Internet-Draft M. Lind TeliaSonera Expires : May 2004 V. Ksinant 6WIND D. Park Samsung Electronics A. Baudot France Telecom December 2003 Scenarios and Analysis for Introducing IPv6 into ISP Networks Status of this Memo This document is an Internet-Draft and is in full conformance with all provisions of Section 10 of RFC2026. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet- Drafts. Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress". The list of current Internet-Drafts can be accessed at http://www.ietf.org/ietf/1id-abstracts.txt The list of Internet-Draft Shadow Directories can be accessed at http://www.ietf.org/shadow.html. Abstract This document first describes different scenarios for the introduction of IPv6 into an existing IPv4 ISP network without disrupting the IPv4 service. Then, this document analyses these scenarios and evaluates the suitability of the already defined transition mechanisms in this context. Known challenges are also identified. Lind, et al. Expires - May 2004 [Page 1] Internet-Draft Introducing IPv6 in ISP Networks December 2003 Table of Contents 1. Introduction..................................................3 1.1 Goal and scope of the document..........................3 1.2 Terminology used........................................3 2. Brief description of a generic ISP network....................4 3. Transition scenarios..........................................6 3.1 Identification of scenarios.............................6 3.1.1 Assumptions............................................6 3.1.2 Customer requirements and ISP offerings................7 3.1.3 Stage 1 Scenarios: Launch..............................8 3.1.4 Stage 2a Scenarios: Backbone...........................8 3.1.5 Stage 2b Scenarios: Customer connection................8 3.1.6 Stage 3 scenarios: Complete............................9 3.1.7 Stage 2a and 3 combination scenarios...................9 3.2 Transition Scenarios....................................9 3.3 Actions needed when deploying IPv6 in an ISP network...10 4. Backbone transition actions..................................11 4.1 Steps in transitioning backbone networks...............11 4.2 Configuration of backbone equipment....................13 4.3 Routing................................................13 4.3.1 IGP...................................................13 4.3.2 EGP...................................................14 4.3.3 Routing protocols transport...........................15 4.4 Multicast..............................................15 5. Customer connection transition actions.......................15 5.1 Steps in transitioning customer connection networks....15 5.2 Access control requirements............................17 5.3 Configuration of customer equipment....................17 5.4 Requirements for Traceability..........................18 5.5 Multi-homing...........................................18 5.6 Ingress filtering in the customer connection network...19 6. Network and service operation actions........................19 7. Future Stages................................................20 8. Example networks.............................................20 8.1 Example 1..............................................22 8.2 Example 2..............................................22 8.3 Example 3..............................................23 9. Security Considerations......................................23 10. Acknowledgements.............................................23 11. Informative references.......................................23 Lind, et al. Expires - May 2004 [Page 2] Internet-Draft Introducing IPv6 in ISP Networks December 2003 1. Introduction 1.1 Goal and scope of the document When an ISP deploys IPv6, its goal is to provide IPv6 connectivity to its customers. The new IPv6 service must be added to an already existing IPv4 service and the introduction of the IPv6 must not interrupt this IPv4 service. The case of an IPv6-only service provider is not addressed in this document. An ISP offering an IPv4 service will find that there are different ways to add IPv6 to this service. This document discusses a small set of scenarios for the introduction of IPv6 in an ISP IPv4 network. It evaluates the suitability of the existing transition mechanisms in the context of these deployment scenarios, and it points out the lack of functionality essential to the ISP operation of an IPv6 service.. The present document is focused on services that include both IPv6 and IPv4 and does not cover issues surrounding an IPv6-only service. It is also outside the scope of this document to describe different types of access or network technologies. 1.2 Terminology used This section defines and clarifies the terminology used in this document: "CPE" : Customer Premise Equipment "PE" : Provider Edge equipment "Network and service operation": : This is the part of the ISP network which hosts the services required for the correct operation of the ISP network. These services usually include management, supervision, accounting, billing and customer management applications. "Customer connection": : This is the part of the network which is used by a customer when connecting to an ISP network. It includes the CPEs, the last hop links and the parts Lind, et al. Expires - May 2004 [Page 3] Internet-Draft Introducing IPv6 in ISP Networks December 2003 of the PE interfacing to the last hop links. "Backbone" : This is the rest of the ISP network infrastructure. It includes the parts of the PE interfacing to the core, the core routers of the ISP and the border routers used in order to exchange routing information with other ISPs (or other administrative entities). "Dual-stack network": A network which supports natively both IPv4 and IPv6. 2. Brief description of a generic ISP network A generic ISP network topology can be divided into two main parts; the backbone network and the customer connection networks connecting the customers. The backbone is the part of the network that interconnects the different customer connection networks and provides transport to the rest of the Internet via exchange points or other means. The backbone network can be built on different technologies but in this document the only interest is whether it is capable of carrying IPv6 traffic natively or not. Since there is no clear definition of "backbone", it is defined in this document as being all routers that are a part of the same routed domain in the transport network. This means that all routers up to (and including, at least partially) the PE router are a part of the backbone. The customer connection networks provide connectivity to enterprise and private customers. Other ISPs might as well be customers and connected to the ISP's customer connection network. As with the backbone the absence or presence of native IPv6 capability is the only thing of real interest in the customer connection network technology. It is noticeable that, in some cases (e.g. incumbent national or regional operators), a given customer connection network may have to be shared between different ISPs. According to the type of the customer connection network used (e.g. involving only layer 2 devices, or involving non-IP technology), this constraint may result in architectural considerations that may be relevant in this document. Lind, et al. Expires - May 2004 [Page 4] Internet-Draft Introducing IPv6 in ISP Networks December 2003 "Network and service operation" building blocks refer to the basic main functions needed for a regular backbone operation. This building block is dealing with: network management, customers' authentication and accounting, address assignment and naming. It represents the minimum functions needed to provide a customer service, referring to both network infrastructure operation, and administrative management of customers. It doesn't matter if these customer networks have a single node or a large routed network. What is of interest is if routing information is exchanged or not since it will affect the ISP's network. The existence of customer premise equipment will also affect how a service can be delivered. In addition to the ISP's actual network components there are a lot of support and backend systems that have to be considered. The basic components in an ISP network are depicted in Figure 1. ------------ ---------- | Network and| | | | service |--| Backbone | | operation | | |\ ------------ ---------- \ . / | \ \ . / | \ \_Peering( Direct & IX ) . / | \ . / | \ . / | \ ---------- / ---------- \ ----------- | Customer | / | Customer | \ | Customer | |Connection|--/ |Connection| \--|Connection| | 1 | | 2 | | 3 | ---------- ---------- ---------- | | | ISP Network ------------------------------------------------------- | | | Customer Networks +--------+ +--------+ +--------+ | | | | | | |Customer| |Customer| |Customer| | | | | | | +--------+ +--------+ +--------+ Figure 1: ISP network topology Lind, et al. Expires - May 2004 [Page 5] Internet-Draft Introducing IPv6 in ISP Networks December 2003 3. Transition scenarios 3.1 Identification of scenarios This section describes different stages an ISP might consider when introducing IPv6 connectivity in the existing IPv4 network and the different scenarios that might occur in the respective stages. The stages here are snapshots of an ISP's network with respect to IPv6 maturity. Since an ISP's network is constantly evolving, a stage is a measure of how far an ISP has come in turn of implementing necessary functionality to offer IPv6 to the customers. It is possible to freely transition between different stages. However, a network segment can only be in one stage at a time but an ISP network as a whole can be in different stages. There are different transition paths between the first and final stage. The transition between two stages does not have to be immediate but can occur gradually. Each stage has different IPv6 properties. An ISP can therefore, based on the requirements it has, decide into which stage it will transform its network. This document is not aimed to cover very small or small ISPs or hosting providers/data centers; only the scenarios applicable to the ISPs eligible for a /32 IPv6 prefix allocation from a RIR are covered. 3.1.1 Assumptions The stages are derived from the generic description of an ISP network in section 2. A combination of different building blocks that constitute an ISP environment will lead to a number of scenarios, which an ISP can choose from. The scenarios of most relevance to this document are considered to be the ones that in the most efficient and feasible way maximize the ability for an ISP to offer IPv6 to its customers. The most predominant case today is considered to be an operator with a core and access IPv4 network delivering IPv4 service to a customer that is running IPv4 as well. At some point in the future, it is expected that the customer will want to have an IPv6 service, in Lind, et al. Expires - May 2004 [Page 6] Internet-Draft Introducing IPv6 in ISP Networks December 2003 addition to the already existing IPv4 service. This IPv6 service may be offered by the same ISP, or by a different one. Anyway the challenge for the ISP is to deliver the requested IPv6 service over the existing infrastructure and at the same time keep the IPv4 service intact. 3.1.2 Customer requirements and ISP offerings Looking at the scenarios from the end customer's perspective there might be a demand for three different services; the customer might demand IPv4 service, IPv6 service or both services. This can lead to two scenarios depending on the stage the ISP's network is in. If an ISP only offers IPv4 or IPv6 service and a customer wants to connect a device or network that only supports one service and if that service is not offered, it will lead to a dead-end. If the customer or ISP instead connects a dual stack device it is possible to circumvent the lack of the missing service in the customer connection network by using some kind of tunneling mechanism. This scenario will only be considered in the perspective of the ISP offering a mechanism to bridge the customer connection and reach the IPv6 backbone. In the case where the customer connects a single stack device to a corresponding single stack customer connection network or when the customer connects a single stack device to a dual stack customer connection network is covered by the all over dual stack case. Therefore, neither of these cases need further be explored separately in this document but can be seen as a part of a full dual stack case. After eliminating a number of cases explained above, there are four stages that are most probable and where an ISP will find its network in. Which stage a network is in depends on whether or not some part of the network previously has been upgraded to support IPv6 or if it is easier to enable IPv6 in one part or another. For instance, routers in the backbone might have IPv6 support or might be easily upgradeable, while the hardware in the customer connection network might have to be completely replaced in order to handle IPv6 traffic. Note that in every stage except Stage 1, the ISP can offer both IPv4 and IPv6 services to the customer. Lind, et al. Expires - May 2004 [Page 7] Internet-Draft Introducing IPv6 in ISP Networks December 2003 The four most probable stages are: o Stage 1 Launch o Stage 2a Backbone o Stage 2b Customer connection o Stage 3 Complete Generally the ISP is able to upgrade current IPv4 network to IPv4/IPv6 dual-stack network via Stage 2b but the IPv6 service can also be implemented at a small cost with simple tunnel mechanisms on the existing system. When designing a new network, Stage 3 might be the first and last step since there are no legacy concerns. Absence of IPv6 capability in the network equipment can still be a limiting factor nevertheless. 3.1.3 Stage 1 Scenarios: Launch The first stage is an IPv4 only ISP with an IPv4 customer. This is the most common case today and has to be the starting point for the introduction of IPv6. From this stage, an ISP can move (transition) to any other stage with the goal to offer IPv6 to its customer. The immediate first step consists of getting a prefix allocation (typically a /32) from the appropriate RIR according to allocation procedures. 3.1.4 Stage 2a Scenarios: Backbone Stage 2a is an ISP with customer connection networks that are IPv4 only and a backbone that supports both IPv4 and IPv6. In particular, the ISP considers it possible to make the backbone IPv6 capable either through software or hardware upgrade, or a combination of both. In this stage the customer should have support for both IPv4 and IPv6. The ISP has to provide IPv6 connectivity through its IPv4 customer connection networks. In particular, the existence of NATs and firewalls in the path (at the CPE, or in the customer's network) need to be considered. 3.1.5 Stage 2b Scenarios: Customer connection Stage 2b is an ISP with a backbone network that is IPv4 and an customer connection network that supports both IPv4 and IPv6. Since the service to the customer is native IPv6 there is no requirement Lind, et al. Expires - May 2004 [Page 8] Internet-Draft Introducing IPv6 in ISP Networks December 2003 for the customer to support both IPv4 and IPv6. This is the biggest difference in comparison to the previous stage. The need to exchange IPv6 traffic or similar still exists but might be more complicated than in the previous case since the backbone isn't IPv6 enabled. After completing stage 2b the original IPv4 backbone still is unchanged. This doesn't imply that there is no IPv6 backbone just that the IPv6 backbone is an overlay to or partially separated from the IPv4 backbone. Generally, the ISP will continue providing IPv4 connectivity; in many cases private addresses and NATs will continue to be used. 3.1.6 Stage 3 scenarios: Complete Stage 3 can be said to be the final step in introducing IPv6, at least in the scope of this document. This is an all over IPv6 service with native support for IPv6 and IPv4 in both backbone and customer connection networks. This stage is identical to the previous stage in the customer's perspective since the customer connection network hasn't changed. The requirement for exchanging IPv6 traffic is identical to stage 2. 3.1.7 Stage 2a and 3 combination scenarios Some ISPs may use different access technologies of varying IPv6 maturity. This may result in a combination of the stages 2a and 3: some customer connections do not support IPv6, but some do; and the backbone is dual-stack. This is equivalent to stage 2a, but requiring support for native IPv6 customer connections on some access technologies. 3.2 Transition Scenarios Given the different stages it is clear that the ISP has to be able to transition from one stage to another. The initial stage, in this document, is the IPv4 only service and network. The end stage is the dual IPv4/IPv6 service and network. As mentioned in the scope, this document does not cover the IPv6 only service and network and its implications on IPv4 customers. This has nothing to do with the usability of an IPv6 only service. Lind, et al. Expires - May 2004 [Page 9] Internet-Draft Introducing IPv6 in ISP Networks December 2003 The transition starts out with the IPv4 ISP and then moves to one of three choices. These choices are the different transition scenarios. One way would be to upgrade the backbone first which leads to stage 2a. Another way to go could be to upgrade the customer connection network which leads to stage 2b. The final possibility is to introduce IPv6 in both the backbone and customer connections as needed which would lead to stage 3. The choice is dependent on many different issues. For example it might not be possible to upgrade the backbone or the customer connection network without large investments in new equipment which could lead to any of the two first choices. In another case it might be easier to take the direct step to a complete IPv6/IPv4 network due to routing protocol issues or similar. If a partially upgraded network (stage 2a or 2b) was chosen as the first step, a second step remains before the network is all over native IPv6/IPv4. This is the transition to an all over dual stack network. This step is perhaps not necessary for stage 2b with an already native IPv6 service to the customer but might still occur when the timing is right. For stage 2a it is more obvious that a transition to a dual stack network is necessary since it has to be done to offer a native IPv6 service. As most of the ISPs keep evolving continuously their backbone IPv4 networks (new firmware versions in the routers, new routers), they will be able to get them IPv6 ready, without additional investment, except the staff training. It may be a slower transition path, but useful since it allows an IPv6 introduction without any actual customer demand. This will probably be better than making everything at the last minute with a higher investment. 3.3 Actions needed when deploying IPv6 in an ISP network When looking at the transitions described above, it appears that it is possible to split the work required by each transition in a small set of actions. Each action is mostly independent from the others and some actions may be common to several transitions. The analysis of the possible transitions leads to a small list of actions: * backbone transition actions: Lind, et al. Expires - May 2004 [Page 10] Internet-Draft Introducing IPv6 in ISP Networks December 2003 - Connect dual-stack customer connection networks to other IPv6 networks through an IPv4 backbone, - Transform an IPv4 backbone into a dual-stack one. This action can be performed directly or through intermediate steps, * customer connection transition actions: - Connect IPv6 customers to an IPv6 backbone through an IPv4 network, - Transform an IPv4 customer connection network into a dual- stack one, * network and service operation transition actions: - configure IPv6 functions into either backbone or network and service operation devices - upgrade regular network management and monitoring applications to take IPv6 into account - [Network and service operation actions - To be completed.] More detailed descriptions of each action follow. 4. Backbone transition actions 4.1 Steps in transitioning backbone networks In terms of physical equipment, backbone networks consist mainly in core and edge high-speed routers. Border routers provide peering with other providers. Filtering, routing policy and policing type functions are generally managed on border routers. The initial step is an IPv4-only backbone, and the final step is a whole dual-stack backbone. In between, intermediate steps may be identified: Tunnels Tunnels IPv4-only ----> or ---> or + DS -----> Full DS IPv6 dedicated IPv6 dedicated routers links links Lind, et al. Expires - May 2004 [Page 11] Internet-Draft Introducing IPv6 in ISP Networks December 2003 The first step involves tunnels or dedicated links but existing routers are left unchanged. Only a small set of routers then have IPv6 capabilities. Configured tunnels are adequate for use during this step. When MPLS is already deployed in the backbone, it may be desirable to provide IPv6-over-MPLS connectivity. However, the problem is that setting up an IPv6 Label Switched Path (LSP) requires some signaling through the MPLS network; both LDP and RSVP-TE can set up IPv6 LSPs, but this would require a software upgrade in the MPLS core network. A workaround is to use BGP for signaling and/or to perform IPv6- over-IPv4/MPLS or IPv6-over-IPv4-over-IPv4/MPLS encapsulation, for example, as described in [BGPTUNNEL]. There seem to be multiple possibilities, some of which may be more preferable than others. More analysis is needed in order to determine which are the best approach(es): 1) require that MPLS networks deploy native IPv6 support or use configured tunneling for IPv6. 2) require that MPLS networks support setting up IPv6 LSPs, and IPv6 connectivity is set up using them, or configured tunneling is used. 3) use only configured tunneling over the IPv4 LSPs; this seems practical with small-scale deployments when the number of tunnels is low. 4) use something like [BGPTUNNEL] to perform IPv6-over- IPv4/MPLS encapsulation for IPv6 connectivity. In the second step, some dual stack routers are added in this network in a progressive manner. The final stage is reached when all or most routers are dual-stack. According to many reasons (technical, financial, etc), an ISP may move forward from step to step or reach directly the final one. One of the important criteria in this evolution is the number of IPv6 customers the ISP gets on its initial deployments. If few customers connect to the first IPv6 infrastructure, then the ISP is likely to remain on the initial steps for a long time. Lind, et al. Expires - May 2004 [Page 12] Internet-Draft Introducing IPv6 in ISP Networks December 2003 In short, each step remains possible, but no one is mandatory. 4.2 Configuration of backbone equipment In the backbone, the number of devices is small and IPv6 configuration mainly deals with routing protocols parameters, interface addresses, loop-back addresses, ACLs... These IPv6 parameters are not supposed to be automatically configured. 4.3 Routing ISPs need routing protocols to advertise the reachability and to find the shortest working paths both internally and externally. OSPFv2 and IS-IS are typically used as an IPv4 IGP. RIPv2 is typically not in use in operator networks. BGP is the only IPv4 EGP. Static routes are used in both. Note that it is possible to configure a given network so that it results in having an IPv6 topology different from the IPv4 topology. For example, some links or interfaces may be dedicated to IPv4-only or IPv6-only traffic, or some routers may be dual-stack while some others maybe single stacked (IPv4 or IPv6). In this case, the routing must be able to manage multiple topologies. 4.3.1 IGP Once the IPv6 topology has been determined the choice of IPv6 IGP must be made: either OSPFv3 or IS-IS for IPv6. RIPng is less appropriate in many contexts and is not discussed here. The IGP typically includes the routers' point-to-point and loop-back addresses. The most important decision to make is whether one wishes to have separate routing protocol processes for IPv4 and IPv6. Having them separate requires more memory and CPU for route calculations e.g. when the links flap. On the other hand, the separation provides a better reassurance that if problems come up with IPv6 routing, they will not affect IPv4 routing protocol at all. In the first phases if it is uncertain whether joint IPv4/IPv6 networks work as intended, having separate processes may be desirable and easier to manage. Lind, et al. Expires - May 2004 [Page 13] Internet-Draft Introducing IPv6 in ISP Networks December 2003 Thus the combinations are: - Separate processes: o OSPFv2 for IPv4, IS-IS for IPv6 (-only) o OSPFv2 for IPv4, OSPFv3 for IPv6, or o IS-IS for IPv4, OSPFv3 for IPv6 - The same process: o IS-IS for both IPv4 and IPv6 Note that if IS-IS is used for both IPv4 and IPv6, the IPv4/IPv6 topologies must be "convex", unless the Multiple-topology IS-IS extensions [MTISIS] have been implemented. In simpler networks or with careful planning of IS-IS link costs, it is possible to keep even non-congruent IPv4/IPv6 topologies "convex". Therefore, the use of same process is recommended especially for large ISPs which intend to deploy, in the short-term, a fully dual- stack backbone infrastructure. If the topologies are not similar in the short term, two processes (or Multi-topology IS-IS extensions) are recommended. The IGP is not typically used to carry customer prefixes or external routes. Internal BGP (iBGP), as described in the next section, is most often deployed in all routers to spread the other routing information. As some of the simplest devices, e.g. CPE routers, may not implement other routing protocols than RIPng, in some cases it may be necessary to also run RIPng in a limited fashion in addition to another IGP, and somehow redistribute the routing information to the other routing protocol(s). 4.3.2 EGP BGP is used for both internal BGP and external BGP sessions. BGP can be used for IPv6 with Multi-protocol extensions [RFC 2858], [RFC 2545]. These enable exchanging both IPv6 routing information as establishing BGP sessions using TCP over IPv6. Lind, et al. Expires - May 2004 [Page 14] Internet-Draft Introducing IPv6 in ISP Networks December 2003 It is possible to use a single BGP session to advertise both IPv4 and IPv6 prefixes between two peers. However, typically, separate BGP sessions are used. 4.3.3 Routing protocols transport IPv4 routing information should be carried by IPv4 transport and IPv6 one by IPv6 for several reasons: * The IPv6 connectivity may work when the IPv4 one is down (or vice-versa). * The best route for IPv4 is not always the best one for IPv6. * The IPv4 logical topology and the IPv6 one may be different because, the administrator may want to use different metric values for one physical link for load balancing or tunnels may be used. 4.4 Multicast Currently, IPv6 multicast is not a strong concern for most ISPs. However, some of them consider deploying it. Multicast is achieved using PIM-SM and PIM-SSM protocols. These also work with IPv6. Information about multicast sources is exchanged using MSDP in IPv4, but it is not defined, on purpose, for IPv6. An alternative mechanism is to use only PIM-SSM or an alternative mechanism for conveying the information [EMBEDRP]. To be completed. send feedback/text! 5. Customer connection transition actions 5.1 Steps in transitioning customer connection networks customer connection networks are generally composed of a large number of CPEs connected to a small set of PEs. Transitioning this infrastructure to IPv6 can be made in several steps, but some ISPs may avoid some of the steps depending on their perception of risks. Connecting IPv6 customers to an IPv6 backbone through an IPv4 network can be considered as a first careful step taken by an ISP in order to provide IPv6 services to its IPv4 customers. More, some ISPs also provide IPv6 services to customers who get their IPv4 services from another ISP. Lind, et al. Expires - May 2004 [Page 15] Internet-Draft Introducing IPv6 in ISP Networks December 2003 This IPv6 service can be provided by using tunneling techniques. The tunnel may terminate at the CPE corresponding to the IPv4 service or in some other part of the customer's infrastructure (for instance, on an IPv6 specific CPE or even on a host). Several tunneling techniques have already been defined: configured tunnels with tunnel broker, 6to4, Teredo... The selection of one candidate depends largely on the presence or not of NATs between the two tunnel end-points, and whether the user's IPv4 tunnel end-point address is static or dynamic. In most cases, 6to4 and ISATAP are incompatible with NATs and an UDP encapsulation for configured tunnels has not been specified. Firewalls in the way can also break these types of tunnels. The administrator of the firewall will have to create a hole for the tunnel. It is not a big deal as long as the firewall is controlled either by the customer or the ISP, which is almost always the case. An ISP has practically two kinds of customers in the customer connection networks: small end users (mostly "unmanaged networks"; home and SOHO users), and others. The former category typically has a dynamic IPv4 address which is often NATted; a reasonably static address is also possible. The latter category typically has static IPv4 addresses, and at least some of them are public; however, NAT traversal or configuring the NAT may be required to reach an internal IPv6 access router, though. Tunneling consideration for small and end sites are discussed in [UNMANCON], that may identify solutions relevant to the first category of unmanaged network. These solutions will be further discussed within an ISP context, when available. For the second category, usually: * Configured tunnels as-is are a good solution when an NAT is not in the way and the IPv4 end-point addresses are static. A mechanism to punch through NATs or to forward packets through it may be desirable in some scenarios. If fine-grained access control is needed, an authentication protocol needs to be used. Lind, et al. Expires - May 2004 [Page 16] Internet-Draft Introducing IPv6 in ISP Networks December 2003 * A tunnel brokering solution, to help facilitate the set-up of a bi-directional tunnel, has been proposed: the Tunnel Set-up Protocol. Whether this is the right way needs to be determined. * Automatic tunneling mechanisms such as 6to4 or Teredo are not applicable in this scenario. Some other ISPs may take a more direct approach and avoid the use of tunnels as much as possible. Note that when the customers use dynamic IPv4 addresses, the tunneling techniques may be more difficult at the ISP's end, especially if a protocol not including authentication (like PPP for IPv6) is not used. This may need to be considered in more detail with tunneling mechanisms. 5.2 Access control requirements Access control is usually required in ISP networks because the ISPs need to control to who they are giving access. For instance, it is important to check if the user who tries to connect is really a valid customer. In some cases, it is also important for billing purposes. However, an IPv6 specific access control is not always required. This is for instance the case when a customer of the IPv4 service has automatically access to the IPv6 service. Then, the IPv4 access control also gives access to the IPv6 services. When the provider does not wish to give to its IPv4 customers automatically access to IPv6 services, a specific access control for IPv6 must be performed in parallel to the IPv4 one. It does not mean that a different user authentication must be performed for IPv6, but the authentication process may lead to different results for IPv4 and IPv6 access. Access control traffic may use IPv4 or IPv6 transport. For instance, Radius traffic related to an IPv6 service can be transported over IPv4. 5.3 Configuration of customer equipment The customer connection networks are composed of CPEs and PEs. Usually, each PE connects a large number of CPEs to the backbone Lind, et al. Expires - May 2004 [Page 17] Internet-Draft Introducing IPv6 in ISP Networks December 2003 network infrastructure. This number may reach tens of thousands or more. The configuration of CPEs is an heavy task for the ISP and this is even made harder as the configuration must be done remotely. In this context, the use of auto-configuration mechanisms is very beneficial, even if manual configuration is still an option. The parameters that usually need to be automatically provided to the customers are: - The network prefix delegated by the ISP, - The address of the Domain Name System server (DNS), - Some other parameters such as the address of an NTP server may also be needed sometimes. When access control is required on the ISP network, DHCPv6 can provide the configuration parameters. This is discussed more in details in [DUAL ACCESS]. When access control is not required (unusual case), a stateless mechanism could be used, but no standard definition exists at the moment. 5.4 Requirements for Traceability Most ISPs have some kind of mechanism to trace the origin of traffic in their networks. This has also to be available for IPv6 traffic. This means that specific IPv6 address or prefix has to be tied to a certain customer, or that records of which customer had which address/prefix must be maintained. This also applies to the customers with tunneled connectivity. This can be done for example by mapping a DHCP response to a physical connection and storing this in a database. It can also be done by assigning a static address or refix to the customer. For any traceability to be useful, ingress filtering must be deployed towards all the customers. 5.5 Multi-homing Customers may desire multihoming or multi-connecting for a number of reasons [RFC3582]. Lind, et al. Expires - May 2004 [Page 18] Internet-Draft Introducing IPv6 in ISP Networks December 2003 Multihoming to more than one ISP is a subject still under debate. Deploying multiple addresses from each ISP and using the addresses of the ISP when sending traffic to that ISP is at least one working model; in addition, tunnels may be used for robustness [RFC3178]. Currently, there are no provider-independent addresses for end- sites. Multi-connecting more than once to just one ISP is a simple practice, and this can be done e.g. with BGP with public or private AS numbers and a prefix assigned to the customer. To be further defined as the multihoming situation gets clearer. 5.6 Ingress filtering in the customer connection network Ingress filtering must be deployed everywhere towards the customers, to ensure traceability, prevent DoS attacks using spoofed addresses, prevent illegitimate access to the management infrastructure, etc. The ingress filtering can be done for example using access lists or Unicast Reverse Path Forwarding (uRPF). Mechanisms for these are described in [BCP38UPD]. 6. Network and service operation actions The network and service operation actions fall into different categories listed below: - IPv6 network devices configuration: for initial configuration and updates - IPv6 Network Management - IPv6 Monitoring - IPv6 customer management - built-in "network and service operation" IPv6 security Some of these actions will require an IPv6 native transport layer to be available, while some other will not. In a first step, network devices configuration and regular network management operations can be performed over an IPv4 transport, as IPv6 MIBs are also available over IPv4 transport. Lind, et al. Expires - May 2004 [Page 19] Internet-Draft Introducing IPv6 in ISP Networks December 2003 Nevertheless, some monitoring functions require IPv6 transport availability. This is for instance the case when ICMP messages are used by the monitoring applications. In a second step, IPv6 transport can be provided for any of these network and service operation facilities. [To be completed, send feedback/text!] 7. Future Stages After a while the ISP might want to transition to a service that is IPv6 only, at least in certain parts of the network. This transition creates a lot of new cases in which to factor in how to maintain the IPv4 service. Providing an IPv6 only service is not much different than the dual IPv4/IPv6 service described in stage 3 except from the need to phase out the IPv4 service. The delivery of IPv4 services over an IPv6 network and the phase out is left for a future document. 8. Example networks In this section, a number of different network examples are presented. They are only example networks and will not necessary match to any existing networks. Nevertheless, the examples will hopefully be useful even in the cases when they do not match the target networks. The purpose of the example networks is to exemplify the applicability of the transition mechanisms described in this document on a number of different example networks with different prerequisites. The example network layout will be the same in all network examples. The networks examples are to be seen as a specific representation of the generic network with a limited number of network devices. An arbitrary number (in this case 7) of routers have been selected to represent the network examples. However, since the network examples follow the implementation strategies recommended for the generic network scenario, it should be possible to scale the example to fit a network with an arbitrary number, e.g. several hundreds or thousands, of routers. The routers in the example are interconnected with each other as well as with another ISP. The connection to another ISP can either be a direct connection or through an exchange point. In addition to Lind, et al. Expires - May 2004 [Page 20] Internet-Draft Introducing IPv6 in ISP Networks December 2003 these connections, there are also a number of customer connection networks connected to the routers. customer connection networks are normally connected to the backbone via access routers, but can in some cases be directly connected to the backbone routers. As described earlier in the generic network scenarios, the customer connection networks are used to connect the customers. customer connection networks can, for example, be xDSL or cable network equipment. | ISP1 | ISP2 +------+ | +------+ | | | | | |Router|--|--|Router| | | | | | +------+ | +------+ / \ +----------------------- / \ / \ +------+ +------+ | | | | |Router|----|Router| | | | | +------+ +------+\ | | \ | Exchange point +------+ +------+ \ +------+ | +------+ | | | | \_| | | | |-- |Router|----|Router|----\|Router|--|--|Switch|-- | | | | | | | | |-- +------+ /+------+ +------+ | +------+ | / | | +-------+/ +-------+ | | | | | |Access1| |Access2| | | | | +-------+ +-------+ ||||| ||||| ISP Network ----|-----------|---------------------- | | Customer Networks +--------+ +--------+ | | | | |Customer| |Customer| | | | | +--------+ +--------+ Figure 2: ISP network example Lind, et al. Expires - May 2004 [Page 21] Internet-Draft Introducing IPv6 in ISP Networks December 2003 8.1 Example 1 In example 1 a network built according to the example topology is present with a native IPv4 backbone, the routers. The backbone is running IS-IS and IBGP as routing protocol for internal and external routes respectively. In the connection to ISP2 and the exchange point MBGP is used to exchange routes. Multicast is present and is using PIM-SM routing. QoS is present and is using DiffServ. Access 1 is xDSL, connected to the backbone through an access router. The xDSL equipment, except for the access router, is considered to be layer 2 only, e.g. Ethernet or ATM. IPv4 addresses are dynamically assigned to the customer using DHCP. No routing information is exchanged with the customer. Access control and traceability is done in the access router. Customers are separated in VLANs or separate ATM PVCs up to the access router. Access 2 is Fiber to the building/home connected directly to the backbone router. The FTTB/H is considered to be layer 3 aware and performs access control and traceability through its layer 3 awareness. IPv4 addresses are dynamically assigned to the customers using DHCP. No routing information is exchanged with the customer. 8.2 Example 2 In example 2 the backbone is running IPv4 with MPLS. Routing protocols used are OSPF and IBGP for internal and external routes. In the connection to ISP2 and the exchange point BGP is used to exchange routes. Multicast and QoS are not present. Access 1 is a fixed line access, e.g. fiber, connected directly to the backbone. CPE is present at the customer and routing information is in some cases exchanged otherwise static routing is used. Access 1 can also be connected to BGP/MPLS-VPN running in the backbone. Access 2 is xDSL connected directly to the backbone router. The xDSL is layer 3 aware. Addresses are dynamically assigned using DHCP. Access control is achieved on the physical layer and traceability is achieved using DHCP snooping. No routing information is exchanged with the customer. Lind, et al. Expires - May 2004 [Page 22] Internet-Draft Introducing IPv6 in ISP Networks December 2003 8.3 Example 3 A transit provider offers IP connectivity to other providers, but not to end users or enterprises. IS-IS and IBGP is used internally and BGP externally. Its accesses connect Tier-2 provider cores. No multicast or QoS is used. 9. Security Considerations This document analyses scenarios and identifies some transition mechanisms that could be used for the scenarios. It does not introduce any new security issues. Security considerations of each mechanism are described in the respective documents. 10. Acknowledgements This draft has greatly benefited from inputs by Pekka Savola, Marc Blanchet, Jordi Palet. 11. Informative references [EMBEDRP] Savola, P., Haberman, B., "Embedding the Address of RP in IPv6 Multicast Address" - draft-ietf-mboned-embeddedrp-00.txt [MTISIS] Przygienda, T., Naiming Shen, Nischal Sheth, "M- ISIS: Multi Topology (MT) Routing in IS-IS" draft-ietf-isis-wg-multi-topology-06.txt [RFC 2858] T. Bates, Y. Rekhter, R. Chandra, D. Katz, "Multiprotocol Extensions for BGP-4" RFC 2858 [RFC 2545] P. Marques, F. Dupont, "Use of BGP-4 Multiprotocol Extensions for IPv6 Inter-Domain Routing" RFC 2545 [BCP38UPD] F. Baker, P. Savola "Ingress Filtering for Multihomed Networks" Lind, et al. Expires - May 2004 [Page 23] Internet-Draft Introducing IPv6 in ISP Networks December 2003 draft-savola-bcp38-multihoming-update-01.txt [RFC3582] J. Abley, B. Black, V. Gill, "Goals for IPv6 Site- Multihoming Architectures" RFC 3582 [RFC3178] J. Hagino, H. Snyder, "IPv6 Multihoming Support at Site Exit Routers" RFC 3178 [BGPTUNNEL] J. De Clercq, G. Gastaud, D. Ooms, S. Prevost, F. Le Faucheur "Connecting IPv6 Islands across IPv4 Clouds with BGP" draft-ooms-v6ops-bgp-tunnel-00.txt [DUAL ACCESS] Y. Shirasaki, S. Miyakawa, T. Yamasaki, A. Takenouchi "A Model of IPv6/IPv4 Dual Stack Internet Access Service" draft-shirasaki-dualstack-service-02.txt [UNMANCON] T.Chown, R. van der Pol, P. Savola, "Considerations for IPv6 Tunneling Solutions in Small End Sites" draft-chown-v6ops-unmanaged-connectivity-00 Authors' Addresses Mikael Lind TeliaSonera Vitsandsgatan 9B SE-12386 Farsta, Sweden Email: mikael.lind@teliasonera.com Vladimir Ksinant 6WIND S.A. Immeuble Central Gare - Bat.C 1, place Charles de Gaulle 78180, Montigny-Le-Bretonneux - France Phone: +33 1 39 30 92 36 Email: vladimir.ksinant@6wind.com Soohong Daniel Park Mobile Platform Laboratory, SAMSUNG Electronics. Lind, et al. Expires - May 2004 [Page 24] Internet-Draft Introducing IPv6 in ISP Networks December 2003 416, Maetan-3dong, Paldal-Gu, Suwon, Gyeonggi-do, Korea Email: soohong.park@samsung.com Alain Baudot France Telecom R&D 42, rue des coutures 14066 Caen - FRANCE Email: alain.baudot@rd.francetelecom.com Intellectual Property Statement The IETF takes no position regarding the validity or scope of any intellectual property or other rights that might be claimed to pertain to the implementation or use of the technology described in this document or the extent to which any license under such rights might or might not be available; neither does it represent that it has made any effort to identify any such rights. Information on the IETF's procedures with respect to rights in standards-track and standards-related documentation can be found in BCP-11. Copies of claims of rights made available for publication and any assurances of licenses to be made available, or the result of an attempt made to obtain a general license or permission for the use of such proprietary rights by implementers or users of this specification can be obtained from the IETF Secretariat. The IETF invites any interested party to bring to its attention any copyrights, patents or patent applications, or other proprietary rights which may cover technology that may be required to practice this standard. Please address the information to the IETF Executive Director. Full Copyright Statement Copyright (C) The Internet Society (2003). All Rights Reserved. This document and translations of it may be copied and furnished to others, and derivative works that comment on or otherwise explain it or assist in its implementation may be prepared, copied, published and distributed, in whole or in part, without restriction of any kind, provided that the above copyright notice and this paragraph are included on all such copies and derivative works. However, this Lind, et al. Expires - May 2004 [Page 25] Internet-Draft Introducing IPv6 in ISP Networks December 2003 document itself may not be modified in any way, such as by removing the copyright notice or references to the Internet Society or other Internet organizations, except as needed for the purpose of developing Internet standards in which case the procedures for copyrights defined in the Internet Standards process must be followed, or as required to translate it into languages other than English. The limited permissions granted above are perpetual and will not be revoked by the Internet Society or its successors or assignees. This document and the information contained herein is provided on an "AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Lind, et al. Expires - May 2004 [Page 26]