--- 1/draft-ietf-mpls-lsp-ping-02.txt 2006-02-05 00:41:09.000000000 +0100 +++ 2/draft-ietf-mpls-lsp-ping-03.txt 2006-02-05 00:41:09.000000000 +0100 @@ -1,21 +1,21 @@ Network Working Group K. Kompella (Juniper) Internet Draft P. Pan (Ciena) -draft-ietf-mpls-lsp-ping-02.txt N. Sheth (Juniper) +draft-ietf-mpls-lsp-ping-03.txt N. Sheth (Juniper) Category: Standards Track D. Cooper (Global Crossing) -Expires: September 2003 G. Swallow (Cisco) +Expires: December 2003 G. Swallow (Cisco) S. Wadhwa (Juniper) R. Bonica (WorldCom) - March 2003 + June 2003 - Detecting MPLS Data Plane Liveness + Detecting MPLS Data Plane Failures *** DRAFT *** Status of this Memo This document is an Internet-Draft and is in full conformance with all provisions of Section 10 of RFC2026. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that @@ -39,46 +39,41 @@ Abstract This document describes a simple and efficient mechanism that can be used to detect data plane failures in Multi-Protocol Label Switching (MPLS) Label Switched Paths (LSPs). There are two parts to this document: information carried in an MPLS "echo request" and "echo reply" for the purposes of fault detection and isolation; and mechanisms for reliably sending the echo reply. -Sub-IP ID Summary +Changes since last revision (This section to be removed before publication.) - (See Abstract above.) - - RELATED DOCUMENTS - - May be found in the "references" section. - - WHERE DOES IT FIT IN THE PICTURE OF THE SUB-IP WORK - - Fits in the MPLS box. - - WHY IS IT TARGETED AT THIS WG + - Changed title to "Detecting MPLS Data Plane Failures" + - removed section 5 "Reliable Reply Path" + - filled in IANA section + - added new top level TLV for Vendor Enterprise Code + - Clarified Downstream Router ID and Downstream Interface Address + - Clarified receiving procedure + - Example for multipath operation - MPLS WG is currently looking at MPLS-specific error detection and - recovery mechanisms. The mechanisms proposed here are for packet- - based MPLS LSPs, which is why the MPLS WG is targeted. +Issues - JUSTIFICATION + (This section to be removed before publication.) - The WG should consider this document, as it allows network operators - to detect MPLS LSP data plane failures in the network. This type of - failures have occurred, and are a source of concern to operators - implementing MPLS networks. + - Question: use two bits from the TLV space to indicate + - Ignore TLV if not understood + - Reflect TLV in reply + - Tweak error codes? Add stack depth? + - More multipath stuff? 1. Introduction This document describes a simple and efficient mechanism that can be used to detect data plane failures in MPLS LSPs. There are two parts to this document: information carried in an MPLS "echo request" and "echo reply"; and mechanisms for transporting the echo reply. The first part aims at providing enough information to check correct operation of the data plane, as well as a mechanism to verify the data plane against the control plane, and thereby localize faults. @@ -106,41 +101,20 @@ The body of this memo contains four main parts: motivation, MPLS echo request/reply packet format, MPLS ping operation, and a reliable return path. It is suggested that first-time readers skip the actual packet formats and read the Theory of Operation first; the document is structured the way it is to avoid forward references. The last section (reliable return path for RSVP LSPs) may be removed in a future revision. -1.3. Changes since last revision - - (This section to be removed before publication.) - - - Clarified definition of downstream router/interface. - - Added text for multipath (mostly just taken from Curtis) - - Mandated the use of Router Alert for sending echo requests - - If reply mode says IPv4 with router alert, and the reply is - labeled, the top label MUST be the router alert label - - Expanded the Theory of Operation, and added a section on ECMP - - Expanded checks on receipt of echo requests, per email on list - -1.4. Issues remaining - - (This section to be removed before publication.) - - - Monitoring mode - - Finalize ECMP format and semantics - - Keep or remove replies via control plane? - - Normalize error codes - 2. Motivation When an LSP fails to deliver user traffic, the failure cannot always be detected by the MPLS control plane. There is a need to provide a tool that would enable users to detect such traffic "black holes" or misrouting within a reasonable period of time; and a mechanism to isolate faults. In this document, we describe a mechanism that accomplishes these goals. This mechanism is modeled after the ping/traceroute paradigm: @@ -221,31 +195,29 @@ 1 MPLS Echo Request 2 MPLS Echo Reply The Reply Mode can take one of the following values: Value Meaning ----- ------- 1 Do not reply 2 Reply via an IPv4 UDP packet 3 Reply via an IPv4 UDP packet with Router Alert - 4 Reply via the control plane An MPLS echo request with "Do not reply" may be used for one-way connectivity tests; the receiving router may log gaps in the sequence numbers and/or maintain delay/jitter statistics. An MPLS echo request would normally have "Reply via an IPv4 UDP packet"; if the normal IPv4 return path is deemed unreliable, one may use "Reply via an IPv4 UDP packet with Router Alert" (note that this requires that all intermediate routers understand and know how to forward MPLS echo - replies) or "Reply via the control plane" (this is currently only - defined for control plane that uses RSVP). + replies). The Return Code is set to zero by the sender. The receiver can set it to one of the following values: Value Meaning ----- ------- 0 The error code is contained in the Error Code TLV 1 Malformed echo request received 2 One or more of the TLVs was not understood 3 Replying router is an egress for the FEC @@ -289,20 +261,21 @@ Types are defined below; Length is the length of the Value field in octets. The Value field depends on the Type; it is zero padded to align to a four-octet boundary. Type # Value Field ------ ----------- 1 Target FEC Stack 2 Downstream Mapping 3 Pad 4 Error Code + 5 Vendor Enterprise Code 3.1. Target FEC Stack A Target FEC Stack is a list of sub-TLVs. The number of elements is determined by the looking at the sub-TLV length fields. Sub-Type # Length Value Field ---------- ------ ----------- 1 5 LDP IPv4 prefix 2 17 LDP IPv6 prefix @@ -320,31 +293,36 @@ corresponding to the top of the label stack, etc. An MPLS echo request MUST have a Target FEC Stack that describes the FEC stack being tested. For example, if an LSR X has an LDP mapping for 192.168.1.1 (say label 1001), then to verify that label 1001 does indeed reach an egress LSR that announced this prefix via LDP, X can send an MPLS echo request with a FEC Stack TLV with one FEC in it, namely of type LDP IPv4 prefix, with prefix 192.168.1.1/32, and send the echo request with a label of 1001. - If LSR X wanted to verify that a label stack of <1001, 23456> is the - right label stack to use to reach an IP VPN prefix of 10/8 in VPN foo - on an egress LSR with loopback address 192.168.1.1 (learned via LDP), - X has two choices. X can send an MPLS echo request with a FEC Stack - TLV with a single FEC of type VPN IPv4 prefix with a prefix of 10/8 - with the Route Distinguisher for VPN foo. Alternatively, X can send - a FEC Stack TLV with two FECs, the first of type LDP IPv4 with a - prefix of 192.168.1.1/32 and the second of type of IP VPN with a - prefix 10/8 in VPN foo. In either case, the MPLS echo request would - have a label stack of <1001, 23456>. (Note: in this example, 1001 is - the "outer" label and 23456 is the "inner" label.) + Say LSR X wanted to verify that a label stack of <1001, 23456> is the + right label stack to use to reach a VPN IPv4 prefix of 10/8 in VPN + foo. Say further that LSR Y with loopback address 192.168.1.1 + announced prefix 10/8 with Route Distinguisher RD-foo-Y (which may in + general be different from the Route Distinguisher that LSR X uses in + its own advertisements for VPN foo), label 23456 and BGP nexthop + 192.168.1.1. Finally, suppose that LSR X receives a label binding of + 1001 for 192.168.1.1 via LDP. X has two choices in sending an MPLS + echo request: X can send an MPLS echo request with a FEC Stack TLV + with a single FEC of type VPN IPv4 prefix with a prefix of 10/8 and a + Route Distinguisher of RD-foo-Y. Alternatively, X can send a FEC + Stack TLV with two FECs, the first of type LDP IPv4 with a prefix of + 192.168.1.1/32 and the second of type of IP VPN with a prefix 10/8 + with Route Distinguisher of RD-foo-Y. In either case, the MPLS echo + request would have a label stack of <1001, 23456>. (Note: in this + example, 1001 is the "outer" label and 23456 is the "inner" label.) 3.1.1. LDP IPv4 Prefix The value consists of four octets of an IPv4 prefix followed by one octet of prefix length in bits. The IPv4 prefix is in network byte order. See [LDP] for an example of a Mapping for an IPv4 FEC. 3.1.2. LDP IPv6 Prefix The value consists of sixteen octets of an IPv6 prefix followed by @@ -393,38 +371,38 @@ | IPv6 tunnel sender address | | | | | | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Must Be Zero | LSP ID | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3.1.5. VPN IPv4 Prefix - The value field consists of a Route Distinguisher, an IPv4 prefix and - a prefix length, as follows: + The value field consists of the Route Distinguisher advertised with + the VPN IPv4 prefix, the IPv4 prefix and a prefix length, 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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Route Distinguisher | | (8 octets) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | IPv4 prefix | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Prefix Length | Must Be Zero | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3.1.6. VPN IPv6 Prefix - The value field consists of a Route Distinguisher, an IPv6 prefix and - a prefix length, as follows: + The value field consists of the Route Distinguisher advertised with + the VPN IPv6 prefix, the IPv6 prefix and a prefix length, 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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Route Distinguisher | | (8 octets) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | IPv6 prefix | | | | | @@ -462,49 +440,60 @@ | Remote PE Address | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | VC ID | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Encapsulation Type | Must Be Zero | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3.2. Downstream Mapping The Downstream Mapping is an optional TLV in an echo request. The - Length is 12 + 4*N octets, where N is the number of Downstream - Labels. The Value of a Downstream Mapping has the following format: + Length is 16 + 4*M + 4*N octets, where M is the Multipath Length, and + N is the number of Downstream Labels. The Value field of a + Downstream Mapping has the following format: 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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ - | Downstream IPv4 Router ID | + | Downstream IPv4 Address | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | MTU | Address Type | DS Index | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Downstream Interface Address | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ - | Hash Key Type | Depth Limit | Multipath Length | + | Hash Key Type | Depth Limit | No of Multipaths | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | IP Address or Next Label | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ . . . (more IP Addresses or Next Labels) . . . +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Downstream Label | Protocol | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ . . . . . . +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Downstream Label | Protocol | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + If the interface to the downstream LSR is numbered, then the + Downstream IPv4 Address can either be the downstream LSR's Router ID + or the interface address of the downstream LSR. In this case, the + Address Type is set to IPv4 and the Downstream Interface Address is + set to the downstream LSR's interface address. If the interface to + the downstream LSR is unnumbered, the Downstream IPv4 Address MUST be + the downstream LSR's Router ID, and the Address Type MUST be + Unnumbered, and the Downstream Interface Address MUST be the index + assigned by the upstream LSR to the interface. + The MTU is the largest MPLS frame (including label stack) that fits on the interface to the Downstream LSR. The Downstream Interface Address Type is one of: Type # Address Type ------ ------------ 1 IPv4 2 Unnumbered 'Protocol' is taken from the following table: @@ -533,23 +522,22 @@ the incoming label L may be swapped with a label stack.) The case where X is the LSR originating the echo request is a special case. X needs to figure out what LSRs would receive the MPLS echo request for a given FEC Stack that X originates with TTL=1. The set of downstream routers at X may be alternative paths (see the discussion below on ECMP) or simultaneous paths (e.g., for MPLS multicast). In the former case, the Multipath sub-field is used as a hint to the sender as to how it may influence the choice of these - alternatives. The Multipath Length is the total length of the - Multipath field (i.e., 4 + 4*M, where M is the number of IP - Address/Next Label fields). The Hash Key Type is taken from the + alternatives. The "No of Multipaths" is the number of IP + Address/Next Label fields. The Hash Key Type is taken from the following table: Hash Key Type IP Address or Next Label -------------------- ------------------------ 0 no multipath (nothing; M = 0) 1 label M labels 2 IP address M IP addresses 3 label range M/2 low/high label pairs 4 IP address range M/2 low/high address pairs 5 no more labels (nothing; M = 0) @@ -560,60 +548,78 @@ maximum number of labels considered in the hash; this SHOULD be set to zero if unspecified or unlimited. IP Address or Next Label is an IP address from the range 127/8 or an next label which will exercise this particular path. The semantics of the Hash Key Type and IP Address/Next Label are as follows: type 1 - a list of single labels is provided, any one of which - will - cause the hash to match this MP path. + will cause the hash to match this MP path. type 2 - a list of single IP addresses is provided, any one of which will cause the hash to match this MP path. type 3 - a list of label ranges is provided, any one of which will cause the hash to match this MP path. type 4 - a list of IP address ranges is provided, any one of which will cause the hash to match this MP path. type 5 - if no more labels are provided on the stack, this MP path will apply (can only appear once). - type 6 - Any IP addresses matches. Undertlying labels may go + type 6 - Any IP addresses matches. Underlying labels may go elsewhere, but all IP takes only one MP path (can only appear once). type 7 - no matches are possible given the set of "Multipath Exercise TLV" provided by prior hops. If prior hops provide a "Downstream Multipath Mapping TLV" the labels and IP addresses should be picked from the set provided in prior "Multipath Exercise TLV" or "Hash Key Type" of 7 used. + For example, suppose LSR X at hop 10 has two downstream LSRs Y and Z + for the FEC in question. X could return Hash Key Type 4, with + low/high IP addresses of 1.1.1.1->1.1.1.255 for downstream LSR Y and + 2.1.1.1->2.1.1.255 for downstream LSR Z. The head end reflects this + information to LSR Y. Y, which has three downstream LSRs U, V and W, + computes that 1.1.1.1->1.1.1.127 would go to U and 1.1.1.128-> + 1.1.1.255 would go to V. Y would then respond with 3 Downstream + Mappings: to U, with Hash Key Type 4 (1.1.1.1->1.1.1.127); to V, with + Hash Key Type 4 (1.1.1.127->1.1.1.255); and to W, with Hash Key Type + 7. + 3.3. Pad TLV The value part of the Pad TLV contains a variable number (>= 1) of octets. The first octet takes values from the following table; all the other octets (if any) are ignored. The receiver SHOULD verify that the TLV is received in its entirety, but otherwise ignores the contents of this TLV, apart from the first octet. Value Meaning ----- ------- 1 Drop Pad TLV from reply 2 Copy Pad TLV to reply 3-255 Reserved for future use 3.4. Error Code The Error Code TLV is currently not defined; its purpose is to provide a mechanism for a more elaborate error reporting structure, should the reason arise. +3.5. Vendor Enterprise Code + + The Length is always 4; the value is the SMI Enterprise code, in + network octet order, of the vendor with a Vendor Private extension to + any of the fields in the fixed part of the message, in which case + this TLV MUST be present. If none of the fields in the fixed part of + the message have vendor private extensions, this TLV is OPTIONAL. + 4. Theory of Operation An MPLS echo request is used to test a particular LSP. The LSP to be tested is identified by the "FEC Stack"; for example, if the LSP was set up via LDP, and is to an egress IP address of 10.1.1.1, the FEC stack contains a single element, namely, an LDP IPv4 prefix sub-TLV with value 10.1.1.1/32. If the LSP being tested is an RSVP LSP, the FEC stack consists of a single element that captures the RSVP Session and Sender Template which uniquely identifies the LSP. @@ -667,22 +673,30 @@ possible paths. However, full coverage may not be possible. 4.2. Sending an MPLS Echo Request An MPLS echo request is a (possibly) labelled UDP packet. The IP header is set as follows: the source IP address is a routable address of the sender; the destination IP address is a (randomly chosen) address from 127/8; the IP TTL is set to 1. The source UDP port is chosen by the sender; the destination UDP port is set to 3503 (assigned by IANA for MPLS echo requests). The Router Alert option - is set in the IP header. If the echo request is labelled, the MPLS - TTL on all the labels except the outermost should be set to 1. + is set in the IP header. + + If the echo request is labelled, one may (depending on what is being + pinged) set the TTL of the innermost label to 1, to prevent the ping + request going farther than it should. Examples of this include + pinging a VPN IPv4 or IPv6 prefix, an L2 VPN end point or an L2 + circuit ID. This can also be accomplished by inserting a router + alert label above this label; however, this may lead to the undesired + side effect that MPLS echo requests take a different data path than + actual data. In "ping" mode (end-to-end connectivity check), the TTL in the outermost label is set to 255. In "traceroute" mode (fault isolation mode), the TTL is set successively to 1, 2, .... The sender chooses a Sender's Handle, and a Sequence Number. When sending subsequent MPLS echo requests, the sender SHOULD increment the sequence number by 1. However, a sender MAY choose to send a group of echo requests with the same sequence number to improve the chance of arrival of at least one packet with that sequence number. @@ -710,29 +724,32 @@ An LSR X that receives an MPLS echo request first parses the packet to ensure that it is a well-formed packet, and that the TLVs are understood. If not, X SHOULD send an MPLS echo reply with the Return Code set to "Malformed echo request received" or "TLV not understood" (as appropriate), and the Subcode set to the appropriate value. If the echo request is good, X then checks whether it is a valid transit or egress LSR for the FEC in the echo request. If not, X MAY log this fact. If it is, X notes that interface I over which the echo was received, and the label L with which it came. X checks - whether it actually advertised L over interface I for the FEC in the - echo request. + whether it actually advertised L for the FEC in the echo request; X + MAY further check whether it expects L over interface I or not. If the echo request contains a Downstream Mapping TLV, X MUST further - check whether its Router ID matches one of the Downstream IPv4 Router - IDs; and if so, whether the given Downstream Label is in fact the - label that X sent as its mapping for the FEC over the downstream - interface. The result of the checks in the previous and this - paragraph are captured in the Return Code/Subcode. + check whether its Router ID or one of its interface addresses matches + one of the Downstream IPv4 Address; if the Address Type is + Unnumbered, X further checks if the interface I has the given + (upstream) index. If these check out, X determines whether the given + Downstream Label is in fact the label that X sent as its mapping for + the FEC over the downstream interface. The result of the checks in + the previous and this paragraph are captured in the Return + Code/Subcode. If the echo request has a Reply Mode that wants a reply, X uses the procedure in the next subsection to send the echo reply. 4.4. Sending an MPLS Echo Reply An MPLS echo reply is a UDP packet. It MUST ONLY be sent in response to an MPLS echo request. The source IP address is a routable address of the replier; the source port is the well-known UDP port for MPLS ping. The destination IP address and UDP port are copied from the @@ -786,148 +803,46 @@ ping, then no reply will be sent, resulting in possible "false negatives". If in "traceroute" mode, a transit LSR does not support MPLS ping, then no reply will be forthcoming from that LSR for some TTL, say n. The LSR originating the echo request SHOULD try sending the echo request with TTL=n+1, n+2, ..., n+k in the hope that some transit LSR further downstream may support MPLS echo requests and reply. In such a case, the echo request for TTL>n MUST NOT have Downstream Mapping TLVs, until a reply is received with a Downstream Mapping. -5. Reliable Reply Path - - One of the issues that are faced with MPLS ping is to distinguish - between a failure in the forward path (the MPLS path being 'pinged') - and a failure in the return path. Note that this problem exists with - vanilla IP ping as well. In the case of MPLS ping, it is assumed - that the IP control and data planes are reliable. However, it could - be that the forwarding in the return path is via an MPLS LSP. - - In this specification, we give two solutions for this problem. One - is to set the Router Alert option in the MPLS echo reply. When a - router sees this option, it MUST forward the packet as an IP packet. - Note that this may not work if some transit LSR does not support MPLS - ping. - - Another option is to send the echo reply via the control plane. At - present, this is defined only for RSVP-TE LSPs, and described below. - - These options are controlled by the ingress LSR, using the Reply Mode - in the MPLS echo request packet. - -5.1. RSVP-TE Extension - - To test an LSP's liveliness, an ingress LSR sends MPLS echo requests - over the LSP being tested. When an egress LSR receives the message, - it needs to acknowledge the ingress LSR by sending an LSP_ECHO object - in a RSVP Resv message. The object has the following format: - - Class = LSP_ECHO (use form 11bbbbbb for compatibility) - - C-Type = 1 - - 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 - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ - | Sequence Number | - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ - | TimeStamp (seconds) | - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ - | TimeStamp (microseconds) | - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ - | UDP Source Port | Return Code | Return Subcode| - +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ - - The Sequence Number is copied from the Sequence Number of the echo - request. The TimeStamp is set to the time the echo request is - received. The UDP Source Port is copied from the UDP source port of - the MPLS echo request. The FEC is implied by the Session and the - Sender Template Objects. - -5.2. Operation - - For the sake of brevity in the context of this document by "the - control plane" we mean "the RSVP-TE component of the control plane". - - Consider an LSP between an ingress LSR and an egress LSR spanning - multiple LSR hops. - -5.3. Procedures at the ingress LSR - - One must ensure before setting the Reply Mode to "reply via the - control plane" that the egress LSR supports this feature. - - The ingress LSR, say X, builds an MPLS echo request as in section - "Sending an MPLS Echo Request". The FEC Type must be an RVSP Session - Query. X also sets the Reply Mode to "reply via the control plane". - - If X does not receive an Resv message from the egress LSR that - contains an LSP_ECHO object within some period of time, it declares - the LSP as "down". At this point, the ingress LSR may apply the - necessary procedures to fix the LSP. These may include generating a - message to network management, tearing-down and re-building the LSP, - and/or rerouting user traffic to a backup LSP. - - To test an LSP that carries non-IP traffic, before injecting ICMP and - MPLS ping messages into the LSP, the IPv4 Explicit NULL label should - be prepended to such messages. The ingress and egress LSR's must - follow the procedures defined in [LABEL-STACK]. - -5.4. Procedures at the egress LSR - - When the egress LSR receives an MPLS ping message, it follows the - procedures given above. If the Reply Mode is set to "Reply via the - control plane", the LSR can, based on the RSVP SESSION and - SENDER_TEMPLATE objects carried in the MPLS ping message, find the - corresponding LSP in its RSVP-TE database. The LSR then checks to - see if the Resv message for this LSP contains an LSP_ECHO object with - the same source UDP port value. If not, the LSR adds or updates the - LSP_ECHO object and refreshes the Resv message. - -5.5. Procedures for the intermediate LSR's - - At intermediate LSRs, normal RSVP processing procedures will cause - the LSP_ECHO object to be forwarded as RSVP messages are refreshed. - - At the LSR's that support MPLS ping the Resv messages that carry the - LSP_ECHO object MUST be delivered upstream immediately. - - Note that an intermediate LSR using RSVP refresh reduction [RSVP- - REFRESH], the new or changed LSP_ECHO object will cause the LSR to - classify the RSVP message as a trigger message. - -6. Normative References +Normative References [KEYWORDS] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997. [LABEL-STACK] Rosen, E., et al, "MPLS Label Stack Encoding", RFC 3032, January 2001. [RSVP] Braden, R. (Editor), et al, "Resource ReSerVation protocol (RSVP) -- Version 1 Functional Specification," RFC 2205, September 1997. [RSVP-REFRESH] Berger, L., et al, "RSVP Refresh Overhead Reduction Extensions", RFC 2961, April 2001. [RSVP-TE] Awduche, D., et al, "RSVP-TE: Extensions to RSVP for LSP tunnels", RFC 3209, December 2001. -7. Informative References +Informative References [ICMP] Postel, J., "Internet Control Message Protocol", RFC 792. [LDP] Andersson, L., et al, "LDP Specification", RFC 3036, January 2001. -8. Security Considerations +Security Considerations There are at least two approaches to attacking LSRs using the mechanisms defined here. One is a Denial of Service attack, by sending MPLS echo requests/replies to LSRs and thereby increasing their workload. The other is obfuscating the state of the MPLS data plane liveness by spoofing, hijacking, replaying or otherwise tampering with MPLS echo requests and replies. Authentication will help reduce the number of seemingly valid MPLS echo requests, and thus cut down the Denial of Service attacks; @@ -937,39 +852,84 @@ tampering attacks; one hopes to use some mechanism devised or suggested by the RPSec WG. It is not clear how to prevent hijacking (non-delivery) of echo requests or replies; however, if these messages are indeed hijacked, MPLS ping will report that the data plane isn't working as it should. It doesn't seem vital (at this point) to secure the data carried in MPLS echo requests and replies, although knowledge of the state of the MPLS data plane may be considered confidential by some. -9. IANA Considerations +5. IANA Considerations - (To be filled in a later revision) + The TCP and UDP port number 3503 has been allocated by IANA for LSP + echo requests and replies. -10. Acknowledgments + The following sections detail the new name spaces to be managed by + IANA. For each of these name spaces, the space is divided into + assignment ranges; the following terms are used in describing the + procedures by which IANA allocates values: "Standards Action" (as + defined in [IANA]); "Expert Review" and "Vendor Private Use". + + Values from "Expert Review" ranges MUST be registered with IANA, and + MUST be accompanied by an Experimental RFC that describes the format + and procedures for using the code point. + + Values from "Vendor Private" ranges MUST NOT be registered with IANA; + however, the message MUST contain an enterprise code as registered + with the IANA SMI Network Management Private Enterprise Codes. For + each name space that has a Vendor Private range, it must be specified + where exactly the SMI Enterprise Code resides; see below for + examples. In this way, several enterprises (vendors) can use the + same code point without fear of collision. + +5.1. Message Types, Reply Modes, Return Codes + + It is requested that IANA maintain registries for Message Types, + Reply Modes, Return Codes and Return Subcodes. Each of these can + take values in the range 0-255. Assignments in the range 0-191 are + via Standards Action; assignments in the range 192-251 are made via + Expert Review; values in the range 252-255 are for Vendor Private + Use, and MUST NOT be allocated. + + If any of these fields fall in the Vendor Private range, a top-level + Vendor Enterprise Code TLV MUST be present in the message. + +5.2. TLVs + + It is requested that IANA maintain registries for the Type field of + top-level TLVs as well as for sub-TLVs. The valid range for each of + these is 0-65535. Assignments in the range 0-32767 are made via + Standards Action; assignments in the range 32768-64511 are made via + Expert Review; values in the range 64512-65535 are for Vendor Private + Use, and MUST NOT be allocated. + + If a TLV or sub-TLV has a Type that falls in the range for Vendor + Private Use, the Length MUST be at least 4, and the first four octets + MUST be that vendor's SMI Enterprise Code, in network octet order. + The rest of the Value field is private to the vendor. + +Acknowledgments This document is the outcome of many discussions among many people, that include Manoj Leelanivas, Paul Traina, Yakov Rekhter, Der-Hwa Gan, Brook Bailey, Eric Rosen and Ina Minei. The Multipath Exercise sub-field of the Downstream Mapping TLV was adapted from text suggested by Curtis Villamizar. -11. Appendix +Appendix This appendix specifies non-normative aspects of detecting MPLS data plane liveness. -11.1. CR-LDP FEC +5.1. CR-LDP FEC This section describes how a CR-LDP FEC can be included in an Echo Request using the following FEC subtype: Sub-Type # Length Value Field ---------- ------ ----------- 5 6 CR-LDP LSP ID The value consists of the LSPID of the LSP being pinged. An LSPID is a four octet IPv4 address (a local address on the ingress LSR, for @@ -977,30 +937,30 @@ per LSP on a given ingress LSR. 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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Ingress LSR Router ID | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Must Be Zero | LSP ID | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ -11.2. Downstream Mapping for CR-LDP +5.2. Downstream Mapping for CR-LDP If a label in a Downstream Mapping was learned via CR-LDP, the Protocol field in the Mapping TLV can use the following entry: Protocol # Signaling Protocol ---------- ------------------ 5 CR-LDP -12. Authors' Addresses +Authors' Addresses Kireeti Kompella Nischal Sheth Juniper Networks 1194 N.Mathilda Ave Sunnyvale, CA 94089 e-mail: kireeti@juniper.net e-mail: nsheth@juniper.net Ping Pan @@ -1031,21 +991,21 @@ email: swadhwa@unispherenetworks.com phone: +1 978.589.0697 Ronald P. Bonica WorldCom 22001 Loudoun County Pkwy Ashburn, Virginia, 20147 email: ronald.p.bonica@wcom.com phone: +1 703.886.1681 -13. Intellectual Property Rights Notices +Intellectual Property Rights Notices The IETF takes no position regarding the validity or scope of any intellectual property or other rights that might be claimed to pertain to the implementation or use of the technology described in this document or the extent to which any license under such rights might or might not be available; neither does it represent that it has made any effort to identify any such rights. Information on the IETF's procedures with respect to rights in standards-track and standards-related documentation can be found in BCP-11. Copies of claims of rights made available for publication and any assurances of