Internet Research Task Force F. Templin, Ed. (IRTF) Boeing Research & Technology Internet-Draft June 23, 2010 Intended status: Informational Expires: December 25, 2010 The Internet Routing Overlay Network (IRON) draft-templin-iron-06.txt Abstract Since the Internet must continue to support escalating growth due to increasing demand, it is clear that current routing architectures and operational practices must be updated. This document proposes an Internet Routing Overlay Network for supporting sustainable growth through Provider Independent addressing while requiring no changes to end systems and no changes to the existing routing system. This document is a product of the IRTF Routing Research Group (RRG). Status of this Memo This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet- Drafts is at http://datatracker.ietf.org/drafts/current/. 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." This Internet-Draft will expire on December 25, 2010. Copyright Notice Copyright (c) 2010 IETF Trust and the persons identified as the document authors. All rights reserved. This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (http://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must Templin Expires December 25, 2010 [Page 1] Internet-Draft IRON June 2010 include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License. Table of Contents 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 4 3. The Internet Routing Overlay Network (IRON) . . . . . . . . . 5 3.1. IR(EP)s - IRON Routers That Connect End User Networks to the IRON . . . . . . . . . . . . . . . . . . . . . . . 6 3.2. IR(VP)s - IRON Routers That Serve Virtual Prefixes . . . . 7 3.3. IR(GW)s - IRON Routers that Connect VPC Networks to the Internet . . . . . . . . . . . . . . . . . . . . . . . 8 4. IRON Organizational Principles . . . . . . . . . . . . . . . . 9 5. IRON Initialization . . . . . . . . . . . . . . . . . . . . . 10 5.1. IR(VP) and IR(GW) Initialization . . . . . . . . . . . . . 10 5.2. IR(EP) Initialization . . . . . . . . . . . . . . . . . . 11 6. IRON Operation . . . . . . . . . . . . . . . . . . . . . . . . 12 6.1. IR(EP) Operation . . . . . . . . . . . . . . . . . . . . . 12 6.2. IR(VP) Operation . . . . . . . . . . . . . . . . . . . . . 13 6.3. IR(GW) Operation . . . . . . . . . . . . . . . . . . . . . 14 6.4. IRON Reference Operating Scenarios . . . . . . . . . . . . 14 6.4.1. Two Hosts in Different IRON EUNs . . . . . . . . . . . 14 6.4.2. Mixed IRON and Non-IRON Hosts . . . . . . . . . . . . 16 6.5. Mobility, Multihoming and Traffic Engineering Considerations . . . . . . . . . . . . . . . . . . . . . . 18 6.5.1. Mobility Management . . . . . . . . . . . . . . . . . 18 6.5.2. Multihoming . . . . . . . . . . . . . . . . . . . . . 18 6.5.3. Inbound Traffic Engineering . . . . . . . . . . . . . 18 6.5.4. Outbound Traffic Engineering . . . . . . . . . . . . . 18 6.6. Renumbering Considerations . . . . . . . . . . . . . . . . 19 6.7. NAT Traversal Considerations . . . . . . . . . . . . . . . 19 7. Open Research Areas . . . . . . . . . . . . . . . . . . . . . 20 8. Related Initiatives . . . . . . . . . . . . . . . . . . . . . 20 9. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 21 10. Security Considerations . . . . . . . . . . . . . . . . . . . 21 11. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 21 12. References . . . . . . . . . . . . . . . . . . . . . . . . . . 21 12.1. Normative References . . . . . . . . . . . . . . . . . . . 21 12.2. Informative References . . . . . . . . . . . . . . . . . . 22 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . . 23 Templin Expires December 25, 2010 [Page 2] Internet-Draft IRON June 2010 1. Introduction Growth in the number of entries carried in the Internet routing system has led to concerns for unsustainable routing scaling [I-D.narten-radir-problem-statement]. Operational practices such as increased use of multihoming with IPv4 Provider-Independent (PI) addressing are resulting in more and more fine-grained prefixes injected into the routing system from more and more end user networks. Furthermore, the forthcoming depletion of the public IPv4 address space has raised concerns for both increased deaggregation (leading to yet further routing table entries) and an impending address space run-out scenario. At the same time, the IPv6 routing system is beginning to see growth in IPv6 Provider-Aggregated (PA) prefixes [BGPMON] which must be managed in order to avoid the same routing scaling issues the IPv4 Internet now faces. Since the Internet must continue to scale to accommodate increasing demand, it is clear that new routing methodologies and operational practices are needed. Several related works have investigated routing scaling issues and proposed solutions. Virtual Aggregation (VA) [I-D.ietf-grow-va] and Aggregation in Increasing Scopes (AIS) [I-D.zhang-evolution] are global routing proposals that introduce routing overlays with Virtual Prefixes (VPs) to reduce the number of entries required in each router's Forwarding Information Base (FIB) and Routing Information Base (RIB). Routing and Addressing in Networks with Global Enterprise Recursion (RANGER) [RFC5720] examines recursive arrangements of enterprise networks that can apply to a very broad set of use case scenarios [I-D.russert-rangers]. In particular, RANGER supports encapsulation and secure redirection by treating each layer in the recursive hierarchy as a virtual non-broadcast, multiple access (NBMA) "link". RANGER is an architectural framework that includes Virtual Enterprise Traversal (VET) [I-D.templin-intarea-vet] and the Subnetwork Adaptation and Encapsulation Layer (SEAL) (including the SEAL Control Message Protocol (SCMP)) [I-D.templin-intarea-seal] as its functional building blocks. This document proposes an Internet Routing Overlay Network (IRON) with goals of supporting sustainable growth while requiring no changes to the existing routing system. IRON borrows concepts from VA, AIS and RANGER, and further borrows concepts from the Internet Vastly Improved Plumbing (Ivip) [I-D.whittle-ivip-arch] architecture proposal. IRON specifically seeks to provide scalable PI addressing without changing the current BGP [RFC4271] routing system. IRON observes the Internet Protocol standards [RFC0791][RFC2460]. Other network layer protocols that can be encapsulated within IP packets (e.g., OSI/CLNP [RFC1070], etc.) are also within scope. Templin Expires December 25, 2010 [Page 3] Internet-Draft IRON June 2010 The IRON is a global overlay network routing system comprising Virtual Prefix Companies (VPCs) that own and manage Virtual Prefixes (VPs) from which End User Network (EUN) PI prefixes (EPs) are delegated to customer sites. The IRON is motivated by a growing customer demand for multihoming, mobility management and traffic engineering while using stable PI addressing to avoid network renumbering [RFC4192][RFC5887]. The IRON uses the existing IPv4 and IPv6 global Internet routing systems as virtual links for tunneling inner network protocol packets within outer IPv4 or IPv6 headers (see: Section 3). The IRON requires deployment of a small number of new routers that are simple commodity hardware platforms. No modifications to hosts, and no modifications to most routers are required. Note: This document is offered in compliance with Internet Research Task Force (IRTF) document stream procedures [RFC5743]; it is not an IETF product and is not a standard. The views in this document were considered controversial by the IRTF Routing Research Group (RRG) but the RG reached a consensus that the document should still be published. The document will undergo a period of review within the RRG and through selected expert reviewers prior to publication. The following sections discuss details of the IRON architecture. 2. Terminology This document makes use of the following terms: End User Network (EUN) an edge network that connects an organization's devices (e.g., computers, routers, printers, etc.) to the Internet and possibly also the IRON. Internet Service Provider (ISP) a service provider which physically connects customer EUNs to the Internet. Provider Assigned (PA) prefix a network layer address prefix delegated to an EUN by a service provider. Provider Independent (PI) prefix a network layer address prefix delegated to an EUN by a third party independently of the EUN's ISP arrangements. Templin Expires December 25, 2010 [Page 4] Internet-Draft IRON June 2010 Virtual Prefix (VP) a highly-aggregated PI prefix block (e.g., an IPv4 /16, an IPv6 /20, etc.) that is owned and managed by a Virtual Prefix Company (VPC). Virtual Prefix Company (VPC) a company that owns and manages one or several VPs from which it delegates End User Network PI Prefixes (EPs) to EUNs Master Virtual Prefix database (MVPd) a distributed database that maintains VP-to-locator mappings for all VPs in the IRON. End User Network PI prefix (EP) a more-specific PI prefix derived from a VP (e.g., an IPv4 /28, an IPv6 /56, etc.) and delegated to an EUN by a VPC. EP Address (EPA) a network layer address taken from an EP address range and assigned to the interface of an end system in an EUN. EPAs are routable only within EUNs. locator an IP address taken from a non-EP address range and assigned to the interface of a router or end system within a public or private network. Locators taken from public IP address spaces are routable within the global Internet while locators taken from private IP address spaces are routable only within the network where the private IP addressing plan is deployed. Internet Routing Overlay Network (IRON) an overlay network configured over the global Internet. The IRON supports routing through encapsulation of inner packets with EPA addresses within outer headers that use locator addresses. 3. The Internet Routing Overlay Network (IRON) The Internet Routing Overlay Network (IRON) consists of IRON Routers (IRs) that use Virtual Enterprise Traversal (VET) and the Subnetwork Encapsulation and Adaptation Layer (SEAL) to encapsulate inner network layer packets within outer IP headers (see: Figure 1) for transmission over the global Internet. Each such IR connects to the IRON via a VET tunnel virtual interface used for automatic tunneling. Each IR therefore sees all other IRs as virtual single-hop neighbors on the link from the standpoint of the inner network layer protocol, while they may be separated by many physical outer IP hops. Templin Expires December 25, 2010 [Page 5] Internet-Draft IRON June 2010 +-------------------------+ | Outer Packet Header | ~ with locator addresses ~ | (IPv4 or IPv6) | +-------------------------+ +-------------------------+ | Inner Packet Header | --> | Inner Packet Header | ~ with EP addresses ~ --> ~ with EP addresses ~ | (IPv4, IPv6, OSI, etc.) | --> | (IPv4, IPv6, OSI, etc.) | +-------------------------+ +-------------------------+ | | --> | | ~ Inner Packet Body ~ --> ~ Inner Packet Body ~ | | --> | | +-------------------------+ +-------------------------+ Inner packet before Outer packet after before encapsulation after encapsulation Figure 1: Encapsulation of Inner Packets Within Outer IP Headers The IRON is manifested through a business model in which Virtual Prefix Companies (VPCs) own and manage a set of IRs that are distributed throughout the Internet and serve highly-aggregated Virtual Prefixes (VPs). VPCs delegate sub-prefixes from their VPs which they lease to customers as End User Network PI prefixes (EPs). The customers in turn assign the EPs to their customer premises IRs which connect their End User Networks (EUNs) to the IRON. VPCs may have no affiliation with the ISP networks from which customers obtain their basic connectivity. Therefore, VPCs can open for business and begin serving their customers immediately without the need to coordinate their activities with ISPs or with other VPCs. The IRON requires no changes to end systems and no changes to most routers in the Internet. Instead, the IRON comprises IRs that are deployed either as new platforms or as modifications to existing platforms. IRs may be deployed incrementally without disturbing the existing Internet routing system, and act as waypoints (or "cairns") for navigating the IRON. The functional roles of IRs are described in the following sections. 3.1. IR(EP)s - IRON Routers That Connect End User Networks to the IRON An "IR(EP)" is a tunnel endpoint router (or host with embedded gateway function) that logically connects its EUNs to the IRON via tunnels. IR(EP)s obtain EPs from VPCs and use them to number subnets and interfaces within their EUNs. An IR(EP) can be deployed as a Customer Premises Equipment (CPE) router that also physically connects its EUNs to its ISPs, but it may also be a router or even a singleton end system within the EUN when the CPE is a legacy router. Templin Expires December 25, 2010 [Page 6] Internet-Draft IRON June 2010 (This model applies even if the CPE router is a Network Address Translator (NAT) - see Section 6.7). An IR(EP) connects its EUNs to the IRON via tunnels as shown in Figure 2: .-. ,-( _)-. +--------+ .-(_ (_ )-. | IR(EP) |--(_ ISP ) +---+----+ `-(______)-' | <= T \ .-. .-. u \ ,-( _)-. ,-( _)-. n .-(_ (- )-. .-(_ (_ )-. n (_ Internet ) (_ EUN ) e `-(______)- `-(______)-' l ___ | s => (:::)-. +----+---+ .-(::::::::) | Host | .-(::::::::::::)-. +--------+ (:::: The IRON ::::) `-(::::::::::::)-' `-(::::::)-' Figure 2: IR(EP) Connecting EUN to the IRON 3.2. IR(VP)s - IRON Routers That Serve Virtual Prefixes An "IR(VP)" is a tunnel endpoint router that is managed by a VPC and that services one or more VPs from which it delegates EPs to customers. In typical deployments, a VPC will deploy several IR(VP)s (e.g., 10-20) for each VP. These IR(VP)s maintain a distributed database of EP-to-customer mappings that allow correspondents to navigate the IRON. IR(VP)s connect to the IRON in a globally-distributed fashion as shown in Figure 3 so that mapping service clients can discover a server that is nearby. IR(VP)s typically forward only initial packets while redirecting traffic flows, and often will require only a single physical network interface. Hence, IR(VP)s may be deployed on a variety of hardware platforms ranging from traditional high-end routers to commodity general-purpose processors. Templin Expires December 25, 2010 [Page 7] Internet-Draft IRON June 2010 +--------+ +--------+ | IR(VP) | | IR(VP) | | Boston | | Tokyo | +--------+ +--------+ +--------+ | IR(VP) | ___ | Seattle| (:::)-. +--------+ +--------+ .-(::::::::) | IR(VP) | .-(::::::::::::)-. | Paris | (:::: The IRON ::::) +--------+ `-(::::::::::::)-' +--------+ `-(::::::)-' +--------+ | IR(VP) | | IR(VP) | | Moscow | +--------+ | Sydney | +--------+ | IR(VP) | +--------+ | Cairo | +--------+ Figure 3: IR(VP) Global Distribution Example 3.3. IR(GW)s - IRON Routers that Connect VPC Networks to the Internet An "IR(GW)" is a tunnel endpoint router that is managed by a VPC and that acts as a gateway between the VPC's IR(VP)s and the non-IRON Internet (i.e., the portion of the Internet used for routing of prefixes that are not derived from VPs). Each VPC configures one or more IR(GW)s which advertise the company's VPs into the IPv4 and/or IPv6 global Internet BGP routing systems. An IR(GW) may be configured on the same physical platform as an IR(VP), or as a separate standalone platform. An IR(GW) will typically be a BGP router that is capable of exchanging both encapsulated packets over the IRON and unencapsulated packets over the native Internet. The role of an IR(GW) is depicted in Figure 4: Templin Expires December 25, 2010 [Page 8] Internet-Draft IRON June 2010 ,-( _)-. .-(_ (_ )-. (_ Internet ) `-(______)-' | +----+---+ | IR(GW) | +----+---+ _|_ (:::)-. .-(::::::::) +--------+ .-(::::::::::::)-. +--------+ | IR(VP) | (:::: The IRON ::::) | IR(VP) | +--------+ `-(::::::::::::)-' +--------+ `-(::::::)-' +--------+ | IR(VP) | +--------+ Figure 4: IR(GW) Connecting VPC to Native Internet 4. IRON Organizational Principles Each VPC in the IRON manages a set of IR(VP)s that service one or more VPs from which EPs are delegated to IR(EP)s. The set of IR(VP)s that service the same VP forms a logical subnetwork for the VP. Each IR(VP) in the logical subnetwork exchanges routing/mapping information via a dynamic routing protocol instance in order to maintain synchronized Forwarding Information Bases (FIBs). The VP company also maintains a set of IR(GW)s that connect the logical subnetworks to the IPv4 and/or IPv6 Internets. Each IR(GW) advertises the VPC's IPv4 VPs into the IPv4 BGP routing system and advertises the VPC's IPv6 VPs into the IPv6 BGP routing system. IR(GW)s forward packets coming from the Internet and destined to an EPA address into the IRON. In the reverse direction, IR(GW)s forward packets coming from the IRON and destined to a locator address into the Internet. In this way, end systems that use locator addresses can communicate with other end systems that use EPA addresses. EUNs establish at least one IR(EP) to connect the EUN to the IRON. The IR(EP) uses encapsulation to forward packets with EP addresses to an IR(VP) belonging to its VP subnetwork as a default router. The IR(VP) then forwards the packets toward their final destination, and returns a SEAL Control Message Protocol (SCMP) redirect message to inform the IR(EP) of a better next hop if necessary. Templin Expires December 25, 2010 [Page 9] Internet-Draft IRON June 2010 The IRON additionally requires a global mapping database to allow IRs to map VPs to locators which are assigned to the interfaces of other IRs. Each VP in the IRON is therefore represented in a globally distributed Master VP database (MVPd). The MVPd is maintained by a globally-managed assigned numbers authority in the same manner as the Internet Assigned Numbers Authority (IANA) currently maintains the master list of all top-level IPv4 and IPv6 delegations. The database can be replicated across multiple servers for load balancing much in the same way that FTP mirror sites are used to manage software distributions. Each VP in the MVPd is encoded as the tuple: "{address family, prefix, prefix-length, FQDN}", where: o "address family" is one of IPv4, IPv6, OSI/CLNP, etc. o "prefix" is the VP, e.g. - 2001:DB8::/32 (IPv6) [RFC3849], 192.2/16 (IPv4) [RFC5737], etc. o "prefix-length" is the length (in bits) of the associated VP o FQDN is a DNS Fully-Qualified Domain Name For each VP entry in the MVPd, the VPC maintains a FQDN in the DNS to map the VP to a list of IR(VP)s that serve it. IR(VP)s in other VPC networks and IR(EP)s that hold EP delegations from the VP discover the mappings by resolving the FQDN into a list of resource records. Each resource record corresponds to an individual IR(VP), and encodes the tuple : "{address family, locator, WGS 84 coordinates}" where "address family" is the address family of the locator, "locator" is the routing locator assigned to an IR(VP) interface, and "WGS 84 coordinates" identify the physical location of the IR(VP). Together, the MVPd and the FQDNs in the global DNS provide sufficient mapping capabilities to support navigation of the IRON. 5. IRON Initialization IRON initialization entails the startup actions of VPC and EUN equipment. The following sections discuss these startups procedures. 5.1. IR(VP) and IR(GW) Initialization Upon startup, each IR(VP) and IR(GW) managed by the VPC discovers the full set of VPs for the IRON by reading the MVPd (see Section 4). These VPs may be IPv4 or IPv6, but they may also be prefixes of other network layer protocols (e.g., OSI/CLNP NSAP [RFC4548], etc.). Each IR(VP) and IR(GW) reads the MVPd from a nearby server upon startup time, and periodically checks the server for deltas since the database was last read. Upon reading the MVPd, each IR(VP) and Templin Expires December 25, 2010 [Page 10] Internet-Draft IRON June 2010 IR(GW) resolves the FQDN corresponding to each VP into a list of locators. Each locator is a routable IPv4 or IPv6 address assigned to an interface of an IR(VP) that serves the VP. For each VP, each IR(VP) and IR(GW) sorts the list of locators to determine a priority ranking (e.g., based on distance from the locator) and inserts each "VP->locator" mapping into its FIB in order of priority. The FIB entries must be configured such that packets with destination addresses covered by the VP are forwarded to the corresponding locator using encapsulation of the inner network layer packet in an outer IP header. This is accomplished by configuring the routing table entry to use the locator addresses as the L2 address corresponding to an imaginary L3 next-hop address. Note that the VP and locator may be of different address families; hence, possible encapsulations include IPv6-in-IPv4, IPv4-in-IPv6, IPv6-in-IPv6, IPv4-in-IPv4, OSI/CLNP-in-IPv6, OSI/CLNP-in-IPv4, etc. After each IR(VP) and IR(GW) reads in the list of VPs and sorts the information accordingly, it is said to be "synchronized with the IRON". Each IR(VP) next installs all EPs derived from its VPs into its FIB based on the mapping information received from each of its EUN customers. 5.2. IR(EP) Initialization Before its first operational use, each IR(EP) must obtain one or more EPs from a VPC along with a FQDN that can be resolved into a list of locators for the IR(VP)s that service the EPs. The IR(EP) must also obtain a certificate and a public/private key pair from the VPC that it can later use to prove ownership of its EPs. This implies that the VPC must run its own key infrastructure to be used only for the purpose of verifying a customer's claimed right to use an EP. Hence, the VPC need not coordinate its key infrastructure with any other organizations. Upon startup, each IR(EP) resolves a FQDN for the VP in order to discover the locators of the company's IR(VPs). (This resolution closely resembles the ISATAP Potential Router List (PRL) resolution procedure [RFC5214].) The IR(EP) then selects the closest subset of these locators (typically 2-4 routers chosen, e.g., based on topological distance) and adds them to a default router list of FIB entries that each points to a VET interface with the locator as the L2 address of the next-hop. The IR(EP) will then use the default router list for forwarding inner packets with EPA source addresses toward the final destination via encapsulation. Templin Expires December 25, 2010 [Page 11] Internet-Draft IRON June 2010 6. IRON Operation Following IRON initialization, IRs engage in the steady-state process of receiving and forwarding packets. All IRs forward encapsulated packets over the IRON using the mechanisms of VET [I-D.templin-intarea-vet] and SEAL [I-D.templin-intarea-seal], while IR(GW)s additionally forward unencapsulated packets to and from the native IPv6 and IPv4 Internets. IRs also use the SEAL Control Message Protocol (SCMP) to test liveness of other IRs and to receive redirect messages informing them of better routes. Each IR operates as specified in the following sub-sections. 6.1. IR(EP) Operation After an IR(EP) is initialized, it must register its EP-to-locator binding with the IR(VP)s that it has selected as default routers by sending periodic beacons with signed certificates and prefix information to prove ownership of its EPs. Each beacon is a SEAL Control Message Protocol (SCMP) Router Solicitation (SRS) message, and will elicit an SCMP Router Advertisement (SRA) message from the IR(VP). If the IR(EP) ceases to receive SRA messages from an IR(VP), it marks the IR(VP) as unreachable and selects a different IR(VP). If the IR(EP) ceases to receive SRA messages from multiple IR(VP)s, it marks the ISP connection as failed/failing and sends its SRS beacons through a different ISP. The IR(EP) also uses the same SRS/ SRA beaconing procedure to inform its IR(VP)s of a change in locator, e.g., due to a mobility event, a change in its primary ISP, etc. When an end system in an EUN has a packet to send, the packet is forwarded through the EUN until it reaches an IR(EP). This source IR(EP) then forwards the packet either to an IR(VP) or to a destination IR(EP). The source IR(EP) first checks its FIB for the longest matching prefix. If the longest matching prefix is more- specific than "default", the source IR(EP) forwards the packet to the next-hop the same as for ordinary IP forwarding. If the longest match is "default", however, the source IR(EP) forwards the packet to one of the IR(VP)s serving as its default routers. The source IR(EP) uses VET and SEAL to encapsulate each forwarded packet in an outer IP header with the IP address of the next-hop IR as the destination address. The source IR(EP) further uses SCMP to test liveness and/or to accept redirect messages from the next-hop IR. When the source IR(EP) receives an SCMP redirect, it checks the identification field of the encapsulated message to verify that the redirect corresponds to a packet that it had previously sent to the neighbor and accepts the redirect if there is a match. Thereafter, subsequent packets forwarded by the source IR(EP) will follow a route-optimized path. Templin Expires December 25, 2010 [Page 12] Internet-Draft IRON June 2010 An IR(EP) that accepts redirects may be redirected by an IR(VP) in its home VPC network to one or more IR(VP)s in a "foreign" network. In that case, the IR(EP) has no way of knowing if these foreign IR(VP)s are reachable and able to process encapsulated packets. In that case, the IR(EP) should select multiple foreign IR(VPs) (e.g., 2-4) and send "live" packets to one of them while sending corresponding "blank" packets to the others. In turn, each foreign IR(VP) accepts and forwards "live" packets, but drops "blank" packets after sending a redirect. In this way, even if the original packet is lost due to congestion or a short-term outage, the IR(EP) will receive a redirect from at least one of the foreign IR(VP)s. 6.2. IR(VP) Operation After an IR(VP) is initialized, it responds to SRSs by sending SRAs as described in Section 6.1. The IR(VP) then propagates the mapping information conveyed in the SRS message to each of its peer IR(VP)s within the logical subnetwork, e.g., via a dynamic routing protocol instance that all of the VP company's IR(VP)s engage in. When the IR(VP) receives an encapsulated packet from another IR, it examines the inner destination address then forwards the packet as follows: o If the inner destination address matches an EP in its FIB, the IR(VP) 'A' re-encapsulates the packet using VET/SEAL and forwards it to the next-hop IR(EP) 'B'. If the source IR 'C' is accepting redirects, 'A' also sends an SCMP redirect message back to 'C'. 'C' will then send subsequent packets directly to 'B'. o If the inner destination address does not match an EP but matches a VP in its FIB, the IR(VP) 'A' re-encapsulates the packet using VET/SEAL and forwards it to the next-hop IR(VP) 'B' . If the source IR 'C' is accepting redirects, 'A' also sends an SCMP redirect message back to 'C'. 'C' will then send subsequent packets directly to 'B'. o If the inner destination address does not match an EP or a VP in the FIB, the IR(VP) forwards the packet to the public Internet either directly or via an IR(GW). An IR(VP) that accepts redirects may need to forward initial packets via the IR(VP)s of a "foreign" network. In that case, the IR(VP) can send a "live" packet in parallel with corresponding "blanks" the same as for an IR(EP). Templin Expires December 25, 2010 [Page 13] Internet-Draft IRON June 2010 6.3. IR(GW) Operation Each VPC must establish one or more IR(GW) routers which advertise the full set of the company's VP's into the IPv4 and/or IPv6 Internet BGP routing systems. The VPs will be seen as ordinary routing information in the BGP, and any packets originating from the non-IRON IPv4 or IPv6 Internet toward a VP will be forwarded into the VPC's network by an IR(GW). When an IR(GW) receives a packet from the non- IRON Internet but destined to an EP destination, it consults its FIB to determine the best next-hop toward the final destination. The IR(GW) then either forwards the packet to an IR(VP) within the home network or acts as an IR(VP) itself to forward the packet further. 6.4. IRON Reference Operating Scenarios The two major reference operating scenarios that may arise for IRON are 1) when both hosts are within separate IRON EUNs, and 2) when one host is within an IRON EUN and the other is within a non-IRON EUN. The following two sections discuss these reference operating scenarios. 6.4.1. Two Hosts in Different IRON EUNs The initial path when both hosts are within separate IRON EUNs involves both the two IR(EPs) and the IR(VP)s of the two VP companies that serve the EUNs. Route optimization based on redirection will allow shortcuts that eliminate the IR(VP)s from the path. Figure 5 depicts the IRON reference operating scenario for communications between hosts within two separate IRON EUNs: Templin Expires December 25, 2010 [Page 14] Internet-Draft IRON June 2010 ________________________________________ .-( )-. .-( +------------+ +------------+ )-. .-( | | | | )-. .( +======>+ IR(VP(A)) +======>+ IR(VP(B)) +=====+ ). .( // | | | | \\ ). .( // +------------+ +------------+ \\ ). ( // <------------- Initial Path --------------> \\ ) ( // \\ ) ( // .-. ................................ .-. \\ ) ( //,-( _)-. . . ,-( _)-\\ ) ( .||_ (_ )-.. <-- Route Optimized Path --> ..-(_ (_ ||. ) ( _|| ISP A .) (. ISP B || )) ( ||-(______)-. .`-(______)||- ) ( || | . v | vv ) ( +-----+ ----+ The IRON +-----+-----+ ) | IR(EP(A)) | (Overlaid on the native Internet) | IR(EP(B)) | +-----+-----+ +-----+-----+ | ( ) | .-. .-( .-) .-. ,-( _)-. .-(________________________)-. ,-( _)-. .-(_ (_ )-. .-(_ (_ )-. (_ IRON EUN A ) (_ IRON EUN B ) `-(______)-' `-(______)-' | | +---+----+ +---+----+ | Host A | | Host B | +--------+ +--------+ Figure 5: Communications Between Two IRON Hosts In this reference scenario, VP companies A and B have established IR(VP)s within the Internet that serve EPs to EUNs. EUN A has procured EP(A) from VPC A, while EUN B has procured EP(B) from VPC B. The hosts in both EUNs have assigned EPAs taken from their corresponding EPs on their EUN-interior interfaces, and the IR(EPs) have assigned locators taken from their ISPs on their WAN interfaces. When Host A in EUN A has a packet to send to Host B in EUN B, normal routing conveys the packet from Host A to IR(EP(A)). If IR(EP(A )) does not have a more-specific route, it encapsulates the packet and forwards it to an IR(VP) managed by VPC A. (The encapsulated packet uses the locator address of IR(EP(A)) as the outer source address and the locator address of IR(VP(A)) as the outer destination address.) When IR(VP(A)) receives the encapsulated packet, it checks its FIB for a route toward the packet's inner destination address (i.e., 'B'). If IR(VP(A)) does not have an entry for EP(B) in its FIB, it consults its full table of VP-to-locator mappings to discover that Templin Expires December 25, 2010 [Page 15] Internet-Draft IRON June 2010 the next-hop toward EP(B) is via IR(VP(B)). IR(VP(A)) then re- encapsulates the packet and sends it to IR(VP(B)) which has an entry for EP(B) in its FIB with IR(EP(B)) as the next hop. IR(VP(B)) then re-encapsulates the packet and sends it to IR(EP(B)), which decapsulates the packet and forwards it via EUN B to Host B. In this process, when an IR(VP) re-encapsulates the packet and forwards it to a next-hop IR, it also returns an SCMP redirect message to the previous hop IR if the previous hop is willing to accept redirects. The previous hop IR will then install a route in its FIB that uses a better next hop. For example, if IR(EP(A)) is accepting redirects IR(VP(A)) will return a redirect message when it forwards a packet to IR(VP(B)). IR(EP(A)) will then send subsequent packets directly to IR(VP(B)), which will return a redirect message when it forwards the packets to IR(EP(B)). Finally, IR(EP(A)) will have an optimized route that lists IR(EP(B)) as the next hop. Another redirection scenario arises when IR(VP(A)) is itself willing to accept redirects. In that case, IR(EP(A)) may discover IR(EP(B)) as a better next hop toward EUN A based solely on a redirect message from IR(VP(A)) and without involving IR(VP(B)). Note however that this may require IR(VP(A)) to carry thousands or even millions of EP entries in its FIB for all EUNs that it has sent packets to recently, which may negatively impact scalability. 6.4.2. Mixed IRON and Non-IRON Hosts When one host is within an IRON EUN and the other is in a non-IRON EUN (i.e., one that connects to the native Internet instead of the IRON), communications must involve an IR(GW) within the IRON host's VPC. Figure 6 depicts the IRON reference operating scenario for communications between Host A in an IRON EUN and Host B in a non-IRON EUN: Templin Expires December 25, 2010 [Page 16] Internet-Draft IRON June 2010 _______________________________________ .-( )-. )-. .-( +-------)----+ )-. .-( | | )-. .( +======>+ IR(GW(A)) +---------------+ ). .( // | | \ ). .( // +--------)---+ \ ). ( // ) \ ) ( // The IRON ) \ ) ( // .-. ) \ .-. ) ( //,-( _)-. ) \ ,-( _)-. ) ( .||_ (_ )-. ) The Native Internet .-|_ (_ )-. ) ( _|| ISP A ) ) (_ | ISP B )) ( ||-(______)-' ) |-(______)-' ) ( || | )-. v | ) ( +-----+ ----+ )-. +-----+-----+ ) | IR(EP(A)) |)-. | Router B | +-----+-----+ +-----+-----+ | ( ) | .-. .-(____________________________________)-. .-. ,-( _)-. ,-( _)-. .-(_ (_ )-. .-(_ (_ )-. (_ IRON EUN A ) (_ non-IRON EUN ) `-(______)-' `-(___B___)-' | | +---+----+ +---+----+ | Host A | | Host B | +--------+ +--------+ Figure 6: Communications Between Hosts in IRON and Non-IRON EUNs In this reference scenario, when Host A sends a flow of packets to Host B, IR(EP(A)) encapsulates and forwards them to an IR(VP) from its VPC as a default router. In the simple case, the IR(VP) also acts as an IR(GW) (depicted here as IR(GW)(A))) that decapsulates packets coming from IR(EP(A)) and forwards them into the native Internet. In this scenario, no route optimization is possible since EUN B is not connected to the IRON. In the reverse direction, when Host B sends a flow of packets to Host A, normal Internet routing conveys the packets over the native Internet to IR(GW(A)) since IR(GW(A)) advertises the VP that covers EP(A) into the BGP routing system. IR(GW(A)) will then encapsulate the packets and forward them over the IRON to IR(EP(A)), which in turn delivers them to Host A. Templin Expires December 25, 2010 [Page 17] Internet-Draft IRON June 2010 6.5. Mobility, Multihoming and Traffic Engineering Considerations While IR(VP)s can be considered as fixed infrastructure, IR(EP)s may need to move between different network points of attachment, connect to multiple ISPs, or explicitly manage their traffic flows. The following sections discuss mobility, multi-homing and traffic engineering considerations for IR(EP)s. 6.5.1. Mobility Management When an IR(EP) changes its network point of attachment (e.g., due to a mobility event), it configures a new locator. The IR(EP) then registers the new EP-to-locator mapping with its VPC by sending SRS messages the same as described in Section 6.2. (For further study are performance characteristics of various mechanisms that could be used to propagate these registration updates within the VPC network.) The IR(EP) also sends Neighbor Advertisement (NA) messages as registration updates to each neighboring IR from which it has received packets recently. The NA message includes the new locator as the outer source address and includes the previous locator within an NA option field. The neighboring IR will update its neighbor cache so that subsequent packets will flow through the new locator. 6.5.2. Multihoming An IR(EP) registers only the locator of its primary ISP with its VPC even if it connects to multiple ISPs. If the IR(EP) later needs to select a different ISP as its primary, it simply repeats the EP-to- locator registration process the same as if it were reacting to a mobility event as described above. 6.5.3. Inbound Traffic Engineering When an IR(EP) receives packets via its primary ISP (i.e., the ISP for which it is currently registered with the VPC), it may wish to balance the load of incoming traffic between multiple ISP connections. In that case, the IR(EP) may send NA messages to certain neighboring IRs the same as in the case of a mobility event as described in Section 6.5.1. In that way, the IR(EP) can manage its incoming traffic flows for best utilization of its multiple ISPs. 6.5.4. Outbound Traffic Engineering IR(EP)s maintain a list of IR(VP)s that serve as default routers for VET interface encapsulation of inner packets with source addresses taken from an EP prefix. IR(EP)s also maintain a list of neighbors on underlying interfaces that serve as default routers for packets Templin Expires December 25, 2010 [Page 18] Internet-Draft IRON June 2010 with non-EP source addresses. If the inner and outer protocols are of different versions (e.g., OSI/CLNP as the inner version and IPv4 as the outer version) then the default routes of both versions are mutually exclusive and no special arrangements are needed. If the inner and outer protocol versions are the same, however (e.g., IPv6 as both the inner and outer protocol) then a special treatment of the default route is necessary. In particular, the IR(EP) must examine the source address of a packet to be forwarded to determine the proper handling of "default". If the packet uses a source address taken from an EP prefix, the IR(EP) forwards it to an IR(VP) using encapsulation via a VET interface; otherwise, the IR(EP) forwards the packet to a next hop on an underlying link. Using this arrangement of default, when an IR(EP) forwards a packet with an EP source address it can forward it to any of its associated IR(VP)s using VET interface encapsulation over any of its underlying interfaces. The choice of underlying interface can be managed, and the source address assigned to the underlying interface will be used as the outer source address of the encapsulated packet. Each encapsulated packet can therefore be directed through the desired ISP using a topologically-correct outer source address. 6.6. Renumbering Considerations As better link layer technologies and service plans emerge, customers will be motivated to select their service providers through healthy competition between ISPs. If a customer's EUN addresses are tied to a specific ISP, however, the customer may be forced to undergo a painstaking EUN renumbering process if it wishes to changes to a different ISP [RFC4192][RFC5887]. When a customer obtains EP prefixes from a VPC, it can change between ISPs seamlessly and without need to renumber. If the VPC itself applies unreasonable costing structures for use of the EPs, however, the customer may be compelled to seek a different VPC and would again be required to confront a renumbering scenario. The strength of the IRON approach therefore lies within a tradeoff between the requirement for VPCs to conduct ethical business practices with reasonable rates and the ability for customers to painlessly renumber their EUNs. 6.7. NAT Traversal Considerations The Internet today consists of a global public IPv4 routing and addressing system with non-IRON EUNs that use either public or private IPv4 addressing. The latter class of EUNs connect to the public IPv4 Internet via Network Address Translators (NATs). When an Templin Expires December 25, 2010 [Page 19] Internet-Draft IRON June 2010 IR(EP) is located behind a NAT, its selects IR(VP)s using the same procedures as for IR(EP)s with public addresses, i.e., it will send SRS messages to IR(VP)s in order to get SRA messages in return. The only requirement is that the IR(EP) must configure its SEAL encapsulation to use a transport protocol that supports NAT traversal, namely UDP. Since the IR(VP) maintains state about its IR(EP) customers, it can discover locator information for each IR(EP) by examining the UDP port number and IPv4 address in the outer headers of SRS messages. When there is a NAT in the path, the UDP port number and IPv4 address in the SRS message will correspond to state in the NAT box and might not correspond to the actual values assigned to the IR(EP). The IR(VP) can then encapsulate packets destined to hosts serviced by the IR(EP) within outer headers that use this IPv4 address and UDP port number. The NAT box will receive the packets, translate the values in the outer headers to match those assigned to the IR(EP), then forward the packets to the IR(EP). 7. Open Research Areas A number of open research areas exist which would need to be explored in taking the IRON concept into large-scale deployment. Considerations for the scalability of Internet Routing due to multihoming, traffic engineering and provider-independent addressing are discussed in [I-D.narten-radir-problem-statement]. Route optimization considerations for mobile networks are found in [RFC5522]. Finally, security implications for tunneling over large- scale Internetworks and feasibility analysis for maintaining a globally distributed mapping service bear further investigation. 8. Related Initiatives IRON builds upon the concepts RANGER architecture [RFC5720], and therefore inherits the same set of related initiatives. Virtual Aggregation (VA) [I-D.ietf-grow-va] and Aggregation in Increasing Scopes (AIS) [I-D.zhang-evolution] provide the basis for the Virtual Prefix concepts. Internet vastly improved plumbing (Ivip) [I-D.whittle-ivip-arch] has contributed valuable insights, including the use of real-time mapping. Templin Expires December 25, 2010 [Page 20] Internet-Draft IRON June 2010 9. IANA Considerations There are no IANA considerations for this document. 10. Security Considerations Security considerations that apply to tunneling in general are discussed in [I-D.ietf-v6ops-tunnel-security-concerns]. Additional considerations that apply also to IRON are discussed in RANGER [RFC5720], VET [I-D.templin-intarea-vet] and SEAL [I-D.templin-intarea-seal]. IRON assumes that the mapping system (including the MVPd and corresponding FQDNs in the DNS) be well-managed and not vulnerable to subversion. This assumption is no different than for the current state of affairs for client-server communications in the Internet today. IR(EP)s require a means for securely registering their EP-to-locator bindings with their VPC and with correspondent nodes. Each VPC provides its customer IR(EP)s with a secure means for registering and re-registering their mappings, and the use of SEAL encapsulation provides a nonce with each packet to allow correspondent nodes to authenticate binding updates from IR(EP)s. 11. Acknowledgements This ideas behind this work have benefited greatly from discussions with colleagues; some of which appear on the RRG and other IRTF/IETF mailing lists. Mohamed Boucadair, Wesley Eddy and Robin Whittle provided review input. 12. References 12.1. Normative References [RFC0791] Postel, J., "Internet Protocol", STD 5, RFC 791, September 1981. [RFC2460] Deering, S. and R. Hinden, "Internet Protocol, Version 6 (IPv6) Specification", RFC 2460, December 1998. Templin Expires December 25, 2010 [Page 21] Internet-Draft IRON June 2010 12.2. Informative References [BGPMON] Analyses, B., "BGPmon.net - Monitoring Your Prefixes, http://bgpmon.net/stat.php", June 2010. [I-D.ietf-grow-va] Francis, P., Xu, X., Ballani, H., Jen, D., Raszuk, R., and L. Zhang, "FIB Suppression with Virtual Aggregation", draft-ietf-grow-va-02 (work in progress), March 2010. [I-D.ietf-v6ops-tunnel-security-concerns] Hoagland, J., Krishnan, S., and D. Thaler, "Security Concerns With IP Tunneling", draft-ietf-v6ops-tunnel-security-concerns-02 (work in progress), March 2010. [I-D.narten-radir-problem-statement] Narten, T., "On the Scalability of Internet Routing", draft-narten-radir-problem-statement-05 (work in progress), February 2010. [I-D.russert-rangers] Russert, S., Fleischman, E., and F. Templin, "Operational Scenarios for IRON and RANGER", draft-russert-rangers-03 (work in progress), June 2010. [I-D.templin-intarea-seal] Templin, F., "The Subnetwork Encapsulation and Adaptation Layer (SEAL)", draft-templin-intarea-seal-15 (work in progress), June 2010. [I-D.templin-intarea-vet] Templin, F., "Virtual Enterprise Traversal (VET)", draft-templin-intarea-vet-15 (work in progress), June 2010. [I-D.whittle-ivip-arch] Whittle, R., "Ivip (Internet Vastly Improved Plumbing) Architecture", draft-whittle-ivip-arch-04 (work in progress), March 2010. [I-D.zhang-evolution] Zhang, B. and L. Zhang, "Evolution Towards Global Routing Scalability", draft-zhang-evolution-02 (work in progress), October 2009. [RFC1070] Hagens, R., Hall, N., and M. Rose, "Use of the Internet as a subnetwork for experimentation with the OSI network Templin Expires December 25, 2010 [Page 22] Internet-Draft IRON June 2010 layer", RFC 1070, February 1989. [RFC3849] Huston, G., Lord, A., and P. Smith, "IPv6 Address Prefix Reserved for Documentation", RFC 3849, July 2004. [RFC4192] Baker, F., Lear, E., and R. Droms, "Procedures for Renumbering an IPv6 Network without a Flag Day", RFC 4192, September 2005. [RFC4271] Rekhter, Y., Li, T., and S. Hares, "A Border Gateway Protocol 4 (BGP-4)", RFC 4271, January 2006. [RFC4548] Gray, E., Rutemiller, J., and G. Swallow, "Internet Code Point (ICP) Assignments for NSAP Addresses", RFC 4548, May 2006. [RFC5214] Templin, F., Gleeson, T., and D. Thaler, "Intra-Site Automatic Tunnel Addressing Protocol (ISATAP)", RFC 5214, March 2008. [RFC5522] Eddy, W., Ivancic, W., and T. Davis, "Network Mobility Route Optimization Requirements for Operational Use in Aeronautics and Space Exploration Mobile Networks", RFC 5522, October 2009. [RFC5720] Templin, F., "Routing and Addressing in Networks with Global Enterprise Recursion (RANGER)", RFC 5720, February 2010. [RFC5737] Arkko, J., Cotton, M., and L. Vegoda, "IPv4 Address Blocks Reserved for Documentation", RFC 5737, January 2010. [RFC5743] Falk, A., "Definition of an Internet Research Task Force (IRTF) Document Stream", RFC 5743, December 2009. [RFC5887] Carpenter, B., Atkinson, R., and H. Flinck, "Renumbering Still Needs Work", RFC 5887, May 2010. Templin Expires December 25, 2010 [Page 23] Internet-Draft IRON June 2010 Author's Address Fred L. Templin (editor) Boeing Research & Technology P.O. Box 3707 MC 7L-49 Seattle, WA 98124 USA Email: fltemplin@acm.org Templin Expires December 25, 2010 [Page 24]