<------ Implicit 'Internet' PvD ------>
        +----+     +----+         _   __              _  _
        |    |     |    |        ( `    )            ( `   )_
        | PC +-----+    |--------------------------(         `)
        |    |     |    |      (         )        (_ Internet  _)
        +----+     |    |     (           )         `- __  _) -
                   |Home|    (   Service    )
                   |gate|    (  Provider's   )
                   |-way|    (   Network     -
        +-----+    |    |    `_            )        +---------+
        | Set |    |    |      (          )         |ISP Video|
        | Top +----+    |---------------------------+on Demand|
        | Box |    |    |       (_     __)          | Service |
        +-----+    +----+         `- --             +---------+
              <-- Explicit 'Video-on-Demand' PvD -->
            

An example of PvD use within a home network.

Figure 3

In this case, the number of PvDs that a single operator could provision is based on the number of independently provisioned services which they offer. Some examples may include:

• Real-time packet voice
• Streaming video
• Interactive video (n-way video conferencing)
• Interactive gaming
• Best effort / Internet access

5. Reference Model for the PvD-aware Node

5.1. Constructions and Maintenance of Separate PvDs

It is assumed that normally, the configuration information contained in a single PvD shall be sufficient for a node to fulfill a network connection request by an application, and hence there should be no need to attempt to merge information across different PvDs.

Nevertheless, even when a PvD lacks some necessary configuration information, merging of information associated with different PvD(s) shall not be done automatically as this will typically lead to the issues described in [RFC6418].

A node may use other sources, for example: node local policy, user input or other mechanisms not defined by the IETF for any of the following:

• Construction of a PvD in its entirety (analogous to statically configuring IP on an interface)
• Supplementing some, or all learned PvDs with particular configuration elements
• Merging of information from different PvDs (if this is explicitly allowed by policy)

As an example, a node administrator could inject a DNS server which is not ISP-specific into PvDs for use on any of the networks that the node could attach to. Such creation / augmentation of PvD(s) could be static or dynamic. The specific mechanism(s) for implementing this are outside of scope of this document.

5.2. Consistent use of PvDs for Network Connections

PvDs enable PvD-aware nodes to consistently use the correct set of configuration elements to serve specific network requests from beginning to end. This section provides examples of such use.

5.2.1. Name Resolution

When a PvD-aware node needs to resolve the name of the destination for use by a connection request, the node could use one, or multiple PvDs for a given name lookup.

The node shall chose a single PvD if, for example, the node policy required the use of a particular PvD for a specific purpose (e.g. to download an MMS message using a specific APN over a cellular connection). To make this selection, the node could use a match between the PvD DNS suffix and an FQDN which is being resolved or match of PvD ID, as determined by the node policy.

The node may pick multiple PvDs, if for example, the PvDs are for general purpose Internet connectivity, and the node is attempting to maximize the probability of connectivity similar to the Happy Eyeballs [RFC6555] approach. In this case, the node could perform DNS lookups in parallel, or in sequence. Alternatively, the node may use only one PvD for the lookup, based on the PvD connectivity properties, user configuration of preferred Internet PvD, etc.

If an application implements an API that provides a way of explicitly specifying the desired interface or PvD, that interface or PvD should be used for name resolution (and the subsequent connection attempt), provided that the host's configuration permits this.

In either case, by default a node uses information obtained via a name service lookup to establish connections only within the same PvD as the lookup results were obtained.

For clarification, when it is written that the name service lookup results were obtained "from a PvD", it should be understood to mean that the name service query was issued against a name service which is configured for use in a particular PvD. In that sense, the results are "from" that particular PvD.

Some nodes may support transports and / or APIs which provide an abstraction of a single connection, aggregating multiple underlying connections. MPTCP [RFC6182] is an example of such a transport protocol. For connections provided by such transports/APIs, a PvD-aware node may use different PvDs for servicing that logical connection, provided that all operations on the underlying connections are performed consistently within their corresponding PvD(s).

5.2.2. Next-hop and Source Address Selection

For the purpose of this example, let us assume that the preceding name lookup succeeded in a particular PvD. For each obtained destination address, the node shall perform a next-hop lookup among routers associated with that PvD. As an example, the node could determine such associations via matching the source address prefixes/specific routes advertized by the router against known PvDs, or receiving an explicit PvD affiliation advertized through a new Router Discovery [RFC4861] option.

For each destination, once the best next-hop is found, the node selects the best source address according to rules defined in [RFC6724], but with the constraint that the source address must belong to a range associated with the used PvD. If needed, the node would use prefix policy from the same PvD for selecting the best source address from multiple candidates.

When destination / source pairs are identified, they are sorted using the [RFC6724] destination sorting rules and prefix policy table from the used PvD.

5.2.3. Listening Applications

Consider a host connected to several PvDs, running an application that opens a listening socket / transport API object. The application is authorized by the host policy to use a subset of connected PvDs that may or may not be equal to the complete set of the connected PvDs. As an example, in the case where there are different PvDs on the Wi-Fi and cellular interfaces, for general Internet traffic the host could use only one, preferred PvD at a time (and accordingly, advertise to remote peers the host name and addresses associated with that PvD), or it could use one PvD as the default for outgoing connections, while still allowing use of the other PvDs simultaneously.

Another example is a host with an established VPN connection. Here, security policy could be used to permit or deny application's access to the VPN (and other) PvD(s).

For non-PvD aware applications, the operating system has policies that determine the authorized set of PvDs and the preferred outgoing PvD. For PvD-aware applications, both the authorized set of PvDs and the default outgoing PvD can be determined as the common subset produced between the OS policies and the set of PvD IDs or characteristics provided by the application.

Application input could be provided on per-application, per-transport-API-object or per-transport-API-call basis. The API for application input may have an option for specifying whether the input should be treated as a preference instead of a requirement.

5.2.3.1. Processing of Incoming Traffic

Unicast IP packets are received on a specific IP address associated with a PvD. For multicast packets, the host can derive the PvD association from other configuration information, such as an explicit PvD property or local policy.

The node OS or middleware may apply more advanced techniques for determining the resultant PvD and / or authorization of the incoming traffic. Those techniques are outside of scope of this document.

If the determined receiving PvD of a packet is not in the allowed subset of PvDs for the particular application / transport API object, the packet should be handled in the same way as if there were no listener.

5.2.3.1.1. Connection-oriented APIs

For connection-oriented APIs, when the initial incoming packet is received, the packet PvD is remembered for the established connection and used for handling of outgoing traffic for that connection. While typically, connection-oriented APIs use a connection-oriented transport protocol, such as TCP, it is possible to have a connection-oriented API that uses a generally connectionless transport protocol, such as UDP.

For APIs/protocols that support multiple IP traffic flows associated with a single transport API connection object (for example, multi path TCP), the processing rules may be adjusted accordingly.

5.2.3.1.2. Connectionless APIs

For connectionless APIs, the host should provide an API that PvD-aware applications can use to query the PvD associated with the packet. For outgoing traffic on this transport API object, the OS should use the selected outgoing PvDs, determined as described above.

5.2.4. Enforcement of Security Policies

By themselves, PvDs do not define, and cannot be used for communication of, security policies. When implemented in a network, this architecture provides the host with information about connected networks. The actual behavior of the host then depends on the host's policies (provisioned through mechanisms out of scope of this document), applied taking received PvD information into account. In some scenarios, e.g. a VPN, such policies could require the host to use only a particular VPN PvD for some / all of the application's traffic (VPN 'disable split tunneling' also known as 'force tunneling' behavior), or apply such restrictions only to selected applications and allow the simultaneous use of the VPN PvD together with the other connected PvDs by the other or all applications (VPN 'split tunneling' behavior).

5.3. Connectivity Tests

Although some PvDs may appear as valid candidates for PvD selection (e.g. good link quality, consistent connection parameters, etc.), they may provide limited or no connectivity to the desired network or the Internet. For example, some PvDs provide limited IP connectivity (e.g., scoped to the link or to the access network), but require the node to authenticate through a web portal to get full access to the Internet. This may be more likely to happen for PvDs which are not trusted by a given PvD-aware node.

An attempt to use such a PvD may lead to limited network connectivity or application connection failures. To prevent the latter, a PvD-aware node may perform a connectivity test for the PvD before using it to serve application network connection requests. In current implementations, some nodes already implement this e.g., by trying to reach a dedicated web server (see [RFC6419]).

Section 5.2 describes how a PvD-aware node shall maintain and use multiple PvDs separately. The PvD-aware node shall perform a connectivity test and, only after validation of the PvD, consider using it to serve application connections requests. Ongoing connectivity tests are also required, since during the IP session, the end-to-end connectivity could be disrupted for various reasons (e.g. L2 problems, IP QoS issues); hence, a connectivity monitoring function is needed to check the connectivity status and remove the PvD from the set of usable PvDs if necessary.

There may be cases where a connectivity test for PvD selection may not be appropriate and should be complemented, or replaced, by PvD selection based on other factors. For example, this could be realized by leveraging some 3GPP and IEEE mechanisms, which would allow the exposure of some PvD characteristics to the node (e.g. 3GPP Access Network Discovery and Selection Function (ANDSF) [TS23402], IEEE 802.11u [IEEE802.11u]/ANQP).

5.4. Relationship to Interface Management and Connection Managers

Current devices, such as mobile handsets make use of proprietary mechanisms and custom applications to manage connectivity in environments with multiple interfaces and multiple sets of network configuration. These mechanisms or applications are commonly known as connection managers [RFC6419].

Connection managers sometimes rely on policy servers to allow a node that is connected to multiple networks to perform network selection. They can also make use of routing guidance from the network (e.g. 3GPP ANDSF [TS23402]). Although connection managers solve some connectivity problems, they rarely address network selection problems in a comprehensive manner. With proprietary solutions, it is challenging to present coherent behavior to the end user of the device, as different platforms present different behaviors even when connected to the same network, with the same type of interface, and for the same purpose. The architecture described in this document should improve the hosts behavior by providing the hosts with tools and guidance to make informed network selection decisions.

6. PvD support in APIs

For all levels of PvD support in APIs described in this chapter, it is expected that the notifications about changes in the set of available PvDs are exposed as part of the API surface.

6.1. Basic

Applications are not PvD-aware in any manner and only submit connection requests. The node performs PvD selection implicitly, without any application participation, based purely on node-specific administrative policies and / or choices made by the user from a user interface provided by the operating environment, not by the application.

As an example, PvD selection can be done at the name service lookup step by using the relevant configuration elements, such as those described in [RFC6731]. As another example, PvD selection could be made based on application identity or type (i.e., a node could always use a particular PvD for a VOIP application).

6.2. Intermediate

Applications indirectly participate in PvD selection by specifying hard requirements and soft preferences. As an example, a real time communication application intending to use the connection for the exchange of real time audio / video data may indicate a preference or a requirement for connection quality, which could affect PvD selection (different PvDs could correspond to Internet connections with different loss rates and latencies).

Another example is the connection of an infrequently executed background activity, which checks for application updates and performs large downloads when updates are available. For such connections, a cheaper or zero cost PvD may be preferable, even if such a connection has a higher relative loss rate or lower bandwidth. The node performs PvD selection based on applications' inputs and policies and / or user preferences. Some / all properties of the resultant PvD may be exposed to applications.

6.3. Advanced

PvDs are directly exposed to applications for enumeration and selection. Node polices and / or user choices may still override the applications' preferences and limit which PvD(s) can be enumerated and / or used by the application, irrespective of any preferences which the application may have specified. Depending on the implementation, such restrictions (imposed by node policy and / or user choice) may or may not be visible to the application.

7. PvD Trust for PvD-Aware Node

7.1. Untrusted PvDs

Implicit and explicit PvDs for which no trust relationship exists are considered untrusted. Only PvDs which meet the requirements in Section 7.2 are trusted; any other PvD is untrusted.

In order to avoid the various forms of misinformation that could occur when PvDs are untrusted, nodes that implement PvD separation cannot assume that two explicit PvDs with the same identifier are actually the same PvD. A node that makes this assumption will be vulnerable to attacks where, for example, an open Wifi hotspot might assert that it was part of another PvD and thereby attempt to draw traffic intended for that PvD onto its own network.

Since implicit PvD identifiers are synthesized by the node, this issue cannot arise with implicit PvDs.

Mechanisms exist (for example, [RFC6731]) whereby a PvD can provide configuration information that asserts special knowledge about the reachability of resources through that PvD. Such assertions cannot be validated unless the node has a trust relationship with the PvD; therefore, assertions of this type must be ignored by nodes that receive them from untrusted PvDs. Failure to ignore such assertions could result in traffic being diverted from legitimate destinations to spoofed destinations.

7.2. Trusted PvDs

Trusted PvDs are PvDs for which two conditions apply: First, a trust relationship must exist between the node that is using the PvD configuration and the source that provided that configuration; this is the authorization portion of the trust relationship. Second, there must be some way to validate the trust relationship. This is the authentication portion of the trust relationship. Two mechanisms for validating the trust relationship are defined.

It shall be possible to validate the trust relationship for all advertised elements of a trusted PvD, irrespective of whether the PvD elements are communicated as a whole, e.g., in a single DHCP option, or separately, e.g., in supplementary RA options. The feasibility of mechanisms to implement a trust relationship for all PvD elements will be determined in the respective companion design documents.

7.2.1. Authenticated PvDs

One way to validate the trust relationship between a node and the source of a PvD is through the combination of cryptographic authentication and an identifier configured on the node. In some cases, the two could be the same; for example, if authentication is by a shared secret, the secret would have to be associated with the PvD identifier. Without a PvD Identifier / shared key tuple, authentication would be impossible, and hence authentication and authorization are combined.

However, if authentication is done using a public key mechanism such as a TLS certificate or DANE, authentication by itself is not enough since theoretically any PvD could be authenticated in this way. In addition to authentication, the node would need configuration to trust the identifier being authenticated. Validating the authenticated PvD name against a list of PvD names configured as trusted on the node would constitute the authorization step in this case.

7.2.2. PvDs Trusted by Attachment

In some cases, a trust relationship may be validated by some means other than those described in Section 7.2.1 simply by virtue of the connection through which the PvD was obtained. For instance, a handset connected to a mobile network may know through the mobile network infrastructure that it is connected to a trusted PvD. Whatever mechanism was used to validate that connection constitutes the authentication portion of the PvD trust relationship. Presumably, such a handset would be configured from the factory (or else through mobile operator or user preference settings) to trust the PvD, and this would constitute the authorization portion of this type of trust relationship.

8. Contributors

The following individuals contributed to this document (listed in no specific order): Alper Yegin (alper.yegin@yegin.org), Aaron Yi Ding (yding@cs.helsinki.fi), Zhen Cao (caozhenpku@gmail.com), Dapeng Liu (liudapeng@chinamobile.com), Dave Thaler (dthaler@microsoft.com), Dmitry Anipko (dmitry.anipko@microsoft.com), Hui Deng (denghui@chinamobile.com), Jouni Korhonen (jouni.nospam@gmail.com), Juan Carlos Zuniga (JuanCarlos.Zuniga@InterDigital.com), Konstantinos Pentikousis (k.pentikousis@huawei.com), Marc Blanchet (marc.blanchet@viagenie.ca), Margaret Wasserman (margaretw42@gmail.com), Pierrick Seite (pierrick.seite@orange.com), Suresh Krishnan (suresh.krishnan@ericsson.com), Teemu Savolainen (teemu.savolainen@nokia.com), Ted Lemon (ted.lemon@nominum.com) and Tim Chown (tjc@ecs.soton.ac.uk).

9. Acknowledgments

The authors would like to thank (in no specific order) Ian Farrer, Marcus Stenberg and Mikael Abrahamsson for their review and comments.

10. IANA Considerations

This memo does not include any IANA requests.

11. Security Considerations

There are at least three different forms of attacks that can be performed using configuration sources that support multiple provisioning domains.

Tampering with provided configuration information:

An attacker may attempt to modify information provided inside the PvD container option. These attacks can easily be prevented by using message integrity features provided by the underlying protocol used to carry the configuration information. E.g. SEND [RFC3971] would detect any form of tampering with the RA contents and the DHCPv6 [RFC3315] AUTH option that would detect any form of tampering with the DHCPv6 message contents. This attack can also be performed by a compromised configuration source by modifying information inside a specific PvD, in which case the mitigations proposed in the next subsection may be helpful.

Rogue configuration source:

A compromised configuration source, such as a router or a DHCPv6 server, may advertise information about PvDs that it is not authorized to advertise. e.g. A coffee shop WLAN may advertise configuration information purporting to be from an enterprise and may try to attract enterprise related traffic. The only real way to prevent this is is for the PvD related configuration container to contain embedded authentication and authorization information from the owner of the PvD. This provides the client with a way of detecting the attack by verifying the authentication and authorization information provided inside the PvD container option, after verifying its trust of the PvD owner (e.g. a certificate with a well-known / common trust anchor).

Replay attacks:

A compromised configuration source or an on-link attacker may try to capture advertised configuration information and replay it on a different link, or at a future point in time. This can be avoided by including a replay protection mechanism such as a timestamp or a nonce inside the PvD container to ensure the validity of the provided information.

12. References

12.1. Normative References

[RFC2119]

Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997.

12.2. Informative References

[I-D.bhandari-dhc-class-based-prefix]	Systems, C., Halwasia, G., Gundavelli, S., Deng, H., Thiebaut, L. and J. Korhonen, "DHCPv6 class based prefix", Internet-Draft draft-bhandari-dhc-class-based-prefix-04, February 2013.
[I-D.korhonen-dmm-prefix-properties]	Korhonen, J., Patil, B., Gundavelli, S., Seite, P. and D. Liu, "IPv6 Prefix Mobility Management Properties", Internet-Draft draft-korhonen-dmm-prefix-properties-03, October 2012.
[IEEE802.11u]	IEEE, "IEEE Standard 802.11u-2011 (Amendment 9: Interworking with External Networks)", 2011.
[RFC3315]	Droms, R., Bound, J., Volz, B., Lemon, T., Perkins, C. and M. Carney, "Dynamic Host Configuration Protocol for IPv6 (DHCPv6)", RFC 3315, July 2003.
[RFC3971]	Arkko, J., Kempf, J., Zill, B. and P. Nikander, "SEcure Neighbor Discovery (SEND)", RFC 3971, March 2005.
[RFC4861]	Narten, T., Nordmark, E., Simpson, W. and H. Soliman, "Neighbor Discovery for IP version 6 (IPv6)", RFC 4861, September 2007.
[RFC5739]	Eronen, P., Laganier, J. and C. Madson, "IPv6 Configuration in Internet Key Exchange Protocol Version 2 (IKEv2)", RFC 5739, February 2010.
[RFC5996]	Kaufman, C., Hoffman, P., Nir, Y. and P. Eronen, "Internet Key Exchange Protocol Version 2 (IKEv2)", RFC 5996, September 2010.
[RFC6182]	Ford, A., Raiciu, C., Handley, M., Barre, S. and J. Iyengar, "Architectural Guidelines for Multipath TCP Development", RFC 6182, March 2011.
[RFC6418]	Blanchet, M. and P. Seite, "Multiple Interfaces and Provisioning Domains Problem Statement", RFC 6418, November 2011.
[RFC6419]	Wasserman, M. and P. Seite, "Current Practices for Multiple-Interface Hosts", RFC 6419, November 2011.
[RFC6555]	Wing, D. and A. Yourtchenko, "Happy Eyeballs: Success with Dual-Stack Hosts", RFC 6555, April 2012.
[RFC6724]	Thaler, D., Draves, R., Matsumoto, A. and T. Chown, "Default Address Selection for Internet Protocol Version 6 (IPv6)", RFC 6724, September 2012.
[RFC6731]	Savolainen, T., Kato, J. and T. Lemon, "Improved Recursive DNS Server Selection for Multi-Interfaced Nodes", RFC 6731, December 2012.
[RFC7078]	Matsumoto, A., Fujisaki, T. and T. Chown, "Distributing Address Selection Policy Using DHCPv6", RFC 7078, January 2014.
[TS23402]	3GPP, "3GPP TS 23.402; Architecture enhancements for non-3GPP accesses; release 12", 2014.

Author's Address