]> Sandelman Software Works

mcr+ietf@sandelman.ca

Cloudflare Ltd.

jhoyland@cloudflare.com

Security Security Area Working Group Internet-Draft The terms on-path attacker and MITM Attack have been used in a variety of ways, sometimes interchangeably, and sometimes meaning different things. Increasingly people have become uncomfortable with the gendered term "Man" in the middle and have sought alternatives. This document offers an update on terminology for network attacks, retaining some acronyms terms while redefining the expansion, and clarifying the different kinds of attacks. Consistent terminology is important in describing what kinds of attacks a particular protocol defends against, and which kinds the protocol does not.

Introduction A number of terms have been used to describe attacks against networks. In the paper, the attacker is assumed to be able to:

• view messages as they are transmitted
• selectively delete messages
• selectively insert or modify messages

Some authors refer to such an attacker as an "on-path" attacker [reference], or a "Man-in-the-Middle" [reference]. Despite a broad consensus on what is meant by a MITM attack, there is less agreement on the how to describe its variants. The term "passive attacker" has been used in many cases to describe situations where the attacker can only observe messages, but can not intercept, modify or delete any messages. Another variant is the case where an eavesdropper is not on the network path between the actual correspondants, and thus cannot drop messages, they may be able to inject packets faster than the correspondants, and thus beat legitimate packets in a race. As summarised, there are three broad variations of the MITM attacker:

1. An on-path attacker that can view, delete and modify messages. This is the Dolev-Yao attack.
2. An off-path attacker that can view messages and insert new messages.
3. An off-path attacker that can only view messages.

Three kinds of attack The attacks are numbered in this section as no consensus on naming the attacks yet. In the diagrams below, the sender is named "Alice", and the recipient is named "Bob", as is typical in many cryptographic protocols , as first introduced by . Alice and Bob were named as expansions of "A" and "B", which would otherwise be very abstract concepts of the two end points. The attacker has historically been named "Mallory", but this document proposes that the expansion be named "Meddler"

Alice communicating with Bob

Active On-Path Attacker, or Meddler in the Middle (MITM) In this attack, the attacker is involved with the forwarding of the packets. A firewall or network router is ideally placed for this attack.

On Path Attacker | Bob | '-------' ╰─────────╯ '-----' ]]>

In this case the Meddler can:

• view all packets
• selectively forward or drop any packet
• modify any packets that is forwarded
• insert additional packets

Passive On-Path attack In this attack, the attacker is not involved with the forwarding of the packets. The attacker receives a copy of packets that are sent along the path. This could be from, for instance, a mirror port or SPAN . Alternatively, a copy of traffic may be obtained via passive (optical) tap . This kind of attack is often associated with Pervasive Monitoring .

Passive On-Path attack | Bob | '-------' | '-----' | v ╭─────────╮ │ Meddler │ ╰─────────╯ ]]>

In this the meddler can:

• view all packets

Note that they have no way to inject new packets, and this attack may occur seconds to decades after the data was exchanged.

Passive On-Path attack with bypass In some cases, the Meddler is be able to send messages to Bob via another route. Due to some other factor (such as shorter or higher cost routing), these messages arrive at Bob prior to the original message from Alice.

Passive On-Path attack with bypass | Bob | '-------' | │ │ │ │ '-----' | │ │ ╰───╯ ^ v │ │ │ ╭─────────╮ ╰──╯ │ │ Meddler │────────────────────╯ ╰─────────╯ ]]>

In that the Meddler can:

• view all packets
• insert additional/copied packets into the stream

But the Meddler is unable to drop or modify the original packets. Bob however, may be unable to distinguish packets from Alice vs packets sent from the Meddler that purport to be from Alice. To be effective or useful, this type of attack needs to occur in real time.

Passive Off-path attacker The third kind of attack is one in which the Meddler can not see any packets from Alice. This is usually what is meant by an "off-path" attack. The meddler can forge packets purporting to be from Alice, but can never see Alice's actual packets.

Passive Off-path attacker | Bob | '-------' '-----' ^ │ ╭─────────╮ │ │ Meddler │───────────────────╯ ╰─────────╯ ]]>

In this the Meddler can:

• insert additional packets

Existing uses of the terms

IETF QUIC terms ended up the following taxonomy:

on-path:

[Dolev-Yao] MITM, Active On-Path attacker

Limited on-path (cannot delete):

Active Off-Path attacker

Off-path:

Passive Off-Path attacker

Security Considerations This document introduces a set of terminology that will be used in many Security Considerations sections.

IANA Considerations This document makes no IANA requests.

Acknowledgements The SAAG mailing list.

Changelog

References Normative References This Glossary provides definitions, abbreviations, and explanations of terminology for information system security. The 334 pages of entries offer recommendations to improve the comprehensibility of written material that is generated in the Internet Standards Process (RFC 2026). The recommendations follow the principles that such writing should (a) use the same term or definition whenever the same concept is mentioned; (b) use terms in their plainest, dictionary sense; (c) use terms that are already well-established in open publications; and (d) avoid terms that either favor a particular vendor or favor a particular technology or mechanism over other, competing techniques that already exist or could be developed. This memo provides information for the Internet community. Informative References MIT Lab. for Computer Science and Department of Mathematics, Cambridge, MA MIT Lab. for Computer Science and Department of Mathematics, Cambridge, MA MIT Lab. for Computer Science and Department of Mathematics, Cambridge, MA Pervasive monitoring is a technical attack that should be mitigated in the design of IETF protocols, where possible.

Monster in the Middle As a special case for the MITM, if the Meddler steals cookies (whether they are HTTP Cookies, IKE nonces, or TCP SYN Cookies), then this kind of attack is a Monster in The Middle. This is otherwise known as a: nom-nom-nom-nom attack.

Contributors

ekr@rtfm.com

lberger@labn.net

aland@deployingradius.com

huitema@huitema.net