Network Working Group S. Previdi, Ed. Internet-Draft C. Filsfils Intended status: Standards Track Cisco Systems, Inc. Expires: December 11, 2014 B. Field Comcast I. Leung Rogers Communications June 9, 2014 IPv6 Segment Routing Header (SRH) draft-previdi-6man-segment-routing-header-01 Abstract Segment Routing (SR) allows a node to steer a packet through a controlled set of instructions, called segments, by prepending a SR header to the packet. A segment can represent any instruction, topological or service-based. SR allows to enforce a flow through any path (topological, or application/service based) while maintaining per-flow state only at the ingress node to the SR domain. Segment Routing can be applied to the IPv6 data plane with the addition of a new type of Routing Extension Header. This draft describes the Segment Routing Extension Header Type and how it is used by SR capable nodes. Requirements Language The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119 [RFC2119]. 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." Previdi, et al. Expires December 11, 2014 [Page 1] Internet-Draft IPv6 Segment Routing Header (SRH) June 2014 This Internet-Draft will expire on December 11, 2014. Copyright Notice Copyright (c) 2014 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 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. Structure of this document . . . . . . . . . . . . . . . . . 3 2. Segment Routing Documents . . . . . . . . . . . . . . . . . . 3 3. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 3.1. Data Planes supporting Segment Routing . . . . . . . . . 4 3.2. Illustration . . . . . . . . . . . . . . . . . . . . . . 4 4. Abstract Routing Model . . . . . . . . . . . . . . . . . . . 8 4.1. Segment Routing Global Block (SRGB) . . . . . . . . . . . 9 4.2. Traffic Engineering with SR . . . . . . . . . . . . . . . 9 4.3. Segment Routing Database . . . . . . . . . . . . . . . . 10 5. IPv6 Instantiation of Segment Routing . . . . . . . . . . . . 10 5.1. Segment Identifiers (SIDs) and SRGB . . . . . . . . . . . 10 5.1.1. Node-SID . . . . . . . . . . . . . . . . . . . . . . 11 5.1.2. Adjacency-SID . . . . . . . . . . . . . . . . . . . . 11 5.2. Segment Routing Extension Header (SRH) . . . . . . . . . 11 5.2.1. SRH and RFC2460 behavior . . . . . . . . . . . . . . 14 5.2.2. SRH Optimization . . . . . . . . . . . . . . . . . . 15 6. SRH Procedures . . . . . . . . . . . . . . . . . . . . . . . 16 6.1. Segment Routing Operations . . . . . . . . . . . . . . . 16 6.2. Segment Routing Node Functions . . . . . . . . . . . . . 16 6.2.1. Ingress SR Node . . . . . . . . . . . . . . . . . . . 17 6.2.2. Transit Non-SR Capable Node . . . . . . . . . . . . . 18 6.2.3. SR Intra Segment Transit Node . . . . . . . . . . . . 18 6.2.4. SR Segment Endpoint Node . . . . . . . . . . . . . . 18 6.3. FRR Flag Settings . . . . . . . . . . . . . . . . . . . . 19 7. SRH Security . . . . . . . . . . . . . . . . . . . . . . . . 19 7.1. Threat model . . . . . . . . . . . . . . . . . . . . . . 19 7.2. Applicability of RFC 5095 to SRH . . . . . . . . . . . . 20 7.3. Security fields in SRH . . . . . . . . . . . . . . . . . 21 7.4. Nodes within the SR domain . . . . . . . . . . . . . . . 22 Previdi, et al. Expires December 11, 2014 [Page 2] Internet-Draft IPv6 Segment Routing Header (SRH) June 2014 7.5. Nodes outside of the SR domain . . . . . . . . . . . . . 22 7.6. SR path exposure . . . . . . . . . . . . . . . . . . . . 22 8. SR and Tunneling . . . . . . . . . . . . . . . . . . . . . . 23 9. Example Use Case . . . . . . . . . . . . . . . . . . . . . . 23 10. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 26 11. Manageability Considerations . . . . . . . . . . . . . . . . 26 12. Security Considerations . . . . . . . . . . . . . . . . . . . 26 13. Contributors . . . . . . . . . . . . . . . . . . . . . . . . 26 14. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 26 15. References . . . . . . . . . . . . . . . . . . . . . . . . . 26 15.1. Normative References . . . . . . . . . . . . . . . . . . 26 15.2. Informative References . . . . . . . . . . . . . . . . . 26 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 28 1. Structure of this document Section 3 gives an introduction on SR for IPv6 networks. Section 4 describes the Segment Routing abstract model. Section 5 defines the Segment Routing Header (SRH) allowing instantiation of SR over IPv6 dataplane. Section 6 details the procedures of the Segment Routing Header. Section 7 describes the security aspect of SR-IPv6. 2. Segment Routing Documents Segment Routing terminology is defined in [I-D.filsfils-spring-segment-routing]. Segment Routing use cases are described in [I-D.filsfils-spring-segment-routing-use-cases]. Segment Routing IPv6 use cases are described in [I-D.ietf-spring-ipv6-use-cases]. Segment Routing protocol extensions are defined in [I-D.ietf-isis-segment-routing-extensions], and [I-D.psenak-ospf-segment-routing-ospfv3-extension]. 3. Introduction Segment Routing (SR), defined in [I-D.filsfils-spring-segment-routing], allows a node to steer a packet through a controlled set of instructions, called segments, by prepending a SR header to the packet. A segment can represent any Previdi, et al. Expires December 11, 2014 [Page 3] Internet-Draft IPv6 Segment Routing Header (SRH) June 2014 instruction, topological or service-based. SR allows to enforce a flow through any path (topological or service/application based) while maintaining per-flow state only at the ingress node to the SR domain. Segments can be derived from different components: IGP, BGP, Services, Contexts, Locators, etc. The list of segment forming the path is called the Segment List and is encoded in the packet header. SR allows the use of strict and loose source based routing paradigms without requiring any additional signaling protocols in the infrastructure hence delivering an excellent scalability property. The source based routing model described in [I-D.filsfils-spring-segment-routing] is inherited from the ones proposed by [RFC1940] and [RFC2460]. The source based routing model offers the support for explicit routing capability. 3.1. Data Planes supporting Segment Routing Segment Routing (SR), can be instantiated over MPLS ([I-D.filsfils-spring-segment-routing-mpls]) and IPv6. This document defines its instantiation over the IPv6 data-plane based on the use- cases defined in [I-D.ietf-spring-ipv6-use-cases]. Segment Routing for IPv6 (SR-IPv6) is required in networks where MPLS data-plane is not used or, when combined with SR-MPLS, in networks where MPLS is used in the core and IPv6 is used at the edge (home networks, datacenters). This document defines a new type of Routing Header (originally defined in [RFC2460]) called the Segment Routing Header (SRH) in order to convey the Segment List in the packet header as defined in [I-D.filsfils-spring-segment-routing]. Mechanisms through which segment are known and advertised are outside the scope of this document. 3.2. Illustration In the context of Figure 1 where all the links have the same IGP cost, let us assume that a packet P enters the SR domain at an ingress edge router I and that the operator requests the following requirements for packet P: The local service S offered by node B must be applied to packet P. The links AB and CE cannot be used to transport the packet P. Any node N along the journey of the packet should be able to determine where the packet P entered the SR domain and where it Previdi, et al. Expires December 11, 2014 [Page 4] Internet-Draft IPv6 Segment Routing Header (SRH) June 2014 will exit. The intermediate node should be able to determine the paths from the ingress edge router to itself, and from itself to the egress edge router. Per-flow State for packet P should only be created at the ingress edge router. The operator can forbid, for security reasons, anyone outside the operator domain to exploit its intra-domain SR capabilities. I---A---B---C---E \ | / \ / \ | / F \|/ D Figure 1: An illustration of SR properties All these properties may be realized by instructing the ingress SR edge router I to push the following abstract SR header on the packet P. +---------------------------------------------------------------+ | | | | Abstract SR Header | | | | | | {SD, SB, SS, SF, SE}, Ptr, SI, SE | Transported | | ^ | | Packet | | | | | P | | +---------------------+ | | | | | +---------------------------------------------------------------+ Figure 2: Packet P at node I The abstract SR header contains a source route encoded as a list of segments {SD, SB, SS, SF, SE}, a pointer (Ptr) and the identification of the ingress and egress SR edge routers (segments SI and SE). A segment identifies a topological instruction or a service instruction. A segment can either be global or local. The instruction associated with a global segment is recognized and executed by any SR-capable node in the domain. The instruction associated with a local segment is only supported by the specific node that originates it. Let us assume some IGP (i.e.: ISIS and OSPF) extensions to define a "Node Segment" as a global instruction within the IGP domain to Previdi, et al. Expires December 11, 2014 [Page 5] Internet-Draft IPv6 Segment Routing Header (SRH) June 2014 forward a packet along the shortest path to the specified node. Let us further assume that within the SR domain illustrated in Figure 1, segments SI, SD, SB, SE and SF respectively identify IGP node segments to I, D, B, E and F. Let us assume that node B identifies its local service S with local segment SS. With all of this in mind, let us describe the journey of the packet P. The packet P reaches the ingress SR edge router. I pushes the SR header illustrated in Figure 2 and sets the pointer to the first segment of the list (SD). SD is an instruction recognized by all the nodes in the SR domain which causes the packet to be forwarded along the shortest path to D. Once at D, the pointer is incremented and the next segment is executed (SB). SB is an instruction recognized by all the nodes in the SR domain which causes the packet to be forwarded along the shortest path to B. Once at B, the pointer is incremented and the next segment is executed (SS). SS is an instruction only recognized by node B which causes the packet to receive service S. Once the service applied, the next segment is executed (SF) which causes the packet to be forwarded along the shortest path to F. Once at F, the pointer is incremented and the next segment is executed (SE). SE is an instruction recognized by all the nodes in the SR domain which causes the packet to be forwarded along the shortest path to E. E then removes the SR header and the packet continues its journey outside the SR domain. All of the requirements are met. First, the packet P has not used links AB and CE: the shortest-path from I to D is I-A-D, the shortest-path from D to B is D-B, the shortest-path from B to F is B-C-F and the shortest-path from F to E Previdi, et al. Expires December 11, 2014 [Page 6] Internet-Draft IPv6 Segment Routing Header (SRH) June 2014 is F-E, hence the packet path through the SR domain is I-A-D-B-C-F-E and the links AB and CE have been avoided. Second, the service S supported by B has been applied on packet P. Third, any node along the packet path is able to identify the service and topological journey of the packet within the SR domain. For example, node C receives the packet illustrated in Figure 3 and hence is able to infer where the packet entered the SR domain (SI), how it got up to itself {SD, SB, SS, SE}, where it will exit the SR domain (SE) and how it will do so {SF, SE}. +---------------------------------------------------------------+ | | | | SR Header | | | | | | {SD, SB, SS, SF, SE}, Ptr, SI, SE | Transported | | ^ | | Packet | | | | | P | | +--------+ | | | | | +---------------------------------------------------------------+ Figure 3: Packet P at node C Fourth, only node I maintains per-flow state for packet P. The entire program of topological and service instructions to be executed by the SR domain on packet P is encoded by the ingress edge router I in the SR header in the form of a list of segments where each segment identifies a specific instruction. No further per-flow state is required along the packet path. The per-flow state is in the SR header and travels with the packet. Intermediate nodes only hold states related to the IGP global node segments and the local IGP adjacency segments. These segments are not per-flow specific and hence scale very well. Typically, an intermediate node would maintain in the order of 100's to 1000's global node segments and in the order of 10's to 100 of local adjacency segments. Typically the SR IGP forwarding table is expected to be much less than 10000 entries. Fifth, the SR header is inserted at the entrance to the domain and removed at the exit of the operator domain. For security reasons, the operator can forbid anyone outside its domain to use its intra- domain SR capability. Previdi, et al. Expires December 11, 2014 [Page 7] Internet-Draft IPv6 Segment Routing Header (SRH) June 2014 4. Abstract Routing Model At the entrance of the SR domain, the ingress SR edge router pushes the SR header on top of the packet. At the exit of the SR domain, the egress SR edge router removes the SR header. The abstract SR header contains an ordered list of segments, a pointer identifying the next segment to process and the identifications of the ingress and egress SR edge routers on the path of this packet. The pointer identifies the segment that MUST be used by the receiving router to process the packet. This segment is called the active segment. A property of SR is that the entire source route of the packet, including the identity of the ingress and egress edge routers is always available with the packet. This allows for interesting accounting and service applications. We define three SR-header operations: "PUSH": an SR header is pushed on an IP packet, or additional segments are added at the head of the segment list. The pointer is moved to the first entry of the added segments. "NEXT": the active segment is completed, the pointer is moved to the next segment in the list. "CONTINUE": the active segment is not completed, the pointer is left unchanged. In the future, other SR-header management operations may be defined. As the packet travels through the SR domain, the pointer is incremented through the ordered list of segments and the source route encoded by the SR ingress edge node is executed. A node processes an incoming packet according to the instruction associated with the active segment. Any instruction might be associated with a segment: for example, an intra-domain topological strict or loose forwarding instruction, a service instruction, etc. At minimum, a segment instruction must define two elements: the identity of the next-hop to forward the packet to (this could be the same node or a context within the node) and which SR-header management operation to execute. Previdi, et al. Expires December 11, 2014 [Page 8] Internet-Draft IPv6 Segment Routing Header (SRH) June 2014 Each segment is known in the network through a Segment Identifier (SID). The terms "segment" and "SID" are interchangeable. 4.1. Segment Routing Global Block (SRGB) In the SR abstract model, a segment is identified by a Segment Routing Identifier (SID). The SR abstract model doesn't mandate a specific format for the SID (IPv6 address or other formats). In Segment Routing IPv6 the SID is an IPv6 address. Therefore, the SRGB is materialized by the global IPv6 address space which represents the set of IPv6 routable addresses in the SR domain. The following rules apply: o Each node of the SR domain MUST be configured with the Segment Routing Global Block (SRGB). o All global segments must be allocated from the SRGB. Any SR capable node MUST be able to process any global segment advertised by any other node within the SR domain. o Any segment outside the SRGB has a local significance and is called a "local segment". An SR-capable node MUST be able to process the local segments it originates. An SR-capable node MUST NOT support the instruction associated with a local segment originated by a remote node. 4.2. Traffic Engineering with SR An SR Traffic Engineering policy is composed of two elements: a flow classification and a segment-list to prepend on the packets of the flow. In SR, this per-flow state only exists at the ingress edge node where the policy is defined and the SR header is pushed. It is outside the scope of the document to define the process that leads to the instantiation at a node N of an SR Traffic Engineering policy. [I-D.filsfils-spring-segment-routing-use-cases] illustrates various alternatives: N is deriving this policy automatically (e.g. FRR). N is provisioned explicitly by the operator. N is provisioned by a controller or server (e.g.: SDN Controller). Previdi, et al. Expires December 11, 2014 [Page 9] Internet-Draft IPv6 Segment Routing Header (SRH) June 2014 N is provisioned by the operator with a high-level policy which is mapped into a path thanks to a local CSPF-based computation (e.g. affinity/SRLG exclusion). N could also be provisioned by other means. [I-D.filsfils-spring-segment-routing-use-cases] explains why the majority of use-cases require very short segment-lists, hence minimizing the performance impact, if any, of inserting and transporting the segment list. A SDN controller, which desires to instantiate at node N an SR Traffic Engineering policy, collects the SR capability of node N such as to ensure that the policy meets its capability. 4.3. Segment Routing Database The Segment routing Database (SRDB) is a set of entries where each entry is identified by a SID. The instruction associated with each entry at least defines the identity of the next-hop to which the packet should be forwarded and what operation should be performed on the SR header (PUSH, CONTINUE, NEXT). +---------+-----------+---------------------------------+ | Segment | Next-Hop | SR Header operation | +---------+-----------+---------------------------------+ | Sk | M | CONTINUE | | Sj | N | NEXT | | Sl | NAT Srvc | NEXT | | Sm | FW srvc | NEXT | | Sn | Q | NEXT | | etc. | etc. | etc. | +---------+-----------+---------------------------------+ Figure 4: SR Database Each SR-capable node maintains its local SRDB. SRDB entries can either derive from local policy or from protocol segment advertisement. 5. IPv6 Instantiation of Segment Routing 5.1. Segment Identifiers (SIDs) and SRGB Segment Routing, as described in [I-D.filsfils-spring-segment-routing], defines Node-SID and Adjacency-SID. When SR is used over IPv6 data-plane the following applies. Previdi, et al. Expires December 11, 2014 [Page 10] Internet-Draft IPv6 Segment Routing Header (SRH) June 2014 The SRGB is the global IPv6 address space which represents the set of IPv6 routable addresses in the SR domain. Node SIDs are IPv6 addresses part of the SRGB (i.e.: routable addresses). Adjacency-SIDs are IPv6 addresses which may not be part of the global IPv6 address space. 5.1.1. Node-SID The Node-SID identifies a node. With SR-IPv6 the Node-SID is an IPv6 prefix that the operator configured on the node and that is used as the node identifier. Typically, in case of a router, this is the IPv6 address of the node loopback interface. Therefore, SR-IPv6 does not require any additional SID advertisement for the Node Segment. The Node-SID is in fact the IPv6 address of the node. 5.1.2. Adjacency-SID In the SR architecture defined in [I-D.filsfils-spring-segment-routing] the Adjacency-SID (or Adj-SID) identifies a given interface and is a local segment (i.e.: the value has significance only to the node advertising the Adj-SID). A node may advertise one (or more) Adj-SIDs allocated to a given interface so to force the forwarding of the packet (when received with that particular Adj-SID) into the interface regardless the routing entry for the packet destination. The semantic of the Adj-SID is: Send out the packet to the interface this prefix is allocated to. When SR is applied to IPv6, any SID is in an IPv6 address and therefore, an Adj-SID may have a global significance (i.e.: when the IPv6 address representing the SID is a global address). In other words, a node that advertises the Adj-SID in the form of a global IPv6 address representing the link/adjacency the packet has to be forwarded to, will apply to the Adj-SID a global significance. Advertisement of Adj-SID may be done using multiple mechanisms among which the ones described in ISIS and OSPF protocol extensions: [I-D.ietf-isis-segment-routing-extensions] and [I-D.psenak-ospf-segment-routing-ospfv3-extension]. The distinction between local and global significance of the Adj-SID is given in the encoding of the Adj-SID advertisement. 5.2. Segment Routing Extension Header (SRH) A new type of the Routing Header (originally defined in [RFC2460]) is defined: the Segment Routing Header (SRH) which has a new Routing Type, to be assigned by IANA. Previdi, et al. Expires December 11, 2014 [Page 11] Internet-Draft IPv6 Segment Routing Header (SRH) June 2014 As an example, if an explicit path is to be constructed across a core network running ISIS or OSPF, the segment list will contain SIDs representing the nodes across the path (loose or strict) which, usually, are the IPv6 loopback interface address of each node. If the path is across service or application entities, the segment list contains the IPv6 addresses of these services or application instances. The Segment Routing Header (SRH) is defined as follows: 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Next Header | Hdr Ext Len | Routing Type | Next Segment | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Last Segment | Flags | HMAC Key ID | Policy List Flags | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | | Segment List[0] (128 bits ipv6 address) | | | | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | | | ... | | | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | | Segment List[n] (128 bits ipv6 address) | | | | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | | Policy List[0] (128 bits ipv6 address) | | (optional) | | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | | Policy List[1] (128 bits ipv6 address) | | (optional) | | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | | Policy List[2] (128 bits ipv6 address) | | (optional) | | | Previdi, et al. Expires December 11, 2014 [Page 12] Internet-Draft IPv6 Segment Routing Header (SRH) June 2014 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | | | | | | HMAC (256 bits) | | (optional) | | | | | | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ where: o Next Header: 8-bit selector. Identifies the type of header immediately following the SRH. o Hdr Ext Len: 8-bit unsigned integer, is the length of the SRH header in 8-octet units, not including the first 8 octets. o Routing Type: TBD, to be assigned by IANA. o Next Segment (originally defined as "Segments Left" in [RFC2460]): offset (in multiple of 8 octets not including the first 8 octets) of the next active segment (according to terminology defined in [I-D.filsfils-spring-segment-routing]) in the SRH. Note that this differs from the semantic defined in the Routing Header specification ([RFC2460] defines it as "Segments Left"). Therefore, in the Segment Routing context, the "Segments Left" field is renamed as "Next Segment". o Last Segment: offset (in multiple of 8 octets not including the first 8 octets) of the last segment of the path in the SRH. o Flags: 4 bits of flags. Two flags are defined: Bit-0: Clean-up Bit. Set when the SRH has to be removed from the packet when packet reaches the last segment. Bit-1: Protected Bit. Set when the packet has been rerouted through FRR mechanism by a SR endpoint node. See Section 6.3 for more details. o HMAC Key ID and HMAC field are defined in Section 7. o Policy List flags. Define the type of the IPv6 addresses encoded into the Policy List (see below). The following have been defined: Previdi, et al. Expires December 11, 2014 [Page 13] Internet-Draft IPv6 Segment Routing Header (SRH) June 2014 Bits 0-2: determine the type of the first element after the segment list. Bits 3-5: determine the type of the second element. Bits 6-8: determine the type of the third element. Bits 9-11: determine the type of the fourth element. The following values are used for the type: 0x0: Not present. If value is set to 0x0, it means the element represented by these bits is not present. 0x1: Ingress SR PE address. 0x2: Egress SR PE address. 0x3: Original Source Address. o Segment List[n]: 128 bit IPv6 addresses representing the nth segment of the path. o Policy List. Optional addresses representing specific nodes in the SR path such as: Ingress SR PE: IPv6 address representing the SR node which has imposed the SRH (SR domain ingress). Egress SR PE: IPv6 address representing the egress SR domain node. Original Source Address: IPv6 address originally present in the SA field of the packet. The segments in the Policy List are encoded after the segment list and they are optional. If none are in the SRH, all bits of the Policy List Flags MUST be set to 0x0. 5.2.1. SRH and RFC2460 behavior The SRH being a new type of the Routing Header, it also has the same properties: SHOULD only appear once in the packet. Only the router whose address is in the DA field of the packet header MUST inspect the SRH. Previdi, et al. Expires December 11, 2014 [Page 14] Internet-Draft IPv6 Segment Routing Header (SRH) June 2014 Therefore, Segment Routing in IPv6 networks implies that the segment identifier (i.e.: the IPv6 address of the segment) is moved into the DA of the packet. The DA of the packet changes at each segment termination/completion and therefore the original DA of the packet MUST be encoded as the last segment of the path. As illustrated in Section 3.2, nodes that are within the path of a segment will forward packets based on the DA of the packet without inspecting the SRH. This ensures full interoperability between SR- capable and non-SR-capable nodes. 5.2.2. SRH Optimization In order to optimize the way the SRH and, more precisely, the Segment List is processed by SR nodes, it is desirable that most of the necessary information of the SL is placed at the top of the list so to avoid reading the whole content of the SRH prior to make forwarding decisions. With this in mind, when the SRH is created and the segment list is inserted, the order of the segments in the segment list is as follows: o The Next Segment field points to the next segment to be examined (offset within the SRH). o The first segment being encoded in the DA by the ingress node, it doesn't need to sit in the first position of the list. o Hence, the first element of the segment list is the second segment of the path so that, when the packet reaches the end of the first segment, the node inspecting the SRH will find the second segment at the beginning of the segment list. o The other segments of the path are encoded sequentially after the second segment. o The last segment of the path is the original DA address. o The last segment in the Segment List is used to encode the first segment. This segment will never be inspected anyway (at least not for forwarding purposes). Previdi, et al. Expires December 11, 2014 [Page 15] Internet-Draft IPv6 Segment Routing Header (SRH) June 2014 6. SRH Procedures In this section we describe the different procedures on the SRH. 6.1. Segment Routing Operations When Segment Routing is instantiated over the IPv6 data plane the following applies: o The segment list is encoded in the SRH. o The active segment is in the destination address of the packet. o The Segment Routing CONTINUE operation (as described in [I-D.filsfils-spring-segment-routing]) is implemented as a regular/plain IPv6 operation consisting of DA based forwarding. o The NEXT operation is implemented through the update of the DA with the value represented by the Next Segment field in the SRH. o The PUSH operation is implemented through the insertion of the SRH or the insertion of additional segments in the SRH segment list. 6.2. Segment Routing Node Functions SR packets are forwarded to segments endpoints (i.e.: nodes whose address is in the DA field of the packet). The segment endpoint, when receiving a SR packet destined to itself, does: o Inspect the SRH. o Determine the next segment. o Update the SRH (or, if requested, remove the SRH from the packet). o Update the DA. o Send the packet to the next segment. The procedures applied to the SRH are related to the node function. Following nodes functions are defined: Ingress SR Node. Transit Non-SR Node. Transit SR Intra Segment Node. Previdi, et al. Expires December 11, 2014 [Page 16] Internet-Draft IPv6 Segment Routing Header (SRH) June 2014 SR Endpoint Node. 6.2.1. Ingress SR Node Ingress Node can be a router at the edge of the SR domain or a SR- capable host. The ingress SR node may obtain the segment list by either: Local path computation. Local configuration. Interaction with an SDN controller delivering the path as a complete SRH. Any other mechanism (mechanisms through which the path is acquired are outside the scope of this document). When creating the SRH (either at ingress node or in the SDN controller) the following is done: Next Header and Hdr Ext Len fields are set according to [RFC2460]. Routing Type field is set as TBD (SRH). The DA of the packet is set with the address of the FIRST segment of the path. Next Segment field contains the offset of the SECOND segment of the path which is encoded in the FIRST position of the segment list. The segment list is encoded as follows: The first element of the list contains the second segment (as stated above). All subsequent segments are encoded following the second segment. The original DA of the packet is encoded as the last segment of the path (which is NOT the last segment of the segment list). The last segment of the segment list is the FIRST segment of the path. Last Segment field contains the offset of the last segment of the path (i.e.: the original DA of the packet). The packet is sent out to the first segment. Previdi, et al. Expires December 11, 2014 [Page 17] Internet-Draft IPv6 Segment Routing Header (SRH) June 2014 6.2.1.1. Security at Ingress The procedures related to the Segment Routing security are detailed in Section 7. In the case where the SR domain boundaries are not under control of the network operator (e.g.: when the SR domain edge is in a home network), it is important to authenticate and validate the content of any SRH being received by the network operator. In such case, the security procedure described in Section 7 is to be used. The ingress node (e.g.: the host in the home network) requests the SRH from a control system (e.g.: an SDN controller) which delivers the SRH with its HMAC signature on it. Then, the home network host can send out SR packets (with an SRH on it) that will be validated at the ingress of the network operator infrastructure. The ingress node of the network operator infrastructure, is configured in order to validate the incoming SRH HMACs in order to allow only packets having correct SRH according to their SA/DA addresses. 6.2.2. Transit Non-SR Capable Node SR is interoperable with plain IPv6 forwarding. Any non SR-capable node will forward SR packets solely based on the DA. There's no SRH inspection. This ensures full interoperability between SR and non-SR nodes. 6.2.3. SR Intra Segment Transit Node Only the node whose address is in DA inspects and processes the SRH (according to [RFC2460]). An intra segment transit node is not in the DA and its forwarding is based on DA and its SR-IPv6 FIB. 6.2.4. SR Segment Endpoint Node The SR segment endpoint node is the node whose address is in the DA. The segment endpoint node inspects the SRH and does: Previdi, et al. Expires December 11, 2014 [Page 18] Internet-Draft IPv6 Segment Routing Header (SRH) June 2014 1. IF DA = myself (segment endpoint) 2. IF Next Segment <> Last Segment THEN update DA with Next Segment increment Next Segment 3. ELSE IF Last Segment <> DA THEN update DA with Next Segment IF Clean-up bit is set THEN remove the SRH 4. ELSE give the packet to next PID (application) End of processing. 5. Forward the packet out 6.3. FRR Flag Settings A node supporting SR and doing Fast Reroute (as described in [I-D.filsfils-spring-segment-routing-use-cases], when rerouting packets through FRR mechanisms, SHOULD inspect the rerouted packet header and look for the SRH. If the SRH is present, the rerouting node SHOULD set the Protected bit on all rerouted packets. 7. SRH Security This section analyses the security threat model as well as the security issues and proposed solutions related to the new routing header for segment routing. The SRH is simply another version of the routing header as described in [RFC2460] and is: o inserted when entering the segment routing domain which could be done by a node or by a router; o inspected and acted upon when reaching the destination address of the IP header. Routers on the path that simply forward an IPv6 packet (i.e. the IPv6 destination address is none of theirs) will never inspect and process the SRH. Routers whose one interface IPv6 address equals the destination address field of the SRH will have to parse the SRH and, if supported and if the local configuration allows it, will act on the SRH. 7.1. Threat model Using a SRH, which is basically source routing, has some well-known security issues as described in [RFC4942] section 2.1.1 and [RFC5095]: Previdi, et al. Expires December 11, 2014 [Page 19] Internet-Draft IPv6 Segment Routing Header (SRH) June 2014 o amplification attacks: where a packet could be forged in such a way to cause looping among a set of SR-enabled routers causing unnecessary traffic, hence a denial of service against bandwidth; o reflection attack: where a hacker could force an intermediate node to appear as the immediate attacker, hence hiding the real attacker from naive forensic; o bypass attack: where an intermediate node could be used as a stepping stone (for example in a DMZ) to attack another host (for example in the datacenter or any back-end server. These security issues did lead to obsoleting the routing-header type 0, RH-0, with [RFC5095] because: o it was assumed to be inspected and acted upon by default by each and every router on the Internet; o it contained multiple segments in the payload. Therefore, if intermediate nodes ONLY act on valid and authorized SRH, then there is no security threat similar to RH-0. But as SR is used for added value services, there is also a need to prevent non-participating nodes to use those services; this is called 'service stealing prevention'. The SRH also contains all IPv6 addresses of intermediate SR-nodes, this obviously reveals those addresses to the potentially hostile attackers if those attackers are on the path. 7.2. Applicability of RFC 5095 to SRH In the segment routing architecture described in [I-D.filsfils-spring-segment-routing] there are basically two kinds of nodes (routers and hosts): o nodes within the segment routing domain, which is within one single administrative domain, i.e., where all nodes are trusted anyway else the damage caused by those nodes could be worse than amplification attacks: traffic interception and man-in-the-middle attacks, more server DoS by dropping packets, and so on. o Nodes outside of the segment routing domain, which is outside of the administrative segment routing domain hence they cannot be trusted because there is no physical security for those nodes, i.e., they can be replaced by hostile nodes or can be coerced in wrong behaviors. Previdi, et al. Expires December 11, 2014 [Page 20] Internet-Draft IPv6 Segment Routing Header (SRH) June 2014 7.3. Security fields in SRH The security-related fields in SRH are: o HMAC Key-id, 8 bits wide, if HMAC key-id is null, then there is no HMAC field; o HMAC, 256 bits wide. The HMAC field is the output of the hash of the concatenation of: o the source IPv6 address; o last segment field, an octet whose bit-0 is the clean-up bit flag and others are 0, HMAC key-id, all addresses in the Segment List; o a pre-shared secret between SR nodes in the SR domain (routers, controllers, ...); o if required by the hash algorithm a pad field filled with 0. The purpose of the HMAC field is to verify the validity, the integrity and the authorization of the SRH itself. If an outsider of the SR domain does not have access to a current pre-shared secret, then it cannot compute the right HMAC field and the first SR router on the path processing the SRH and configured to check the validity of the HMAC will simply reject the packet. The HMAC field is located at the end of the SRH simply because only the router on the ingress of the SR domain needs to process it, then all other SR nodes can ignore it (based on local policy) because they can trust the upstream router. This is to speed up forwarding operations because some hardware platforms can only parse in hardware so many bytes. The HMAC Key-id field allows for the simultaneous existence of several hash algorithms (SHA-256, SHA3-256 ... or future ones) as well as pre-shared keys. This allows for pre-shared key roll-over when two pre-shared keys are supported for a while when all SR nodes converged to a fresher pre-shared key. The HMAC key-id is opaque, i.e., it has no syntax except as an index to the right combination of pre-shared key and hash algorithm. It also allows for interoperation among different SR domains if allowed by local policy. When a specific SRH is linked to a time-related service (such as turbo-QoS for a 1-hour period) where the DA, SID are identical, then it is important to refresh the shared-secret frequently as the HMAC validity period expires only when the HMAC key-id and its associated Previdi, et al. Expires December 11, 2014 [Page 21] Internet-Draft IPv6 Segment Routing Header (SRH) June 2014 shared-secret expires. How HMAC key-id and pre-shared secret are synchronized between participating nodes in the SR domain is outside of the scope of this document ([RFC6407] GDOI could be a basis). 7.4. Nodes within the SR domain Those nodes can be trusted to generate SRH and to process SRH received on interfaces that are part of the SR domain. These nodes MUST drop all packets received on an interface that is not part of the SR domain and containing a SRH whose HMAC field cannot be validated by local policies. This includes obviously packet with a SRH generated by a non-cooperative SR domain. If the validation fails, then these packets MUST be dropped, ICMP error messages (parameter problem) SHOULD be generated (but rate limited) and SHOULD be logged. 7.5. Nodes outside of the SR domain Nodes outside of the SR domain cannot be trusted for physical security; hence, they need to request by some means (outside of the scope of this document) a complete SRH for each new connection (i.e. new destination address). The SRH MUST include a HMAC key-id and HMAC field which is computed correctly (see Section 7.3). When an outside node sends a packet with an SRH and towards a SR ingress node, the packet MUST contain the HMAC key-id and HMAC field and the SR ingress node MUST be the destination address. The ingress SR router, i.e., the router with an interface address equals to the destination address, MUST verify the HMAC field with respect to the HMAC key-id. If the validation is successful, then the packet is simply forwarded as usual for a SR packet. As long as the packet travels within the SR domain, no further HMAC check needs to be done. Subsequent routers in the SR domain MAY verify the HMAC field when they process the SRH (i.e. when they are the destination). If the validation fails, then this packet MUST be dropped, an ICMP error message (parameter problem) SHOULD be generated (but rate limited) and SHOULD be logged. 7.6. SR path exposure As the intermediate SR nodes addresses appears in the SRH, if this SRH is visible to an outside then he/she could reuse this knowledge to launch an attack on the intermediate SR nodes or get some insider Previdi, et al. Expires December 11, 2014 [Page 22] Internet-Draft IPv6 Segment Routing Header (SRH) June 2014 knowledge on the topology. This is especially applicable when the path between the source node and the first SR-node in the domain is on the public Internet. The first remark is to state that 'security by obscurity' is never enough; in other words, the security policy of the SR domain MUST assume that the internal topology and addressing is known by the attacker. A simple traceroute will also give the same information (with even more information as all intermediate nodes between SID will also be exposed). IPsec Encapsulating Security Payload (RFC 4303) cannot be use to protect the SRH as per RFC 4303 the ESP header must appear after any routing header (including SRH). To prevent a user to leverage the gained knowledge by intercepting SRH, it it recommended to apply an infrastructure Access Control List (iACL) at the edge of the SR domain. This iACL will drop all packets from outside the SR-domain whose destination is any address of any router inside the domain. This security policy should be tuned for local operations. 8. SR and Tunneling Encapsulation can be realized in two different ways with SR-IPv6: Outer encapsulation. SRH with SA/DA original addresses. Outer encapsulation tunneling is the traditional method where an additional IPv6 header is prepended to the packet. The original IPv6 header being encapsulated, everything is preserved and the packet is switched/routed according to the outer header (that could contain a SRH). SRH allows encoding both original SA and DA and therefore, hence an operator may decide to change the SA/DA at ingress and restore them at egress. This can be achieved without outer encapsulation, by changing SA/DA and encoding the original values in the Segment List (the last segment of the path being the original DA) and in the Policy List (original SA). 9. Example Use Case A more detailed description of use cases are available in [I-D.ietf-spring-ipv6-use-cases]. In this section, a simple SR-IPv6 example is illustrated. Previdi, et al. Expires December 11, 2014 [Page 23] Internet-Draft IPv6 Segment Routing Header (SRH) June 2014 In the topology described in Figure 6 it is assumed an end-to-end SR deployment. Therefore SR is supported by all nodes from A to J. Home Network | Backbone | Datacenter | | | +---+ +---+ +---+ | +---+ | +---|---| C |---| D |---| E |---|---| I |---| | | +---+ +---+ +---+ | +---+ | | | | | | | | | +---+ +---+ +---+ | | | | | | |--| X | | A |---| B | | +---+ +---+ +---+ | +---+ | +---+ +---+ +---+ | | F |---| G |---| H |---|---| J |---| | +---+ +---+ +---+ | +---+ | | | | +-----------+ | SDN | | Orch/Ctlr | +-----------+ Figure 6: Sample SR topology The following workflow applies to packets sent by host A and destined to server X. Previdi, et al. Expires December 11, 2014 [Page 24] Internet-Draft IPv6 Segment Routing Header (SRH) June 2014 . Host A sends a request for a path to server X to the SDN controller or orchestration system. . The SDN controller/orchestrator builds a SRH with: . Segment List: C, F, J, X . HMAC that satisfies the requirements expressed in the request by host A and based on policies applicable to host A. . Host A receives the SRH and insert it into the packet. The packet has now: . SA: A . DA: C . SRH with . SL: F,J,X,C . PL: C (ingress), J (egress) Note that X is the last segment and C is the first segment (encoded at the end of the SL). . When packet arrives in C (first segment), C does: . Validate the HMAC of the SRH. . Update the DA with the next segment (found in SRH): DA is set to F. . Forward the packet to F. . Packet arrives in F which inspects the SRH and find the next segment: . DA is set to J. . Packet travels across G and H nodes which do plain IPv6 forwarding based on DA. No inspection of SRH needs to be done in these nodes. However, any SR capable node is allowed to set the Protected bit in case of FRR protection. . Packet arrives in J where two options are available depending on the settings of the cleanup bit set in the SRH: . If the cleanup bit is set, then node J will strip out the SRH from the packet, set the DA as X and send the packet out. . If the clean-up bit is not set, the DA is set to X and the packet is sent out with the SRH. The packet arrives in the server that may or may not support SR. The return traffic, from server to host, may be sent using the same procedures. Previdi, et al. Expires December 11, 2014 [Page 25] Internet-Draft IPv6 Segment Routing Header (SRH) June 2014 10. IANA Considerations TBD 11. Manageability Considerations TBD 12. Security Considerations Security mechanisms applied to Segment Routing over IPv6 networks are detailed in Section 7. 13. Contributors Eric Vyncke contributed to this document through the writings of Section 7. 14. Acknowledgements The authors would like to thank John Leddy, John Brzozowski, Pierre Francois, Nagendra Kumar, Mark Townsley, Christian Martin, Roberta Maglione and James Connolly for their contribution to this document. 15. References 15.1. Normative References [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997. [RFC2460] Deering, S. and R. Hinden, "Internet Protocol, Version 6 (IPv6) Specification", RFC 2460, December 1998. [RFC5095] Abley, J., Savola, P., and G. Neville-Neil, "Deprecation of Type 0 Routing Headers in IPv6", RFC 5095, December 2007. [RFC6407] Weis, B., Rowles, S., and T. Hardjono, "The Group Domain of Interpretation", RFC 6407, October 2011. 15.2. Informative References Previdi, et al. Expires December 11, 2014 [Page 26] Internet-Draft IPv6 Segment Routing Header (SRH) June 2014 [I-D.filsfils-spring-segment-routing] Filsfils, C., Previdi, S., Bashandy, A., Decraene, B., Litkowski, S., Horneffer, M., Milojevic, I., Shakir, R., Ytti, S., Henderickx, W., Tantsura, J., and E. Crabbe, "Segment Routing Architecture", draft-filsfils-spring- segment-routing-03 (work in progress), June 2014. [I-D.filsfils-spring-segment-routing-mpls] Filsfils, C., Previdi, S., Bashandy, A., Decraene, B., Litkowski, S., Horneffer, M., Milojevic, I., Shakir, R., Ytti, S., Henderickx, W., Tantsura, J., and E. Crabbe, "Segment Routing with MPLS data plane", draft-filsfils- spring-segment-routing-mpls-02 (work in progress), June 2014. [I-D.filsfils-spring-segment-routing-use-cases] Filsfils, C., Francois, P., Previdi, S., Decraene, B., Litkowski, S., Horneffer, M., Milojevic, I., Shakir, R., Ytti, S., Henderickx, W., Tantsura, J., Kini, S., and E. Crabbe, "Segment Routing Use Cases", draft-filsfils- spring-segment-routing-use-cases-00 (work in progress), March 2014. [I-D.ietf-isis-segment-routing-extensions] Previdi, S., Filsfils, C., Bashandy, A., Gredler, H., Litkowski, S., Decraene, B., and J. Tantsura, "IS-IS Extensions for Segment Routing", draft-ietf-isis-segment- routing-extensions-01 (work in progress), June 2014. [I-D.ietf-spring-ipv6-use-cases] Brzozowski, J., Leddy, J., Leung, I., Previdi, S., Townsley, W., Martin, C., Filsfils, C., and R. Maglione, "IPv6 SPRING Use Cases", draft-ietf-spring-ipv6-use- cases-00 (work in progress), May 2014. [I-D.psenak-ospf-segment-routing-ospfv3-extension] Psenak, P., Previdi, S., Filsfils, C., Gredler, H., Shakir, R., and W. Henderickx, "OSPFv3 Extensions for Segment Routing", draft-psenak-ospf-segment-routing- ospfv3-extension-01 (work in progress), February 2014. [RFC1940] Estrin, D., Li, T., Rekhter, Y., Varadhan, K., and D. Zappala, "Source Demand Routing: Packet Format and Forwarding Specification (Version 1)", RFC 1940, May 1996. [RFC4942] Davies, E., Krishnan, S., and P. Savola, "IPv6 Transition/ Co-existence Security Considerations", RFC 4942, September 2007. Previdi, et al. Expires December 11, 2014 [Page 27] Internet-Draft IPv6 Segment Routing Header (SRH) June 2014 Authors' Addresses Stefano Previdi (editor) Cisco Systems, Inc. Via Del Serafico, 200 Rome 00142 Italy Email: sprevidi@cisco.com Clarence Filsfils Cisco Systems, Inc. Brussels BE Email: cfilsfil@cisco.com Brian Field Comcast 4100 East Dry Creek Road Centennial, CO 80122 US Email: Brian_Field@cable.comcast.com Ida Leung Rogers Communications 8200 Dixie Road Brampton, ON L6T 0C1 CA Email: Ida.Leung@rci.rogers.com Previdi, et al. Expires December 11, 2014 [Page 28]