{"draft":"draft-ietf-tcpm-tcp-antispoof-06","doc_id":"RFC4953","title":"Defending TCP Against Spoofing Attacks","authors":["J. Touch"],"format":["ASCII","HTML"],"page_count":"28","pub_status":"INFORMATIONAL","status":"INFORMATIONAL","source":"TCP Maintenance and Minor Extensions","abstract":"Recent analysis of potential attacks on core Internet infrastructure\r\nindicates an increased vulnerability of TCP connections to spurious\r\nresets (RSTs), sent with forged IP source addresses (spoofing).  TCP\r\nhas always been susceptible to such RST spoofing attacks, which were\r\nindirectly protected by checking that the RST sequence number was\r\ninside the current receive window, as well as via the obfuscation of\r\nTCP endpoint and port numbers.  For pairs of well-known endpoints\r\noften over predictable port pairs, such as BGP or between web servers\r\nand well-known large-scale caches, increases in the path\r\nbandwidth-delay product of a connection have sufficiently increased\r\nthe receive window space that off-path third parties can brute-force\r\ngenerate a viable RST sequence number.  The susceptibility to attack\r\nincreases with the square of the bandwidth, and thus presents a\r\nsignificant vulnerability for recent high-speed networks.  This\r\ndocument addresses this vulnerability, discussing proposed solutions\r\nat the transport level and their inherent challenges, as well as\r\nexisting network level solutions and the feasibility of their\r\ndeployment.  This document focuses on vulnerabilities due to spoofed\r\nTCP segments, and includes a discussion of related ICMP spoofing\r\nattacks on TCP connections.  This memo provides information for the Internet community.","pub_date":"July 2007","keywords":["rst","transport control protocol"],"obsoletes":[],"obsoleted_by":[],"updates":[],"updated_by":[],"see_also":[],"doi":"10.17487\/RFC4953","errata_url":null}