Adaptive Cryptographically Synchronized Authentication (ACSA) Final Report

The information in this document may change at any time. No part of this document can be reproduced or transmitted in any form or by any means, electronic or mechanical, for any purpose, without the express written permission of Network Associates, Inc. The U.S. Department of Commerce controls export of cryptographic items such as the ACSA prototype toolkit mentioned herein under the Export Administration Regulations. Exporters of cryptographic items such as the ACSA prototype toolkit mentioned herein must obtain any export authorizations and licenses required under the Export Administration Regulations. LINUX is a trademark of Linus Torvalds. RED HAT is a registered trademark of Red Hat, Inc. All other names and trademarks are the property of their respective owners.Executive Summary Conventional authentication mechanisms do not operate at speeds fast enough to meet the demands of ultra-fast networks [1][23][29]. This disparity presents a great challenge for high-speed applications that demand network authentication. Much work has focused on devising fast authentication algorithms. The Adaptive Cryptographically Synchronized Authentication (ACSA) Project [10] provides a new solution to this challenge by trading off authentication strength and performance to achieve data origin authentication and connectionless integrity. Some of the applications that motivate our work include real-time high-speed video, high-performance distributed computing, high-speed distributed storage, and 3-D virtual reality. Such applications might run on conventional platforms with high-speed network devices (e.g. Gigabit Ethernet or ATM networks on Pentium class machines without cryptographic hardware), on special platforms with conventional cryptographic hardware, or on computationally limited single-processor devices. The ACSA Project offers a practical and flexible solution to high-speed network authentication that can be implemented in software or hardware. Figure 1 illustrates hypothetical strength-performance tradeoffs that ACSA might achieve. For example, the user can select a lower-speed, high-strength authentication mechanism, a higher-speed, lower-strength mechanism, or a high-speed Partial Message Authentication Code (PMAC) that authenticates only a portion of the message. Throughout, we use the term " network authentication " to mean data origin authentication and connectionless integrity, as defined in Internet standards [17]. This type of authentication is without regard to the ordering of the message in a stream of messages. By " authentication strength " we mean the computational difficulty of producing a forgery. By " performance " we mean computational complexity (CPU time, memory space). The ACSA project aims to lighten the processing loads of the sender and receiver by requiring them to spend less processor resources …