51                 13                    64-V           ?
+------------------------+---------+----------------------------+---+
|    (base scheme)       | Zone ID |         Device ID          | V |
+------------------------+---------+----------------------------+---+
		

Figure 4: Addressing Scheme 2

The fields are defined as follows: [Editor's note: The lengths of the fields is for discussion.]

• Zone ID: As in sub-scheme 1.
• Device ID: As in sub-scheme 1.
• V: Virtualization bit(s): 1 or more bits that indicate a virtual context on an autonomic node.

In addition the scheme 1 (Section 5.6.3.1), this scheme allows the direct addressing of specific virtual containers / VMs on an autonomic node. An increasing number of hardware platforms have a distributed architecture, with a base OS for the node itself, and the support for hardware blades with potentially different OSs. The VMs on the blades could be considered as separate autonomic nodes, in which case it would make sense to be able to address them directly. Autonomic Service Agents (ASAs) could be instantiated in either the base OS, or one of the VMs on a blade. This addressing scheme allows for the easy separation of the hardware context.

The location of the V bit(s) at the end of the address allows to announce a single prefix for each autonomic node, while having separate virtual contexts addressable directly.

5.6.4. Usage of the Zone Field

The "zone ID" allows for the introduction of structure in the addressing scheme.

Zone = zero is the default addressing scheme in an autonomic domain. Every autonomic node MUST respond to its ACP address with zone=0. Used on its own this leads to a non-hierarchical address scheme, which is suitable for networks up to a certain size. In this case, the addresses primarily act as identifiers for the nodes, and aggregation is not possible.

If aggregation is required, the 13 bit value allows for up to 8191 zones. The allocation of zone numbers may either happen automatically through a to-be-defined algorithm; or it could be configured and maintained manually. [We could divide the zone space into manual and automatic space - to be discussed.]

If a device learns through an autonomic method or through configuration that it is part of a zone, it MUST also respond to its ACP address with that zone number. In this case the ACP loopback is configured with two ACP addresses: One for zone 0 and one for the assigned zone. This method allows for a smooth transition between a flat addressing scheme and an hierarchical one.

(Theoretically, the 13 bits for the zone ID would allow also for two levels of zones, introducing a sub-hierarchy. We do not think this is required at this point, but a new type could be used in the future to support such a scheme.)

Note: Another way to introduce hierarchy is to use sub-domains in the naming scheme. The node names "node17.subdomainA.example.com" and "node4.subdomainB.example.com" would automatically lead to different ULA prefixes, which can be used to introduce a routing hierarchy in the network, assuming that the subdomains are aligned with routing areas.

5.7. Routing in the ACP

Once ULA address are set up all autonomic entities should run a routing protocol within the autonomic control plane context. This routing protocol distributes the ULA created in the previous section for reachability. The use of the autonomic control plane specific context eliminates the probable clash with the global routing table and also secures the ACP from interference from the configuration mismatch or incorrect routing updates.

The establishment of the routing plane and its parameters are automatic and strictly within the confines of the autonomic control plane. Therefore, no manual configuration is required.

All routing updates are automatically secured in transit as the channels of the autonomic control plane are by default secured.

The routing protocol inside the ACP should be light weight and highly scalable to ensure that the ACP does not become a limiting factor in network scalability. We suggest the use of RPL [RFC6550] as one such protocol which is light weight and scales well for the control plane traffic. See Appendix A for more details on the choice of RPL.

6. Workarounds for Non-Autonomic Nodes

6.1. Connecting a Non-Autonomic Controller / NMS system

The Autonomic Control Plane can be used by management systems, such as controllers or network management system (NMS) hosts (henceforth called simply "NMS hosts"), to connect to devices through it. For this, an NMS host must have access to the ACP. By default, the ACP is a self-protecting overlay network, which only allows access to trusted systems. Therefore, a traditional, non-autonomic NMS system does not have access to the ACP by default, just like any other external device.

If the NMS host is not autonomic, i.e., it does not support autonomic negotiation of the ACP, then it can be brought into the ACP by explicit configuration. On an adjacent autonomic node with ACP, the interface with the NMS host can be configured to be part of the ACP. In this case, the NMS host is with this interface entirely and exclusively inside the ACP. It would likely require a second interface for connections between the NMS host and administrators, or Internet based services. This mode of connecting an NMS host has security consequences: All systems and processes connected to this implicitly trusted interface have access to all autonomic nodes on the entire ACP, without further authentication. Thus, this connection must be physically controlled.

The non-autonomic NMS host must be routed in the ACP. This involves two parts: 1) the NMS host must point default to the AN device for the ULA prefix used inside the ACP, and 2) the prefix used between AN node and NMS host must be announced into the ACP, and distributed there.

The document "Autonomic Network Stable Connectivity" [I-D.eckert-anima-stable-connectivity] explains in more detail how the ACP can be integrated in a mixed NOC environment.

6.2. ACP through Non-Autonomic L3 Clouds

Not all devices in a network may be autonomic. If non-autonomic Layer-2 devices are between autonomic nodes, the communications described in this document should work, since it is IP based. However, non-autonomic Layer-3 devices do not forward link local autonomic messages, and thus break the Autonomic Control Plane.

One workaround is to manually configure IP tunnels between autonomic nodes across a non-autonomic Layer-3 cloud. The tunnels are represented on each autonomic node as virtual interfaces, and all autonomic transactions work across such tunnels.

Such manually configured tunnels are less "indestructible" than an automatically created ACP based on link local addressing, since they depend on correct data plane operations, such as routing and addressing.

7. Building the ACP

7.1. Neighbor discovery via GRASP

Autonomic devices use inscure GRASP to discovery candidate autonomic neighbors across L2 adjacencies. When Alice discovers Bob:

• If Alice is not part an autonomic domain, she starts autonomic enrollment with Bob as the proxy using procedures described in [I-D.ietf-anima-bootstrapping-keyinfra].
• If Alice is part of an autonomic domain, Alice attempts to build the ACP to Bob. Bob will do the same.

7.2. Channel Selection

To avoid attacks, initial discovery of candidate ACP peers can not include any non-protected negotiation. To avoid re-inventing and validating security association mechanisms, the next step after discoving the address of a candidate neighbor can only be to try first to establish a security association with that neighbor using a well-known security association method.

At this time in the lifecycle of autonomic devices, it is unclear whether it is feasible to even decide on a single MTI (mandatory to implement) security association protocol across all autonomic devices.

From the use-cases it is clear that not all type of autonomic devices can or need to connect directly to each other or are able to support or prefer all possible mechanisms. For example, code space limited IoT devices may only support dTLS (because that code exists already on them for end-to-end security use-cases), but low-end in-ceiling L2 switches may only want to support MacSec because that is also supported in HW, and only a more flexible garteway device may need to support both of these mechanisms and potentially more.

To support these requirements, and make ACP channel negotiation also easily extensible, the secure channel selection between Alice and Bob is a potentially two stage process. In the first stage, Alice and Bob directly try to establish a secure channel using the security-association and channel protocols they support. One or more of these protocols may simply be protocols not directly resulting in an ACP channel, but instead establishing a secure negotiation channel through which the final secure channel protocol is decided. If both Alice and Bob support such a negotiation step, then this secured negotiation channel is the first step, and the secure channel protocol is the second step.

If Alice supports multiple security association protocols in the first step, it is a matter of Alices local policy to determine the order in which she will try to build the connection to Bob. To support multiple security association protocols, Alice must be able to simultaneously act as a responder in parallel for all of them - so that she can respond to any order in which Bob wants to prefer building the security association.

When ACP setup between Alice and Bob results in the first secure association to be established, the peer with the higher Device-ID in the certificate will stop building new security associations. The peer with the lower certificate Device-ID is now responsible to continue building its most preferred security association and to tear down all but that most preferred one - unless the secure association is one of a negotation protocols whose rules superceed this.

All this negotiation is in the context of an "L2 interface". Alice and Bob will build ACP connections to each other on every "L2 interface" that they both connect to.

7.3. Security Association protocols

The following sections define the security association protocols that we consider to be important and feasible to specify in this document. In all cases, the mutual authentication is done via the autonomic domain certificate of the peer as follows - unless superceeded by intent:

• The certificate is valid as proven by the security associations protocol exchanges.
• The peers certificate is signed by the same CA as the devices domain certificate.
• The peers OU (Organizational Unit) field in the certificates Subject is the same as in the devices certificate.

7.3.1. ACP via IPsec

To run ACP via IPsec transport mode, no further IANA assignments/definitions are required. All autonomic devices suppoting IPsec MUST support IPsec security setup via IKEv2, transpoort mode encapsulation via the device and peer link-local IPv6 addresses and AES256 encryption. Further parameter options can be negotiated via IKEv2 or via GRASP/TLS.

7.3.2. ACP via GRE/IPsec

In network devices it is often easier to provide virtual interfaces on top of GRE encapsulation than natively on top of a simple IPsec association. On those devices it may be necessary to run the ACP secure channel on top of a GRE connection protected by the IPsec association. The requirements for the IPsec association are the same as described above, but instead of directly carrying the ACP IPv6 packets, the payload is an ACP IPv6 packet inside GREP/IPv6.

Note that without explicit negotiation (eg: via GRASP/TLS), this method is incompatible to direct ACP via IPsec, so it must only be used as an option during GRASP/TLS negotiation.

7.3.3. ACP via dTLS

To run ACP via UDP and dTLS v1.2 [RFC6347] an IANA assigned port [TBD] is used. All autonomic devices supporting ACP via dTLS must support AES256 encryption.

7.3.4. GRASP/TLS negotiation

To explicitly allow negotiation of the ACP channel protocol, GRASP over a TLS connection using the GRASP_LISTEN_PORT and the devices and peers link-local IPv6 address is used. When Alice and Bob support GRASP negotiation, they do prefer it over any other non-explicitly negotiated security association protocol and should wait trying any non-negotiated ACP channel protocol until after it is clear that GRASP/TLS will not work to the peer.

When Alice and Bob successfully establish the GRASP/TSL session, they will initially negotiate the channel mechanism to use. Bob and Alice each have a list of channel mehanisms they support, sorted by preference. They negotiate the best mechansm supported by both of them. In the absence of Intent, the mechanism choosen is the best one for the device with the lower Device-ID.

After agreeing on a channel mechanism, Alice and Bob start the selected Channel protocol. The GRASP/TLS connection can be kept alive or timed out as long as the seelected channel protocol has a secure association between Alice and Bob. When it terminates, it needs to be re-negotiated via GRASP/TLS.

Negotiation of a channel type may require IANA assignments of code points. See IANA Considerations [iana] for the formal definition of those code points.

TBD: The exact negotiation steps in GRASP to achieve this outcome.

7.3.5. ACP Security Profiles

A baseline autonomic device MUST support IPsec and SHOULD support GRASP/TLS and dTLS. A constrained autonomic device MUST support dTLS.

Autonomic devices need to specify in documentation the set of secure ACP mechanisms they suppport.

7.4. GRASP instance details

GRASP run to (insecurely) discover autonomic neighbors are isolated instances from each other and other uses of GRASP - GRASP/TLS sessions of L2 interfaces and GRASP inside the ACP

Received GRASP packets are assigned to an instance of GRASP by the context they are received on:

• GRASP packets received on an ACP (virtual) interfaces are assigned to the ACP instance of GRASP
• GRASP/UDP packets received on L2 interfaces where the device is willing to run ACP across are are assigned to a separate instance of GRASP for that L2 interface. We call those instances of GRASP the "insecure L2 GRASP instances" and the ASA to perform the discovery the "insecure L2 discovery ASA" (IL2D).
• GRASP packets received inside a TLS connection established for GRASP/TLS ACP negotiation are assigned to a separate instance of GRASP for that negotiation

All insecure L2 discovery of candidate ACP neighbors via GRASP and the potentially following GRASP/TLS negotiation is per-L2 interface: If Alice and Bob connect to each other via multiple interfaces, they will independently establish the ACP to each other across each of these interfaces.

For every L2-discovery instance of GRASP and its IL2D, the following rules apply, amending and overriding the recommendations in [I-D.ietf-anima-grasp]:

• GRASP link-local multicast discovery messages MUST use GTSM [RFC5082]. With GTSM, discovery packets are sent with a TTL of 255 and packets received with a TTL smaller than 255 are ignored upon receipt.
• The GRASP loop count of GRASP discovery packets must be set to 1 on sending.
• GRASP MUST send response messages for the discovery objected defined here (overriding the MAY).
• GRASP MUST NOT respond to discovery objectives with the Divert option - objectives learned and cached are solely for local consumption.
• GRASP MUST NOT relay discovery or any other messages across different interfaces.

TBD: The Details of the GRASP objective/packet formats still need to be defined. Eg: Need to define an allocation for the objective of "Autonomic Node".

8. Self-Healing Properties

The ACP is self-healing:

• New neighbors will automatically join the ACP after successful validation and will become reachable using their unique ULA address across the ACP.
• When any changes happen in the topology, the routing protocol used in the ACP will automatically adapt to the changes and will continue to provide reachability to all devices.
• If an existing device gets revoked, it will automatically be denied access to the ACP as its domain certificate will be validated against a Certificate Revocation List during authentication. Since the revocation check is only done at the establishment of a new security association, existing ones are not automatically torn down. If an immediate disconnect is required, existing sessions to a freshly revoked device can be re-set.

The ACP can also sustain network partitions and mergers. Practically all ACP operations are link local, where a network partition has no impact. Devices authenticate each other using the domain certificates to establish the ACP locally. Addressing inside the ACP remains unchanged, and the routing protocol inside both parts of the ACP will lead to two working (although partitioned) ACPs.

There are few central dependencies: A certificate revocation list (CRL) may not be available during a network partition; a suitable policy to not immediately disconnect neighbors when no CRL is available can address this issue. Also, a registrar or Certificate Authority might not be available during a partition. This may delay renewal of certificates that are to expire in the future, and it may prevent the enrolment of new devices during the partition.

After a network partition, a re-merge will just establish the previous status, certificates can be renewed, the CRL is available, and new devices can be enrolled everywhere. Since all devices use the same trust anchor, a re-merge will be smooth.

Merging two networks with different trust anchors requires the trust anchors to mutually trust each other (for example, by cross-signing). As long as the domain names are different, the addressing will not overlap (see Section 5.6).

9. Self-Protection Properties

As explained in Section 5, the ACP is based on channels being built between devices which have been previously authenticated based on their domain certificates. The channels themselves are protected using standard encryption technologies like DTLS or IPsec which provide additional authentication during channel establishment, data integrity and data confidentiality protection of data inside the ACP and in addition, provide replay protection.

An attacker will therefore not be able to join the ACP unless having a valid domain certificate, also packet injection and sniffing traffic will not be possible due to the security provided by the encryption protocol.

The remaining attack vector would be to attack the underlying AN protocols themselves, either via directed attacks or by denial-of-service attacks. However, as the ACP is built using link-local IPv6 address, remote attacks are impossible. The ULA addresses are only reachable inside the ACP context, therefore unreachable from the data plane. Also, the ACP protocols should be implemented to be attack resistant and not consume unnecessary resources even while under attack.

10. The Administrator View

An ACP is self-forming, self-managing and self-protecting, therefore has minimal dependencies on the administrator of the network. Specifically, since it is independent of configuration, there is no scope for configuration errors on the ACP itself. The administrator may have the option to enable or disable the entire approach, but detailed configuration is not possible. This means that the ACP must not be reflected in the running configuration of devices, except a possible on/off switch.

While configuration is not possible, an administrator must have full visibility of the ACP and all its parameters, to be able to do trouble-shooting. Therefore, an ACP must support all show and debug options, as for any other network function. Specifically, a network management system or controller must be able to discover the ACP, and monitor its health. This visibility of ACP operations must clearly be separated from visibility of data plane so automated systems will never have to deal with ACP aspect unless they explicitly desire to do so.

Since an ACP is self-protecting, a device not supporting the ACP, or without a valid domain certificate cannot connect to it. This means that by default a traditional controller or network management system cannot connect to an ACP. See Section 6.1 for more details on how to connect an NMS host into the ACP.

11. Explanations

This section is non-normative and intended to provide further explanations for the choices made in this document.

11.1. Why GRASP to discover autonomic neighbors

None of the considered mechanisms to establish security associations (eg: IPsec or dTLS) include mechanisms to discover candidate neighbors, so these security mechanisms themselves are insufficient for the discovery.

Existing L2 mechanisms such as LLDP (or vendor speccific alternatives like Ciscos CDP) are L2 link-local. If an autonomic device connects via an LLDP capable, but non-autonomic capable L2 switch to another autonomic device, then the non-autonomic L2 switch would not propagate the LLDP messages, so discovery would not work as desired.

Existing L3/L4 link local discovery mechanisms such as mDNS or Web-Services Discovery (http://specs.xmlsoap.org/ws/2005/04/discovery/ws-discovery.pdf) are capable to support the simple discovery required by autonomic devices but have the following downsides compared to GRASP.

There is no clear single ubiquitoous protocol that would apply equally well to all market segments in which autonomic routers are intended to be deployed. Making a choice is therefore difficult.

In some of these protocols, the fact that they operate L3 link local is often seen as a limitation rather than as a necessity for the application.

Various mechanisms are used or considered in these protocols to expand the scope of discovery beyond a single L3 subnet. If autonomic devices would use such a protocol, then autonomic discovery messages could more likely leak into remote networks and give more undesired (insecured) visibility into the presence of autonomic devices and potentially leading to more attempts to establish autonomic associations with those discovered devices. To achieve the maximum resilience with the minimum number of ACP channels, those channels need to follow as closely the physcial hops in the topology as possible.

Visibility of discovery protocols in other domains may be undesirable: Visibility of mDNS messages for example could extend all the way into end user application level service browsers. It is undesirable to see desvices announcing themselves as automic there.

Existing protocols can be more complex compared to GRASP as they have been designed for different purposes, for example to be more flexible and generic. In mDNS, if DNS-SD was used, it would require at least four RRs to be exchanged for a single service: a PTR, a SRV, a TXT and a AAAA RR. Minimizing the number of protocol exchanges by coalescing these RRs is possible but requires additional software design considerations.

GRASP is already required inside the ACP and a highly desirable option for secure ACP channel negotiation (GRASP/TLS). Using it for discovery allows to reuse that already necessary code basis. If any other protocol was used for discovery, then autonomic discovery might be the only purpose for which the protocol code exists in the device.

None of the above arguments individually are strong reasons not to use one of these GRASP alternatives, but together they make it reasonable to first define GRASP as the MTI (Mandatory To Implement) for the discovery step.

12. Security Considerations

An ACP is self-protecting and there is no need to apply configuration to make it secure. Its security therefore does not depend on configuration.

However, the security of the ACP depends on a number of other factors:

• The usage of domain certificates depends on a valid supporting PKI infrastructure. If the chain of trust of this PKI infrastructure is compromised, the security of the ACP is also compromised. This is typically under the control of the network administrator.
• Security can be compromised by implementation errors (bugs), as in all products.

Fundamentally, security depends on correct operation, implementation and architecture. Autonomic approaches such as the ACP largely eliminate the dependency on correct operation; implementation and architectural mistakes are still possible, as in all networking technologies.

13. IANA Considerations

Section 7.3.3 describes ACP over dTLS, which requires a well-known UDP port. We request IANA to assign this UDP port for 'ACP over dTLS'.

Section 7.3.4 describes an option for the channel negotiation, the 'ACP channel type'. We request IANA to create a registry for 'ACP channel type'.

The ACP channel type is a 8-bit unsigned integer. This document only assigns the first value.

     Number | Channel Type                      | RFC
   ---------+-----------------------------------+------------
         0  | GRE tunnel protected with         | this document
            | IPsec transport mode              |
     1-255  | reserved for future channel types | 

Section 5.6.2 describes a 'type' field in the base addressing scheme. We request IANA to create a registry for the 'ACP addressing scheme type'. The initial value and definition will be determined in a later version of this document.

14. Acknowledgements

This work originated from an Autonomic Networking project at Cisco Systems, which started in early 2010. Many people contributed to this project and the idea of the Autonomic Control Plane, amongst which (in alphabetical order): Ignas Bagdonas, Parag Bhide, Alex Clemm, Yves Hertoghs, Bruno Klauser, Max Pritikin, Ravi Kumar Vadapalli.

Further input and suggestions were received from: Rene Struik, Brian Carpenter, Benoit Claise.

15. Change log [RFC Editor: Please remove]

15.1. Initial version

First version of this document: [I-D.behringer-autonomic-control-plane]

15.2. draft-behringer-anima-autonomic-control-plane-00

Initial version of the anima document; only minor edits.

15.3. draft-behringer-anima-autonomic-control-plane-01

• Clarified that the ACP should be based on, and support only IPv6.
• Clarified in intro that ACP is for both, between devices, as well as for access from a central entity, such as an NMS.
• Added a section on how to connect an NMS system.
• Clarified the hop-by-hop crypto nature of the ACP.
• Added several references to GDNP as a candidate protocol.
• Added a discussion on network split and merge. Although, this should probably go into the certificate management story longer term.

15.4. draft-behringer-anima-autonomic-control-plane-02

Addresses (numerous) comments from Brian Carpenter. See mailing list for details. The most important changes are:

• Introduced a new section "overview", to ease the understanding of the approach.
• Merged the previous "problem statement" and "use case" sections into a mostly re-written "use cases" section, since they were overlapping.
• Clarified the relationship with draft-eckert-anima-stable-connectivity

15.5. draft-behringer-anima-autonomic-control-plane-03

• Took out requirement for IPv6 --> that's in the reference doc.
• Added requirement section.
• Changed focus: more focus on autonomic functions, not only virtual out of band. This goes a bit throughout the document, starting with a changed abstract and intro.

15.6. draft-ietf-anima-autonomic-control-plane-00

No changes; re-submitted as WG document.

15.7. draft-ietf-anima-autonomic-control-plane-01

• Added some paragraphs in addressing section on "why IPv6 only", to reflect the discussion on the list.
• Moved the data-plane ACP out of the main document, into an appendix. The focus is now the virtually separated ACP, since it has significant advantages, and isn't much harder to do.
• Changed the self-creation algorithm: Part of the initial steps go into the reference document. This document now assumes an adjacency table, and domain certificate. How those get onto the device is outside scope for this document.
• Created a new section 6 "workarounds for non-autonomic nodes", and put the previous controller section (5.9) into this new section. Now, section 5 is "autonomic only", and section 6 explains what to do with non-autonomic stuff. Much cleaner now.
• Added an appendix explaining the choice of RPL as a routing protocol.
• Formalised the creation process a bit more. Now, we create a "candidate peer list" from the adjacency table, and form the ACP with those candidates. Also it explains now better that policy (Intent) can influence the peer selection. (section 4 and 5)
• Introduce a section for the capability negotiation protocol (section 7). This needs to be worked out in more detail. This will likely be based on GRASP.
• Introduce a new parameter: ACP tunnel type. And defines it in the IANA considerations section. Suggest GRE protected with IPSec transport mode as the default tunnel type.
• Updated links, lots of small edits.

15.8. draft-ietf-anima-autonomic-control-plane-02

• Added explicitly text for the ACP channel negotiation.
• Merged draft-behringer-anima-autonomic-addressing-02 into this document, as suggested by WG chairs.

16. References

[I-D.behringer-anima-autonomic-addressing]	Behringer, M., "An Autonomic IPv6 Addressing Scheme", Internet-Draft draft-behringer-anima-autonomic-addressing-02, October 2015.
[I-D.behringer-anima-reference-model]	Behringer, M., Carpenter, B., Eckert, T., Ciavaglia, L., Liu, B., Jeff, J. and J. Strassner, "A Reference Model for Autonomic Networking", Internet-Draft draft-behringer-anima-reference-model-04, October 2015.
[I-D.behringer-autonomic-control-plane]	Behringer, M., Bjarnason, S., BL, B. and T. Eckert, "An Autonomic Control Plane", Internet-Draft draft-behringer-autonomic-control-plane-00, June 2014.
[I-D.eckert-anima-stable-connectivity]	Eckert, T. and M. Behringer, "Using Autonomic Control Plane for Stable Connectivity of Network OAM", Internet-Draft draft-eckert-anima-stable-connectivity-02, October 2015.
[I-D.ietf-anima-bootstrapping-keyinfra]	Pritikin, M., Richardson, M., Behringer, M. and S. Bjarnason, "Bootstrapping Key Infrastructures", Internet-Draft draft-ietf-anima-bootstrapping-keyinfra-02, March 2016.
[I-D.ietf-anima-grasp]	Bormann, C., Carpenter, B. and B. Liu, "A Generic Autonomic Signaling Protocol (GRASP)", Internet-Draft draft-ietf-anima-grasp-04, March 2016.
[RFC4122]	Leach, P., Mealling, M. and R. Salz, "A Universally Unique IDentifier (UUID) URN Namespace", RFC 4122, DOI 10.17487/RFC4122, July 2005.
[RFC4193]	Hinden, R. and B. Haberman, "Unique Local IPv6 Unicast Addresses", RFC 4193, DOI 10.17487/RFC4193, October 2005.
[RFC5082]	Gill, V., Heasley, J., Meyer, D., Savola, P. and C. Pignataro, "The Generalized TTL Security Mechanism (GTSM)", RFC 5082, DOI 10.17487/RFC5082, October 2007.
[RFC6347]	Rescorla, E. and N. Modadugu, "Datagram Transport Layer Security Version 1.2", RFC 6347, DOI 10.17487/RFC6347, January 2012.
[RFC6550]	Winter, T., Thubert, P., Brandt, A., Hui, J., Kelsey, R., Levis, P., Pister, K., Struik, R., Vasseur, JP. and R. Alexander, "RPL: IPv6 Routing Protocol for Low-Power and Lossy Networks", RFC 6550, DOI 10.17487/RFC6550, March 2012.
[RFC7404]	Behringer, M. and E. Vyncke, "Using Only Link-Local Addressing inside an IPv6 Network", RFC 7404, DOI 10.17487/RFC7404, November 2014.
[RFC7575]	Behringer, M., Pritikin, M., Bjarnason, S., Clemm, A., Carpenter, B., Jiang, S. and L. Ciavaglia, "Autonomic Networking: Definitions and Design Goals", RFC 7575, DOI 10.17487/RFC7575, June 2015.
[RFC7576]	Jiang, S., Carpenter, B. and M. Behringer, "General Gap Analysis for Autonomic Networking", RFC 7576, DOI 10.17487/RFC7576, June 2015.

Appendix A. Background on the choice of routing protocol

In a pre-standard implementation, the "IPv6 Routing Protocol for Low-Power and Lossy Networks (RPL, [RFC6550] was chosen. This Appendix explains the reasoning behind that decision.

Requirements for routing in the ACP are:

• Self-management: The ACP must build automatically, without human intervention. Therefore routing protocol must also work completely automatically. RPL is a simple, self-managing protocol, which does not require zones or areas; it is also self-configuring, since configuration is carried as part of the protocol (see Section 6.7.6 of [RFC6550]).
• Scale: The ACP builds over an entire domain, which could be a large enterprise or service provider network. The routing protocol must therefore support domains of 100,000 nodes or more, ideally without the need for zoning or separation into areas. RPL has this scale property. This is based on extensive use of default routing. RPL also has other scalability improvements, such as selecting only a subset of peers instead of all possible ones, and trickle support for information synchronisation.
• Low resource consumption: The ACP supports traditional network infrastructure, thus runs in addition to traditional protocols. The ACP, and specifically the routing protocol must have low resource consumption both in terms of memory and CPU requirements. Specifically, at edge nodes, where memory and CPU are scarce, consumption should be minimal. RPL builds a destination-oriented directed acyclic graph (DODAG), where the main resource consumption is at the root of the DODAG. The closer to the edge of the network, the less state needs to be maintained. This adapts nicely to the typical network design. Also, all changes below a common parent node are kept below that parent node.
• Support for unstructured address space: In the Autonomic Networking Infrastructure, node addresses are identifiers, and may not be assigned in a topological way. Also, nodes may move topologically, without changing their address. Therefore, the routing protocol must support completely unstructured address space. RPL is specifically made for mobile ad-hoc networks, with no assumptions on topologically aligned addressing.
• Modularity: To keep the initial implementation small, yet allow later for more complex methods, it is highly desirable that the routing protocol has a simple base functionality, but can import new functional modules if needed. RPL has this property with the concept of "objective function", which is a plugin to modify routing behaviour.
• Extensibility: Since the Autonomic Networking Infrastructure is a new concept, it is likely that changes in the way of operation will happen over time. RPL allows for new objective functions to be introduced later, which allow changes to the way the routing protocol creates the DAGs.
• Multi-topology support: It may become necessary in the future to support more than one DODAG for different purposes, using different objective functions. RPL allow for the creation of several parallel DODAGs, should this be required. This could be used to create different topologies to reach different roots.
• No need for path optimisation: RPL does not necessarily compute the optimal path between any two nodes. However, the ACP does not require this today, since it carries mainly non-delay-sensitive feedback loops. It is possible that different optimisation schemes become necessary in the future, but RPL can be expanded (see point "Extensibility" above).

Appendix B. Alternative: An ACP without Separation

Section 5 explains how the ACP is constructed as a virtually separated overlay network. An alternative ACP design can be achieved without the VRFs. In this case, the autonomic virtual addresses are part of the data plane, and subject to routing, filtering, QoS, etc on the data plane. The secure tunnels are in this case used by traffic to and from the autonomic address space. They are still required to provide the authentication function for all autonomic packets.

At IETF 93 in Prague, the suggestion was made to not advance with the data plane ACP, and only continue with the virtually separate ACP. The reason for this decision is that the contextual separation of the ACP provides a range of benefits (more robustness, less potential interactions with user configurations), while it is not much harder to achieve.

This appendix serves to explain the decision; it will be removed in the next version of the draft.

Authors' Addresses