL/TLS Basic Protocol. Session Resumption. Computing the Keys. Client Authentication. PKI as Deployed by SSL. Version Numbers. Negotiating Cipher Suites. Negotiating Compression Method. Attacks Fixed in v3. Exportability. Encoding. Further Reading. Easy Homework. Homework. IV. ELECTRONIC MAIL. 20. Electronic Mail Security. Distribution Lists. Store and Forward. Security Services for Electronic Mail. Establishing Keys. Privacy. Authentication of the Source. Message Integrity. Non-Repudiation. Proof of Submission. Proof of Delivery. Message Flow Confidentiality. Anonymity. Containment. Annoying Text Format Issues. Names and Addresses. Verifying When a Message Was Really Sent. Homework. 21. PEM & S/MIME. Introduction. Structure of a PEM Message. Establishing Keys. Some PEM History. PEM Certificate Hierarchy. Certificate Revocation Lists (CRLs). Reformatting Data to Get Through Mailers. General Structure of a PEM Message. Encryption. Source Authentication and Integrity Protection. Multiple Recipients. Bracketing PEM Messages. Forwarding and Enclosures. Unprotected Information. Message Formats. DES-CBC as MIC Doesn't Work. Differences in S/MIME. S/MIME Certificate Hierarchy. Homework. 22. PGP (Pretty Good Privacy). Introduction. Overview. Key Distribution. Efficient Encoding. Certificate and Key Revocation. Signature Types. Your Private Key. Key Rings. Anomalies. Object Formats. V. LEFTOVERS. 23. Firewalls. Packet Filters. Application Level Gateway. Encrypted Tunnels. Comparisons. Why Firewalls Don't Work. Denial-of-Service Attacks. Should Firewalls Go Away? 24. More Security Systems. NetWare V3. NetWare V4. KryptoKnight. DASS/SPX. Lotus Notes Security. DCE Security. Microsoft Windows Security. Network Denial of Service. Clipper. Homework. 25. Web Issues. Introduction. URLs/URIs. HTTP. HTTP Digest Authentication. Cookies. Other Web Security Problems. Homework. 26. Folklore. Perfect Forward Secrecy. Change Keys Periodically. Multiplexing Flows over a Single SA. Use Different Keys in the Two Directions. Use Different Secret Keys for Encryption vs. Integrity Protection. Use Different Keys for Different Purposes. Use Different Keys for Signing vs. Encryption. Have Both Sides Contribute to the Master Key. Don't Let One Side Determine the Key. Hash in a Constant When Hashing a Password. HMAC Rather than Simple MD. Key Expansion. Randomly Chosen Ivs. Use of Nonces in Protocols. Don't Let Encrypted Data Begin with a Constant. Don't Let Encrypted Data Begin with a Predictable Value. Compress Data Before Encrypting It. Don't Do Encryption Only. Avoiding Weak Keys. Minimal vs. Redundant Designs. Overestimate the Size of Key. Hardware Random Number Generators. Timing Attacks. Put Checksums at the End of Data. Forward Compatibility. Negotiating Parameters. Homework. Bibliography. Glossary. Index.

Network Security: Private Communication in a Public World

The classic guide to network securityi¾now fully updated!"Bob and Alice are back!"Widely regarded as the most comprehensive yet comprehensible guide to network security, the first edition of Network Security received critical acclaim for its lucid and witty explanations of the inner workings of network security protocols. In the second edition, this most distinguished of author teams draws on hard-won experience to explain the latest developments in this field that has become so critical to our global network-dependent society.Network Security, Second Edition brings together clear, insightful, and clever explanations of every key facet of information security, from the basics to advanced cryptography and authentication, secure Web and email services, and emerging security standards. Coverage includes: All-new discussions of the Advanced Encryption Standard (AES), IPsec, SSL, and Web security Cryptography: In-depth, exceptionally clear introductions to secret and public keys, hashes, message digests, and other crucial concepts Authentication: Proving identity across networks, common attacks against authentication systems, authenticating people, and avoiding the pitfalls of authentication handshakes Core Internet security standards: Kerberos 4/5, IPsec, SSL, PKIX, and X.509 Email security: Key elements of a secure email system-plus detailed coverage of PEM, S/MIME, and PGP Web security: Security issues associated with URLs, HTTP, HTML, and cookies Security implementations in diverse platforms, including Windows, NetWare, and Lotus Notes The authors go far beyond documenting standards and technology: They contrast competing schemes, explain strengths and weaknesses, and identify the crucial errors most likely to compromise secure systems. Network Security will appeal to a wide range of professionals, from those who design or evaluate security systems to system administrators and programmers who want a better understanding of this important field. It can also be used as a textbook at the graduate or advanced undergraduate level.Prentice Hall Series in Computer Networking and Distributed Systems

Network security: private communication in a public world, second edition

We present protocols that allow a user Alice, knowing only her name and password, and not carrying a smart card, to “log in to the network” from a “generic” workstation, i.e., one that has all the necessary software installed, but none of the configuration information usually assumed to be known a priori in a security scheme, such as Alice’s public and private keys, her certificate, and the public keys of one or more CAs. By “logging in”, we mean the workstation retrieves this information on behalf of the user. This would be straightforward if Alice had a cryptographically strong password. We propose protocols that are secure even if Alice’s password is guessable. We concentrate on the initial retrieval of Alice’s private key from some server Bob on the network. We discuss various protocols for doing this that avoid off-line password guessing attacks by someone eavesdropping or impersonating Alice or Bob. We discuss auditable vs. unauditable on-line attacks, and present protocols that allow Bob to be stateless, avoid denial-of-service attacks, allow for salt, and are minimal in computation and number of messages.

Secure Password-Based Protocol for Downloading a Private Key.

Since IP packet reassembly requires resources, a denial of service attack can be mounted by swamping a receiver with IP fragments. In this paper we argue how this attack need not affect protocols that do not rely on IP fragmentation, and argue how most protocols, e.g., those that run on top of TCP, can avoid the need for fragmentation. However, protocols such as IPsec's IKE protocol, which both runs on top of UDP and requires sending large packets, depend on IP packet reassembly. Photuris, an early proposal for IKE, introduced the concept of a stateless cookie, intended for DoS protection. However, the stateless cookie mechanism cannot protect against a DoS attack unless the receiver can successfully receive the cookie, which it will not be able to do if reassembly resources are exhausted. Thus, without additional design and/or implementation defenses, an attacker can successfully, through a fragmentation attack, prevent legitimate IKE handshakes from completing. Defense against this attack requires both protocol design and implementation defenses. The IKEv2 protocol was designed to make it easy to design a defensive implementation. This paper explains the defense strategy designed into the IKEv2 protocol, along with the additional needed implementation mechanisms. It also describes and contrasts several other potential strategies that could work for similar UDP-based protocols.

/pdf/dos-protection-for-udp-based-protocols-2yiz0lvzeb.pdf

DoS protection for UDP-based protocols

In this paper we present PDM (Password Derived Moduli), a new approach to strong password-based protocols usable either for mutual authentication or for downloading security information such as the user's private key We describe how the properties desirable for strong password mutual authentication differ from the properties desirable for credentials download In particular, a protocol used solely for credentials download can be simpler and less expensive than one used for mutual authentication since some properties (such as authentication of the server) are not necessary for credentials download The features necessary for mutual authentication can be easily added to a credentials download protocol, but many of the protocols designed for mutual authentication are not as desirable for use in credentials download as protocols like PDM and basic EKE and SPEKE because they are unnecessarily expensive when used for that purpose PDM's performance is vastly more expensive at the client than any of the protocols in the literature, but it is more efficient at the server We claim that performance at the server, since a server must handle a large and potentially unpredictable number of clients, is more important than performance at the client, assuming that client performance is "good enough" We describe PDM for credentials download, and then show how to enhance it to have the properties desirable for mutual authentication In particular, the enhancement we advocate for allowing PDM to avoid storing a password-equivalent at the server is less expensive than existing schemes, and our approach can be used as a more efficient (at the server) variant of augmented EKE and SPEKE than the currently published schemes PDM is important because it is a very different approach to the problem than any in the literature, we believe it to be unencumbered by patents, and because it can be a lot less expensive at the server than existing schemes

/pdf/pdm-a-new-strong-password-based-protocol-5cvsd24e04.pdf

Charlie Kaufman

Papers

Network Security: Private Communication in a Public World

Network security: private communication in a public world, second edition

Secure Password-Based Protocol for Downloading a Private Key.

DoS protection for UDP-based protocols

PDM: a new strong password-based protocol