RFC 8922 | Transport Security Survey | October 2020 |
Enghardt, et al. | Informational | [Page] |
This document provides a survey of commonly used or notable network security protocols, with a focus on how they interact and integrate with applications and transport protocols. Its goal is to supplement efforts to define and catalog Transport Services by describing the interfaces required to add security protocols. This survey is not limited to protocols developed within the scope or context of the IETF, and those included represent a superset of features a Transport Services system may need to support.¶
This document is not an Internet Standards Track specification; it is published for informational purposes.¶
This document is a product of the Internet Engineering Task Force (IETF). It represents the consensus of the IETF community. It has received public review and has been approved for publication by the Internet Engineering Steering Group (IESG). Not all documents approved by the IESG are candidates for any level of Internet Standard; see Section 2 of RFC 7841.¶
Information about the current status of this document, any errata, and how to provide feedback on it may be obtained at https://www.rfc-editor.org/info/rfc8922.¶
Copyright (c) 2020 IETF Trust and the persons identified as the document authors. All rights reserved.¶
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.¶
Services and features provided by transport protocols have been cataloged in [RFC8095]. This document supplements that work by surveying commonly used and notable network security protocols, and identifying the interfaces between these protocols and both transport protocols and applications. It examines Transport Layer Security (TLS), Datagram Transport Layer Security (DTLS), IETF QUIC, Google QUIC (gQUIC), tcpcrypt, Internet Protocol Security (IPsec), Secure Real-time Transport Protocol (SRTP) with DTLS, WireGuard, CurveCP, and MinimaLT. For each protocol, this document provides a brief description. Then, it describes the interfaces between these protocols and transports in Section 4 and the interfaces between these protocols and applications in Section 5.¶
A Transport Services system exposes an interface for applications to access various (secure) transport protocol features. The security protocols included in this survey represent a superset of functionality and features a Transport Services system may need to support both internally and externally (via an API) for applications [TAPS-ARCH]. Ubiquitous IETF protocols such as (D)TLS, as well as non-standard protocols such as gQUIC, are included despite overlapping features. As such, this survey is not limited to protocols developed within the scope or context of the IETF. Outside of this candidate set, protocols that do not offer new features are omitted. For example, newer protocols such as WireGuard make unique design choices that have implications for and limitations on application usage. In contrast, protocols such as secure shell (SSH) [RFC4253], GRE [RFC2890], the Layer 2 Tunneling Protocol (L2TP) [RFC5641], and Application Layer Transport Security (ALTS) [ALTS] are omitted since they do not provide interfaces deemed unique.¶
Authentication-only protocols such as the TCP Authentication Option (TCP-AO) [RFC5925] and the IPsec Authentication Header (AH) [RFC4302] are excluded from this survey. TCP-AO adds authentication to long-lived TCP connections, e.g., replay protection with per-packet Message Authentication Codes. (TCP-AO obsoletes TCP MD5 "signature" options specified in [RFC2385].) One primary use case of TCP-AO is for protecting BGP connections. Similarly, AH adds per-datagram authentication and integrity, along with replay protection. Despite these improvements, neither protocol sees general use and both lack critical properties important for emergent transport security protocols, such as confidentiality and privacy protections. Such protocols are thus omitted from this survey.¶
This document only surveys point-to-point protocols; multicast protocols are out of scope.¶
This survey is intended to help identify the most common interface surfaces between security protocols and transport protocols, and between security protocols and applications.¶
One of the goals of the Transport Services effort is to define a common interface for using transport protocols that allows software using transport protocols to easily adopt new protocols that provide similar feature sets. The survey of the dependencies security protocols have upon transport protocols can guide implementations in determining which transport protocols are appropriate to be able to use beneath a given security protocol. For example, a security protocol that expects to run over a reliable stream of bytes, like TLS, restricts the set of transport protocols that can be used to those that offer a reliable stream of bytes.¶
Defining the common interfaces that security protocols provide to applications also allows interfaces to be designed in a way that common functionality can use the same APIs. For example, many security protocols that provide authentication let the application be involved in peer identity validation. Any interface to use a secure transport protocol stack thus needs to allow applications to perform this action during connection establishment.¶
While this survey provides similar analysis to that which was performed for transport protocols in [RFC8095], it is important to distinguish that the use of security protocols requires more consideration.¶
It is not a goal to allow software implementations to automatically switch between different security protocols, even where their interfaces to transport and applications are equivalent. Even between versions, security protocols have subtly different guarantees and vulnerabilities. Thus, any implementation needs to only use the set of protocols and algorithms that are requested by applications or by a system policy.¶
Different security protocols also can use incompatible notions of peer identity and authentication, and cryptographic options. It is not a goal to identify a common set of representations for these concepts.¶
The protocols surveyed in this document represent a superset of functionality and features a Transport Services system may need to support. It does not list all transport protocols that a Transport Services system may need to implement, nor does it mandate that a Transport Service system implement any particular protocol.¶
A Transport Services system may implement any secure transport protocol that provides the described features. In doing so, it may need to expose an interface to the application to configure these features.¶
The following terms are used throughout this document to describe the roles and interactions of transport security protocols (some of which are also defined in [RFC8095]):¶
This section contains brief transport and security descriptions of various security protocols currently used to protect data being sent over a network. These protocols are grouped based on where in the protocol stack they are implemented, which influences which parts of a packet they protect: Generic application payload, application payload for specific application-layer protocols, both application payload and transport headers, or entire IP packets.¶
Note that not all security protocols can be easily categorized, e.g., as some protocols can be used in different ways or in combination with other protocols. One major reason for this is that channel security protocols often consist of two components:¶
For some protocols, such as tcpcrypt, these two components are tightly integrated. In contrast, for IPsec, these components are implemented in separate protocols: AH and the Encapsulating Security Payload (ESP) are record protocols, which can use keys supplied by the handshake protocol Internet Key Exchange Protocol Version 2 (IKEv2), by other handshake protocols, or by manual configuration. Moreover, some protocols can be used in different ways: While the base TLS protocol as defined in [RFC8446] has an integrated handshake and record protocol, TLS or DTLS can also be used to negotiate keys for other protocols, as in DTLS-SRTP, or the handshake protocol can be used with a separate record layer, as in QUIC [QUIC-TRANSPORT].¶
The following protocols provide security that protects application payloads sent over a transport. They do not specifically protect any headers used for transport-layer functionality.¶
TLS (Transport Layer Security) [RFC8446] is a common protocol used to establish a secure session between two endpoints. Communication over this session prevents "eavesdropping, tampering, and message forgery." TLS consists of a tightly coupled handshake and record protocol. The handshake protocol is used to authenticate peers, negotiate protocol options such as cryptographic algorithms, and derive session-specific keying material. The record protocol is used to marshal and, once the handshake has sufficiently progressed, encrypt data from one peer to the other. This data may contain handshake messages or raw application data.¶
DTLS (Datagram Transport Layer Security) [RFC6347] [DTLS-1.3] is based on TLS, but differs in that it is designed to run over unreliable datagram protocols like UDP instead of TCP. DTLS modifies the protocol to make sure it can still provide equivalent security guarantees to TLS with the exception of order protection/non-replayability. DTLS was designed to be as similar to TLS as possible, so this document assumes that all properties from TLS are carried over except where specified.¶
The following protocols provide application-specific security by protecting application payloads used for specific use cases. Unlike the protocols above, these are not intended for generic application use.¶
Secure RTP (SRTP) is a profile for RTP that provides confidentiality, message authentication, and replay protection for RTP data packets and RTP control protocol (RTCP) packets [RFC3711]. SRTP provides a record layer only, and requires a separate handshake protocol to provide key agreement and identity management.¶
The commonly used handshake protocol for SRTP is DTLS, in the form of DTLS-SRTP [RFC5764]. This is an extension to DTLS that negotiates the use of SRTP as the record layer and describes how to export keys for use with SRTP.¶
ZRTP [RFC6189] is an alternative key agreement and identity management protocol for SRTP. ZRTP Key agreement is performed using a Diffie-Hellman key exchange that runs on the media path. This generates a shared secret that is then used to generate the master key and salt for SRTP.¶
The following security protocols provide protection for both application payloads and headers that are used for Transport Services.¶
QUIC is a new standards-track transport protocol that runs over UDP, loosely based on Google's original proprietary gQUIC protocol [QUIC-TRANSPORT] (See Section 3.3.2 for more details). The QUIC transport layer itself provides support for data confidentiality and integrity. This requires keys to be derived with a separate handshake protocol. A mapping for QUIC of TLS 1.3 [QUIC-TLS] has been specified to provide this handshake.¶
Google QUIC (gQUIC) is a UDP-based multiplexed streaming protocol designed and deployed by Google following experience from deploying SPDY, the proprietary predecessor to HTTP/2. gQUIC was originally known as "QUIC"; this document uses gQUIC to unambiguously distinguish it from the standards-track IETF QUIC. The proprietary technical forebear of IETF QUIC, gQUIC was originally designed with tightly integrated security and application data transport protocols.¶
Tcpcrypt [RFC8548] is a lightweight extension to the TCP protocol for opportunistic encryption. Applications may use tcpcrypt's unique session ID for further application-level authentication. Absent this authentication, tcpcrypt is vulnerable to active attacks.¶
MinimaLT [MinimaLT] is a UDP-based transport security protocol designed to offer confidentiality, mutual authentication, DoS prevention, and connection mobility. One major goal of the protocol is to leverage existing protocols to obtain server-side configuration information used to more quickly bootstrap a connection. MinimaLT uses a variant of TCP's congestion control algorithm.¶
The following protocols provide protection for IP packets. These are generally used as tunnels, such as for Virtual Private Networks (VPNs). Often, applications will not interact directly with these protocols. However, applications that implement tunnels will interact directly with these protocols.¶
IKEv2 [RFC7296] and ESP [RFC4303] together form the modern IPsec protocol suite that encrypts and authenticates IP packets, either for creating tunnels (tunnel-mode) or for direct transport connections (transport-mode). This suite of protocols separates out the key generation protocol (IKEv2) from the transport encryption protocol (ESP). Each protocol can be used independently, but this document considers them together, since that is the most common pattern.¶
WireGuard [WireGuard] is an IP-layer protocol designed as an alternative to IPsec for certain use cases. It uses UDP to encapsulate IP datagrams between peers. Unlike most transport security protocols, which rely on Public Key Infrastructure (PKI) for peer authentication, WireGuard authenticates peers using pre-shared public keys delivered out of band, each of which is bound to one or more IP addresses. Moreover, as a protocol suited for VPNs, WireGuard offers no extensibility, negotiation, or cryptographic agility.¶
OpenVPN [OpenVPN] is a commonly used protocol designed as an alternative to IPsec. A major goal of this protocol is to provide a VPN that is simple to configure and works over a variety of transports. OpenVPN encapsulates either IP packets or Ethernet frames within a secure tunnel and can run over either UDP or TCP. For key establishment, OpenVPN can either use TLS as a handshake protocol or use pre-shared keys.¶
Across the different security protocols listed above, the primary dependency on transport protocols is the presentation of data: either an unbounded stream of bytes, or framed messages. Within protocols that rely on the transport for message framing, most are built to run over transports that inherently provide framing, like UDP, but some also define how their messages can be framed over byte-stream transports.¶
The following protocols all depend upon running on a transport protocol that provides a reliable, in-order stream of bytes. This is typically TCP.¶
Application Payload Security Protocols:¶
Transport-Layer Security Protocols:¶
The following protocols all depend on the transport protocol to provide message framing to encapsulate their data. These protocols are built to run using UDP, and thus do not have any requirement for reliability. Running these protocols over a protocol that does provide reliability will not break functionality but may lead to multiple layers of reliability if the security protocol is encapsulating other transport protocol traffic.¶
Application Payload Security Protocols:¶
Transport-Layer Security Protocols:¶
Packet Security Protocols:¶
Of the protocols listed above that depend on the transport for message framing, some do have well-defined mappings for sending their messages over byte-stream transports like TCP.¶
Application Payload Security Protocols:¶
Packet Security Protocols:¶
One protocol surveyed, tcpcrypt, has a direct dependency on a feature in the transport that is needed for its functionality. Specifically, tcpcrypt is designed to run on top of TCP and uses the TCP Encryption Negotiation Option (TCP-ENO) [RFC8547] to negotiate its protocol support.¶
QUIC, CurveCP, and MinimaLT provide both transport functionality and security functionality. They depend on running over a framed protocol like UDP, but they add their own layers of reliability and other Transport Services. Thus, an application that uses one of these protocols cannot decouple the security from transport functionality.¶
This section describes the interface exposed by the security protocols described above. We partition these interfaces into pre-connection (configuration), connection, and post-connection interfaces, following conventions in [TAPS-INTERFACE] and [TAPS-ARCH].¶
Note that not all protocols support each interface. The table in Section 5.4 summarizes which protocol exposes which of the interfaces. In the following sections, we provide abbreviations of the interface names to use in the summary table.¶
Configuration interfaces are used to configure the security protocols before a handshake begins or keys are negotiated.¶
The application can provide its identity, credentials (e.g., certificates), and private keys, or mechanisms to access these, to the security protocol to use during handshakes.¶
The application can choose the algorithms that are supported for key exchange, signatures, and ciphersuites.¶
The application enables or configures extensions that are to be negotiated by the security protocol, such as Application-Layer Protocol Negotiation (ALPN) [RFC7301].¶
The application provides the ability to save and retrieve session state (such as tickets, keying material, and server parameters) that may be used to resume the security session.¶
The application provides access to a separate module that will provide authentication, using the Extensible Authentication Protocol (EAP) [RFC3748] for example.¶
Either the handshake protocol or the application directly can supply pre-shared keys for use in encrypting (and authenticating) communication with a peer.¶
During a handshake, the security protocol will conduct identity validation of the peer. This can offload validation or occur transparently to the application.¶
The handshake protocol may interact with the transport protocol or application to validate the address of the remote peer that has sent data. This involves sending a cookie exchange to avoid DoS attacks. (This list omits protocols that depend on TCP and therefore implicitly perform SAV.)¶
The security protocol may be instructed to tear down its connection and session information. This is needed by some protocols, e.g., to prevent application data truncation attacks in which an attacker terminates an underlying insecure connection-oriented protocol to terminate the session.¶
The handshake protocol may be instructed to update its keying material, either by the application directly or by the record protocol sending a key expiration event.¶
The handshake protocol may provide an interface for producing shared secrets for application-specific uses.¶
The record protocol can signal that its keys are expiring due to reaching a time-based deadline or a use-based deadline (number of bytes that have been encrypted with the key). This interaction is often limited to signaling between the record layer and the handshake layer.¶
The record protocol can be signaled that it is being migrated to another transport or interface due to connection mobility, which may reset address and state validation and induce state changes such as use of a new Connection Identifier (CID).¶
The following table summarizes which protocol exposes which interface.¶
Protocol | IPK | ALG | EXT | CM | AD | PSKI | IV | SAV | CT | KU | SSKE | KE | ME |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
TLS | x | x | x | x | x | x | x | x | x | ||||
DTLS | x | x | x | x | x | x | x | x | x | x | x | ||
ZRTP | x | x | x | x | x | x | |||||||
QUIC | x | x | x | x | x | x | x | x | x | x | |||
tcpcrypt | x | x | x | x | x | x | x | ||||||
MinimaLT | x | x | x | x | x | x | x | x | x | ||||
CurveCP | x | x | x | ||||||||||
IPsec | x | x | x | x | x | x | x | x | x | x | x | ||
WireGuard | x | x | x | x | x | ||||||||
OpenVPN | x | x | x | x | x | x |
x = Interface is exposed
(blank) = Interface is not exposed¶
This document has no IANA actions.¶
This document summarizes existing transport security protocols and their interfaces. It does not propose changes to or recommend usage of reference protocols. Moreover, no claims of security and privacy properties beyond those guaranteed by the protocols discussed are made. For example, metadata leakage via timing side channels and traffic analysis may compromise any protocol discussed in this survey. Applications using Security Interfaces should take such limitations into consideration when using a particular protocol implementation.¶
Analysis of how features improve or degrade privacy is intentionally omitted from this survey. All security protocols surveyed generally improve privacy by using encryption to reduce information leakage. However, varying amounts of metadata remain in the clear across each protocol. For example, client and server certificates are sent in cleartext in TLS 1.2 [RFC5246], whereas they are encrypted in TLS 1.3 [RFC8446]. A survey of privacy features, or lack thereof, for various security protocols could be addressed in a separate document.¶
The authors would like to thank Bob Bradley, Frederic Jacobs, Mirja Kühlewind, Yannick Sierra, Brian Trammell, and Magnus Westerlund for their input and feedback on this document.¶