INTERNET-DRAFT Richard G. Ogier IETF MANET Working Group Fred L. Templin 28 November 2001 Bhargav Bellur Mark G. Lewis SRI International Topology Broadcast Based on Reverse-Path Forwarding (TBRPF) Status of This Memo This document is an Internet-Draft and is in full conformance with all provisions of Section 10 of RFC 2026 except that the right to produce derivative works is not granted. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet- Drafts. Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress." The list of current Internet-Drafts can be accessed at http://www.ietf.org/ietf/1id-abstracts.txt The list of Internet-Draft Shadow Directories can be accessed at http://www.ietf.org/shadow.html. Abstract TBRPF is a proactive, link-state routing protocol designed for mobile ad-hoc networks, which provides hop-by-hop routing along minimum-hop paths to each destination. Each node running TBRPF computes a source tree (providing paths to all reachable nodes) based on partial topology information stored in its topology table, using a modification of Dijkstra's algorithm. To minimize overhead, each node reports only *part* of its source tree to neighbors. This is in contrast to other protocols (e.g., STAR) in which each node reports its *entire* source tree to neighbors. TBRPF uses a combination of periodic and differential updates to keep all neighbors informed of the reportable part of its source tree. Each node also has the option to report additional topology information (up to the full topology), to provide improved robustness in highly mobile networks. TBRPF performs neighbor discovery using "differential" HELLO messages which report only *changes* in the status of neighbors. This results in HELLO messages that are much smaller than those of other link- state routing protocols such as OSPF and OLSR. Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 1] INTERNET-DRAFT TBRPF 28 November 2001 Contents 1. Introduction ................................................... 1 2. Changes ........................................................ 1 3. TBRPF Terminology .............................................. 2 4. Applicability Section .......................................... 3 4.1. Networking Context ........................................ 4 4.2. Protocol Characteristics and Mechanisms ................... 4 5. TBRPF Overview ................................................. 6 5.1. Overview of Neighbor Discovery ............................ 6 5.2. Overview of Topology Discovery and Route Computation ...... 7 6. TBRPF Packets .................................................. 9 6.1. TBRPF Packet Header ....................................... 9 6.2. TBRPF Packet Body ......................................... 11 7. TBRPF Neighbor Discovery ....................................... 13 7.1. HELLO Message Format ...................................... 13 7.2. Neighbor Table ............................................ 14 7.3. Sending HELLO Messages .................................... 15 7.4. Processing a Received HELLO Message ....................... 15 7.5. Expiration of Timer nbr_life .............................. 16 7.6. Link-Layer Failure Indication ............................. 16 7.7. Configurable Parameters ................................... 17 8. Topology Discovery and Route Computation ....................... 17 8.1. Conceptual Data Structures and Messages ................... 17 8.2. TOPOLOGY UPDATE Messages .................................. 19 8.3. TBRPF Operation ........................................... 19 8.4. Pseudocode ................................................ 21 8.5. Configurable Parameters ................................... 26 9. Application of TBRPF In Mobile Ad-hoc Networks (MANETs) ........ 26 9.1. Data Link Layer Assumptions ............................... 26 9.2. Network Layer Assumptions ................................. 27 9.3. Optional Automatic Address Resolution ..................... 27 9.4. Support for Multiple Interfaces and/or Alias Addresses .... 27 9.5. Support for Hierarchical Network Prefixes ................. 28 9.6. Support for non-MANET Hosts ............................... 28 9.7. Internet Protocol Considerations .......................... 28 9.7.1. IPv4 Operation ..................................... 28 9.7.2. IPv6 Operation ..................................... 29 10. IANA Considerations ........................................... 29 11. Implementation Status ......................................... 29 12. Patent Rights Statement ....................................... 30 13. Acknowledgments ............................................... 30 14. References .................................................... 30 Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page ii] INTERNET-DRAFT TBRPF 28 November 2001 1. Introduction TBRPF is a proactive, link-state routing protocol that provides hop- by-hop routing along minimum-hop paths to each destination. Each node running TBRPF computes a source tree (providing minimum-hop paths to all reachable nodes) based on partial topology information stored in its topology table, using a modification of Dijkstra's algorithm. To minimize overhead, each node reports only *part* of its source tree to neighbors. This is in contrast to other protocols (e.g., FTSP [7] and STAR [4]) in which each node reports its *entire* source tree to neighbors. TBRPF uses a combination of periodic and differential updates to keep all neighbors informed of the reportable part of its source tree. Each node also has the option to report addition topology information (up to the full topology), to provide improved robustness in highly mobile networks. TBRPF performs neighbor discovery using "differential" HELLO messages which report only *changes* in the status of neighbors. This results in HELLO messages that are much smaller than those of other link- state routing protocols such as OSPF and OLSR. TBRPF consists of two modules: the neighbor discovery module and the topology discovery and route computation module. An overview of these modules is given in Section 5. 2. Changes Major changes from version 02 to version 03: - TBRPF no longer uses ACKs or NACKs to provide reliable delivery of messages. Instead, it uses a combination of periodic and differen- tial topology updates. - NEW PARENT messages are no longer used for the selection of parents. Instead, parents select themselves, thus avoiding the overhead of NEW PARENT messages. - The full topology (FT) mode of TBRPF has been merged with the par- tial topology (PT) mode. Each node is required to report the minimum topology information (as in PT), but has the option of reporting additional topology information. Major changes from version 01 to version 02: - The neighbor discovery protocol has been modified so that it is simpler and is independent of the routing protocol. Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 1] INTERNET-DRAFT TBRPF 28 November 2001 - Sections 9 and 10 have been shortened considerably. - The IANA considerations section has been revised to bring it up to date with recent discussions on the MANET email list. Major changes from version 00 to version 01: - A partial-topology mode (TBRPF-PT) has been introduced to achieve a further reduction in control traffic in large, dense networks. - The neighbor discovery protocol has been improved significantly. - The unicast mode for transmitting TBRPF messages has been elim- inated: each TBRPF packet is now broadcast to all neighbors. If a message is to be processed by only a few neighbor nodes, their identities are included in the message. - All topology updates are now sent reliably to all neighbors using NACKs. - Two new configurable parameters, MIN_UPDATE_INTERVAL and MIN_FORW_UPDATE_INTERVAL, have been introduced to minimize fre- quent generation and forwarding of topology updates. - TBRPF-FT includes a new message type NEW PARENT REPLY, which is sent in response to one or more NEW PARENT messages. A single message is sent in response to multiple NEW PARENT messages that are received at about the same time. - A new packet formatting scheme is used that provides optimal pack- ing of TBRPF protocol elements with minimal insertion of header and null padding octets. A packet compression mechanism is included that eliminates all null padding octets when enabled. - Formats have been added for new TBRPF-FT and TBRPF-PT message types. Formats have been revised for some existing TBRPF-FT mes- sage types. - An implicit NARP mechanism has been provided for binding one or more IP addresses to a Router ID. - Optional checksum facilities have been provided for data integrity. 3. TBRPF Terminology The keywords "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC2119 [2]. The Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 2] INTERNET-DRAFT TBRPF 28 November 2001 following terms are also used to describe TBRPF: node A router that implements TBRPF, identified by a unique router ID (RID), also called a node ID. link A logical connection from one node to another, identified by a pair (u,v), where u and v represent nodes. Nodes u and v are called the "head" and "tail" of the link, respectively. A link is said to be bidirectional or 2-way if each node can receive messages from the other. A bidirectional link need not use the same interfaces or media in both directions. neighbor A node v is said to be a neighbor of node u if there is a bidirectional link (u,v) between them. topology The topology of the network is described by a graph G = (V, E), where V is the set of nodes u and E is the set of links (u,v) in the network. directed tree A subset of (directed) links (u,v) that does not contain any loops. The root of a directed tree is the only node u such that the tree contains no link whose tail is u. source tree The directed tree computed by each node that provides shortest paths to all other nodes. Not the same as a broadcast tree. topology update A message or part of a message that reports a state change for one or more links. update source The node that originated a given topology update. parent The parent of a node i for an update source u is the next node on the computed minimum-hop path to node u. 4. Applicability Section This section describes the networking context and protocol charac- teristics of the TBRPF protocol as specified in the MANET Routing Protocol Applicability Statement [16]. Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 3] INTERNET-DRAFT TBRPF 28 November 2001 4.1. Networking Context TBRPF is appropriate for any large MANET network, especially when nodes are highly mobile and bandwidth is limited. 4.2. Protocol Characteristics and Mechanisms * Does the protocol provide shortest path routes? Yes. * Does the protocol provide support for unidirectional links? (if so, how?) No. It uses only bidirectional links (as in 802.11). * Does the protocol require the use of tunneling? (if so, how?) No. * Does the protocol require using some form of source routing? (if so, how?) No. The protocol uses hop-by-hop routing. However, since the entire path is known, source routing can be used as an option. * Does the protocol require the use of periodic messaging? (if so, how?) Each node transmits HELLO and topology update messages periodi- cally. * Does the protocol require the use of reliable or sequenced packet delivery? (if so, how?) No. * Does the protocol provide support for routing through a multi- technology routing fabric? (if so, how?) Yes. Each network interface is assigned a unique IP address. * Does the protocol provide support for multiple hosts per router? (if so, how?) Yes. The concept of a Router ID (RID) [3,14,17] can be used to associate multiple hosts with a single RID. Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 4] INTERNET-DRAFT TBRPF 28 November 2001 * Does the protocol support the IP addressing architecture? (if so, how?) Yes. * Does the protocol require link or neighbor status sensing (if so, how?) Yes. A new, efficient neighbor discovery protocol is presented, in which each HELLO message contains only the IDs of nodes that have recently been heard but with which a 2-way link has not yet been established. * Does the protocol depend on a central entity? (if so, how?) No. * Does the protocol function reactively? (if so, how?) The protocol reacts to topology changes, but not to traffic demand. * Does the protocol function proactively? (if so, how?) Yes. It maintains optimal paths to all reachable destinations at all times. * Does the protocol provide loop-free routing? (if so, how?) Yes. Since routing is based on link states, it provides loop-free routing when the topology is stable. * Does the protocol provide for sleep period operation? (if so, how?) TBRPF operates correctly even if nodes can go into and out of sleep mode at arbitrary times. Other features can be added, such as assigning awake routers to store packets for sleeping nodes. * Does the protocol provide some form of security? (if so, how?) No. Other other protocols (such as IMEP [14]) can be used to pro- vide security. * Does the protocol provide support for utilizing multi-channel, link-layer technologies? (if so, how?) Yes. A network using multiple radio channels can be represented by a single logical topology in which any link can use any channel or combination of channels. Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 5] INTERNET-DRAFT TBRPF 28 November 2001 5. TBRPF Overview TBRPF consists of two main modules: the neighbor discovery module, and the topology discovery and route computation module. 5.1. Overview of Neighbor Discovery The TBRPF Neighbor Discovery (TND) protocol allows each node to quickly detect the neighboring nodes with which the node has a direct, bidirectional link, i.e., a link such that the node at each end of the link can hear the other node. The protocol also detects when a bidirectional link to some neighbor no longer exists. The key feature of TND is that it uses "differential" HELLO messages which report only *changes* in the status of neighbors. This results in HELLO messages that are much smaller than those of other link- state routing protocols such as OSPF and OLSR, in which each HELLO message includes the IDs of *all* neighbors. As a result, HELLO mes- sages can be sent more frequently in highly mobile networks without increasing overhead significantly. TND is designed to be fully modular and independent of the routing protocol. As a result, TND can be used by other routing protocols, and TBRPF can use another neighbor discovery protocols in place of TND, e.g., one provided by the link layer. Each node sends a HELLO message every HELLO_INTERVAL seconds, with a small jitter. Each HELLO message contains three (possibly empty) lists of router IDs, formatted as the following three message sub- types: NEIGHBOR REQUEST, NEIGHBOR REPLY, and NEIGHBOR LOST. A NEIGH- BOR REQUEST message is always included in a HELLO, even if its list of router IDs is empty. However, a NEIGHBOR REPLY or NEIGHBOR LOST message is included only if its list of router IDs is nonempty. Each HELLO message also contains the current HELLO sequence number (HSEQ), which is incremented with each transmitted HELLO. For convenience, we say that "node i sends a NEIGHBOR REQUEST message for node j" if node i sends such a message that includes the ID of node j, and simi- larly for NEIGHBOR REPLY and NEIGHBOR LOST messages. Each node maintains a neighbor table, which stores the state informa- tion for other nodes. The state (or level) of each node can be 1- WAY, 2-WAY, or LOST. When node i changes the state of a neighbor j, it sends the appropriate message (NEIGHBOR REQUEST/UP/LOST) in at most NBR_HOLD_COUNT (typically 3) consecutive HELLOs. This ensures that node j will either receive the message, or will miss NBR_HOLD_COUNT HELLOs and thus declare node i to be LOST. This tech- nique also makes it unnecessary for a node to include each 1-WAY neighbor in HELLOs indefinitely, unlike OSPF and OLSR. The NEIGHBOR REQUEST message includes a list of neighbors from which Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 6] INTERNET-DRAFT TBRPF 28 November 2001 HELLO messages have recently been heard but for which a 2-WAY link is not currently established. To avoid establishing a link that is likely to be short lived (i.e., to employ hysteresis), a node i must receive at least HELLO_ACQUIRE_COUNT (typically 2) of the last HELLO_ACQUIRE_WINDOW (typically 3) HELLOs from another node j before declaring the node to be 1-WAY. In this case, node i sends a NEIGH- BOR REQUEST message for node j in each of its next NBR_HOLD_COUNT HELLO messages, or until a NEIGHBOR REPLY message for node i is received from node j. If node j receives a NEIGHBOR REQUEST from node i, then node j declares the link to node i to be 2-WAY (if it is not already 2-WAY), and includes a NEIGHBOR REPLY message for node i in its next NBR_HOLD_COUNT transmitted HELLO messages. Upon receiving the NEIGH- BOR REPLY message, node i declares the link to node j to be 2-WAY. If node i receives a HELLO from a 1-WAY or 2-WAY neighbor j whose HSEQ indicates that at least NBR_HOLD_COUNT HELLOs were missed, or if node i receives no HELLO from node j within NBR_HOLD_TIME seconds, then node i changes the state of node j to LOST, and sends a NEIGHBOR LOST message for node i in its next NBR_HOLD_COUNT transmitted HELLO messages (unless the link changes state before these transmissions are complete). Node j will either receive the message or will miss NBR_HOLD_COUNT HELLOs; in either case it will declare node i to be LOST. In this manner, both nodes will agree that the link is no longer bidirectional, even if node j can still hear HELLOs from node i. 5.2. Overview of Topology Discovery and Route Computation TBRPF is a proactive, link-state routing protocol that provides hop- by-hop routing along minimum-hop paths to each destination. Each node running TBRPF computes a source tree (providing paths to all reach- able nodes) based on partial topology information stored in its topology table, using a modification of Dijkstra's algorithm. To minimize overhead, each node reports only part of its source tree to neighbors. This is in contrast to FTSP [7] and STAR [4], in which each node reports its *entire* source tree to neighbors, which is redundant since the source trees of different neighbors often overlap considerably. The main idea behind the current version of TBRPF came from PTSP [7], another protocol in which each node reports only part of its source tree. However, PTSP differs in many ways from TBRPF: e.g., PTSP uses NEW PARENT messages and reliable updates, unlike the current version of TBRPF (but like previous versions of TBRPF [1]). To decide which part of its source tree T to report to neighbors, a node running TBRPF first computes its "reportable node set" RN. Roughly speaking, node i includes node u in RN if it estimates that the minimum-hop path from some neighbor j (of node i) to destination Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 7] INTERNET-DRAFT TBRPF 28 November 2001 u goes through node i. In other words, node i includes node u in RN if it estimates that it is the next hop (or parent) of some neighbor to destination u. This technique avoids the need for the NEW PARENT messages that were used in PTSP and previous versions of TBRPF. The part of T that a node reports to neighbors is called the "report- able subtree" and is denoted RT. RT consists of links (u,v) of T such that u is in RN. Each node reports RT to neighbors in *periodic* topology updates (e.g., every 5 seconds), and reports changes (additions and deletions) to RT in more frequent *differen- tial* updates (e.g., every 1 second). Periodic updates inform new neighbors of RT, and ensure that each neighbor eventually learns RT even if it does not receive all updates. Differential updates ensure the fast propagation of each topology update to all nodes that are affected by the update. A received topology update is not forwarded, but *may* result in a change to RT, which will be reported in the next differential update. Whenever possible, topology updates are included in the same packet as a HELLO message, to minimize the number of control packets sent. TBRPF does not require reliable or sequenced delivery of messages, and does not use ACKs or NACKs. TBRPF does not use sequence numbers for topology updates, thus reduc- ing message overhead and avoiding wraparound problems. Instead, a technique similar to SPTA [20] is used in which, for each link (u,v) reported by one or more neighbors, only the parent p(u) for u is believed regarding the the state of the link. (However, in SPTA each node reports the full topology.) Using this technique, each node maintains a topology graph TG, consisting of "believable" links that are reported by neighbors, and computes T as the shortest-path tree within TG. To allow immediate rerouting, the restriction that each link (u,v) in TG must be reported by p(u) is relaxed temporarily if p(u) changes to a neighbor that is not reporting the link. Each node is required to report RT, but MAY report additional links, e.g., to provide increased robustness in highly mobile networks. More precisely, a node may maintain any subgraph H of TG that con- tains T, and report the reportable subgraph RH, which consists of links (u,v) of H such that u is in RN. For example, H can equal TG, which would provide each node with the full network topology if this is done by all nodes. H can also be a biconnected subgraph that con- tains T, which would provide each node with two disjoint paths to each other node, if this is done by all nodes. In a MANET, some nodes, called non-relay nodes, may choose not to forward data packets for other nodes. Non-relay nodes are supported by TBRPF in a very simple manner: they run TBRPF (and transmit HEL- LOs) but do not do not transmit topology update messages. As a result, no node will compute a path that uses a non-relay node as an intermediate node. Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 8] INTERNET-DRAFT TBRPF 28 November 2001 6. TBRPF Packets Nodes send TBRPF protocol data in *packets*. Each packet includes a packet header, optional header extensions, and a packet body that comprises one or more *message(s)* and padding options as needed. The total length of all packet components MUST be less than the max- imum length represented by an unsigned 16-bit integer, i.e. 64KB. 6.1. TBRPF Packet Header TBRPF packet headers are variable-length (minimum two octets) as determined by "flag" bits found in the first octet. Implementations MUST ensure that the first octet of the TBRPF packet header is aligned on an even 8-octet boundary in order to guarantee alignment of n-octet data elements up to 8-octets in length. All further align- ment requirements discussed throughout this section are relative to the first octet of the TBRPF packet header. Therefore, bit ordering beginning at bit 0 is shown ONLY in the TBRPF packet header diagram; Other diagrams throughout this document omit the bit ordering nota- tion. The format for the TBRPF packet header and a description of the individual fields are given below: 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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+- - - - - - - - - - - - - - - - - - |C|L|R|Z| Vers | Reserved | Optional Extension Fields +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+- - - - - - - - - - - - - - - - - - Flags (4 bits): Three flag bits (C, L, R) specify which header extension fields (if any) are present. Any header extension fields enabled by these bits MUST appear in canonical order (i.e. first C, then L, then R). A fourth flag bit (Z) specifies either aligned or compressed format for the TBRPF message. The flag bits and their usage are described in detail below: C - Checksum Field Included: If the underlying delivery service provides a checksum facility the sender MAY set C = '0' and omit the checksum extension field. Otherwise, the sender MUST set C = '1' and include a 16-bit checksum field beginning in the first octet of the header extension. When C = '1', the sender MUST calculate the checksum of the packet in the manner described in [11] and write the result into the checksum field. The checksum is cal- culated across ALL data bytes in the packet header and body, but DOES NOT include a pseudo-header of the underlying delivery service. Therefore, if other header extension fields are also included (see below), the sender MUST fill in such additional fields BEFORE the checksum is calculated. Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 9] INTERNET-DRAFT TBRPF 28 November 2001 Receivers MUST examine the C bit to determine whether the checksum field is present. If C = '1', a receiver MUST verify the checksum encoded in the 16-bit message checksum field as described in [11] WHETHER OR NOT a checksum was also performed by the underlying delivery service. Receivers MUST discard any TBRPF packet that contains an incorrect checksum as well as any packet with neither a checksum provided by the underlying delivery service nor a checksum in the packet header. L - Length Field Included: If the underlying delivery service provides a length field, the sender MAY set L = '0' and omit the length field. Otherwise, the sender MUST set L = '1' and include a 16-bit unsigned integer length field immediately after any previous header fields. When L = '1', the sender MUST calculate the length of the TBRPF packet (including all header and data bytes) and write the result into the length field in network byte order. Receivers MUST examine the L bit to determine whether the length field is present. If L = '1', a receiver MUST convert the length field to host byte order to determine the length of the TBRPF packet, including the TBRPF packet header. If the underlying delivery service ALSO provides a length field, receivers MUST verify that two lengths match (discarding the packet if they do not.) Similarly, receivers MUST discard any TBRPF packet if neither the underlying delivery service nor the TBRPF packet header provide packet length. R - Router ID (RID) Included: If the underlying delivery service encodes the sender's RID, the sender MAY set R = '0' and omit the RID field. Otherwise, the sender MUST set R = '1' and include a 32-bit unsigned integer RID immediately after any previous header fields. The RID option provides a mechanism for implicit Network-level Address Resolution (NARP) as described in [14]. A receiver that detects an RID option MUST create a NARP binding between the RID and the source address that appears in the network-level header. NB: The first octet of the RID field is NOT GUARANTEED to fall on a natural 32-bit word alignment REGARDLESS of the sense of the 'Z' bit in the TBRPF packet header (see below). Receivers MUST copy the RID field byte-by-byte (rather than as an atomic 32-bit word) when the RID is not aligned on a natural 32-bit boundary. Z - Message Alignment/Compression: If Z = '0', the sender MUST align multi-octet words within the TBRPF packet body on "natural" boundaries (i.e. 64-bit words on integral modulo-8 byte addresses, 32-bit on integral modulo-4 byte addresses and 16-bit words on integral module-2 byte Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 10] INTERNET-DRAFT TBRPF 28 November 2001 addresses.) The sender MUST use Pad1 and/or PadN options (described below) where necessary to achieve natural alignment. If Z = '1', the sender MAY OPTIONALLY omit padding to provide message compression. Since compression may result in multi- octet words aligned on non-integral byte boundaries, receivers MUST process such multi-octet words "byte-by-byte" rather than as atomic n-byte data units. Senders SHOULD avoid message compression to minimize processing overhead for receivers unless maximal message compression is required (e.g. for low- bandwidth links). Receivers MUST be prepared to process both aligned and compressed packets. Version (4 bits): Identifies TBRPF protocol version. The following version field values are defined: TBRPFVERSION_2 2 TBRPFVERSION_1 1 The packet formats defined in this draft represent TBRPF version 2 and MUST be transmitted with the value TBRPFVERSION_2 in the ver- sion field. TBRPF version 2 implementations MAY OPTIONALLY pro- vide backwards compatibility for TBRPFVERSION_1 packets (specified in earlier versions of this draft). Reserved (8 bits): Reserved for future use. 6.2. TBRPF Packet Body The TBRPF packet body consists of the concatenation of a number of elements including one or more TBRPF message(s) and padding options where necessary. These elements are encoded using a method similar to the type-length-value (TLV) encoding method described in pp. 9-10 of Internet RFC 2460 [15] and other documents, except that the length of each element is calculated in a type-specific fashion. Therefore, the "LENGTH" field is subsumed by the "VALUE" field in the symbolic representation. This method encodes elements within the TBRPF packet body using the following format: +-+-+-+-+-+-+-+-+- - - - - | TYPE |RES| VALUE +-+-+-+-+-+-+-+-+- - - - - TYPE (6 bits): A 6-bit identifier with value 0 - 63 that identifies the element. Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 11] INTERNET-DRAFT TBRPF 28 November 2001 RES - Reserved (2 bits): Reserved for future use. VALUE Variable-length field. Format depends on TYPE, as described in the following sections. As specified in [15], the sequence of elements MUST be processed strictly in the order they appear within the TBRPF packet body; a receiver must not, for example, scan through the packet body looking for a particular type of element prior to processing all preceding elements. Additionally, individual elements may have specific align- ment requirements to ensure that multi-octet values fall on natural boundaries. The alignment requirement of an element is specified using the notation "xn+y", meaning the element type must begin at an integer multiple of x octets from the start of the TBRPF packet header, plus y octets. For example: 2n means any 2-octet offset from the start of the TBRPF packet header. 8n+2 means any 8-octet offset from the start of the TBRPF packet header, plus 2 octets. TBRPF packet elements include *padding options* and *messages*. The following subsections describe these elements. 6.2.1. Padding Options (Type = 0 thru 1) As specified in [15], senders may insert two types of padding options where necessary to satisfy alignment requirements for other elements. Padding options may occur anywhere within the TBRPF packet body. When the Z bit in the packet header is '0', senders MUST insert padding options to ensure that the alignment requirements for other elements are met. When Z = '1', padding options MAY be omitted and alignment is NOT guaranteed. The following two padding options are defined: Pad1 option (TYPE = 0) - Alignment requirement: none +-+-+-+-+-+-+-+-+ | 0 |0|0| +-+-+-+-+-+-+-+-+ The Pad1 option inserts one octet of padding into the TBRPF packet body. The RES bits MUST be '0', and the VALUE field is omitted. If more than one octet of padding is required, the PadN option (described next) should be used, rather than multiple Pad1 options. Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 12] INTERNET-DRAFT TBRPF 28 November 2001 PadN option (TYPE = 1) - Alignment requirement: none +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+- - - - - - - - - - - | 1 |0|0| LEN | Zero-valued Octets +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+- - - - - - - - - - - The PadN option inserts two or more octets of padding into the TBRPF packet body. The RES bits MUST be '0' and the first octet of the VALUE field contains an 8-bit unsigned integer length containing a value between 0 - 253 which specifies the number of zero-valued octets that immediately follow, yielding a maximum total of 255 padding octets. 7. TBRPF Neighbor Discovery This section describes the TBRPF Neighbor Discovery (TND) protocol, which allows each node to quickly detect the neighboring nodes with which the node has a direct, bidirectional (2-WAY) link, and to quickly detect the loss of such a link. TND is designed to be independent of the module of TBRPF that per- forms topology discovery and route computation. The interface between these two modules is defined simply by the two functions Link_Up() and Link_Down(), which are called by TND to report a new neighbor or the loss of a neighbor. 7.1. HELLO Message Format The HELLO message has the following three subtypes: - NEIGHBOR REQUEST (TYPE = 2) - NEIGHBOR REPLY (TYPE = 3) - NEIGHBOR LOST (TYPE = 4) Each HELLO subtype has the following format: +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | TYPE |0|0| N | HSEQ | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Neighbor RID(1) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Neighbor RID(2) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ ~ ... ~ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Neighbor RID(N) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 13] INTERNET-DRAFT TBRPF 28 November 2001 - The message body contains N 4-octet router IDs. - HSEQ is the 8-bit HELLO sequence number. A HELLO message is the concatenation of a NEIGHBOR REQUEST message, a NEIGHBOR REPLY message, and a NEIGHBOR LOST message, where each of the last two messages is omitted if its list of router IDs is empty. Thus, a HELLO message always includes a (possibly empty) NEIGHBOR REQUEST. 7.2. Neighbor Table Each node maintains a neighbor table, which stores state information for each known neighbor. The entry for neighbor j contains the fol- lowing variables: nbr_level(j) - The current level (or status) of the link to node j, which can be LOST, 1-WAY, or 2-WAY. nbr_life(j) - The amount of time (in seconds) remaining before nbr_level(j) must be changed to LOST if no further HELLO message from node j is received. Set to NBR_HOLD_TIME whenever a HELLO is received from node j. nbr_hseq(j) - The last value of HSEQ received from node j. Used to determine the number of HELLOs have been missed. nbr_count(j) - The number of times a NEIGHBOR REQUEST/UP/LOST mes- sage has been sent for node j since nbr_level(j) has changed. hello_history(j) - A list of the sequence numbers of the last HELLO_ACQUIRE_WINDOW HELLO messages received from node j. The table entry for a neighbor j may be deleted if no HELLO has been received from node j within the last 2*NBR_HOLD_TIME seconds. (It is kept while the NEIGHBOR LOST for node j is being transmitted.) The absence of an entry for a given node j is equivalent to an entry with nbr_level(j) = LOST and hello_history(j) = NULL. The three possible values of nbr_level(j) at node i have the follow- ing meaning: LOST A neighbor is declared to be LOST if no HELLO message has been received from node j within the last NBR_HOLD_TIME seconds, and remains LOST until HELLO_ACQUIRE_COUNT of the last HELLO_ACQUIRE_WINDOW HELLOs from node j are received. Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 14] INTERNET-DRAFT TBRPF 28 November 2001 1-WAY HELLO_ACQUIRE_COUNT of the last HELLO_ACQUIRE_WINDOW HELLOs from node j have been received, but the link is not 2-WAY. 2-WAY Node i has received from node j either a NEIGHBOR REQUEST or a NEIGHBOR REPLY from node j, and no event has since occurred to change this status. A link that is 2-WAY is considered to be "up". 7.3. Sending HELLO Messages Each node sends a HELLO message periodically every HELLO_INTERVAL seconds, with a random jitter selected from the interval [0, MAX_JITTER]. Each HELLO message always includes a NEIGHBOR REQUEST message, even if its router ID list is empty. The NEIGHBOR REQUEST message includes the sequence number HSEQ, which is incremented each time a HELLO is sent. The HELLO message also includes a NEIGHBOR REPLY message if its router ID list is nonempty, and a NEIGHBOR LOST message if its router ID list is nonempty. The contents of these three messages are determined by the following steps at node i: 1. For each node j such that nbr_level(j) = LOST and nbr_count(j) > 0, include node j in the NEIGHBOR LOST message and decrement nbr_count(j). 2. For each node j such that nbr_level(j) = 1-WAY and nbr_count(j) > 0, include node j in the NEIGHBOR REQUEST message and decrement nbr_count(j). 3. For each node j such that nbr_level(j) = 2-WAY and nbr_count(j) > 0, include node j in the NEIGHBOR REPLY message and decrement nbr_count(j). 7.4. Processing a Received HELLO Message Upon receiving a HELLO message with sequence number HSEQ from node j, node i performs the following steps: 1. If a neighbor table entry does not exist for node j, create one with nbr_level(j) = LOST (temporarily) and nbr_count(j) = 0. Update hello_history(j) to reflect the received HELLO message. 2. If nbr_level(j) = LOST and hello_history(j) indicates that HELLO_ACQUIRE_COUNT of the last HELLO_ACQUIRE_WINDOW HELLO mes- sages from node j have been received: a. If node i does not appear in the NEIGHBOR REQUEST list, set nbr_level(j) = 1-WAY and nbr_count(j) = NBR_HOLD_COUNT (a Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 15] INTERNET-DRAFT TBRPF 28 November 2001 NEIGHBOR REQUEST will be sent). b. If node i appears in the NEIGHBOR REQUEST list, set nbr_level(j) = 2-WAY and nbr_count(j) = NBR_HOLD_COUNT (a NEIGHBOR REPLY will be sent). Call Link_Up(). 3. Else if nbr_level(j) = 1-WAY: a. If node i appears in the NEIGHBOR REQUEST list, then set nbr_level(j) = 2-WAY, nbr_count(j) = NBR_HOLD_COUNT (a NEIGHBOR REPLY will be sent). Call Link_Up(j). b. Else if node i appears in the NEIGHBOR REPLY list, then set nbr_level(j) = 2-WAY and nbr_count(j) = 0. Call Link_Up(j). c. Else if HSEQ - nbr_hseq(j) > NBR_HOLD_COUNT, then set nbr_level(j) = LOST and nbr_count(j) = NBR_HOLD_COUNT (a NEIGH- BOR LOST will be sent). 4. Else if nbr_level(j) = 2-WAY: a. If node i appears in the NEIGHBOR LOST list, set nbr_level(j) = LOST and nbr_count(j) = 0. Call Link_Down(j). b. Else if HSEQ - nbr_hseq(j) > NBR_HOLD_COUNT, set nbr_level(j) = LOST and nbr_count(j) = NBR_HOLD_COUNT (a NEIGHBOR LOST will be sent). c. Else if node i appears in the NEIGHBOR REQUEST list and nbr_count(j) = 0, then set nbr_count(j) = NBR_HOLD_COUNT (a NEIGHBOR REPLY will be sent). 5. Set nbr_life(j) = NBR_HOLD_TIME and nbr_hseq(j) = HSEQ. 7.5. Expiration of Timer nbr_life Upon expiration of the timer nbr_life(j), node i performs the follow- ing step: If nbr_level(j) = 1-WAY or 2-WAY, set nbr_level(j) = LOST and nbr_count(j) = NBR_HOLD_COUNT (a NEIGHBOR LOST will be sent). Call Link_Down(j). 7.6. Link-Layer Failure Indication Some link-layer protocols (e.g., IEEE 802.11) provide an indication that the link to a particular neighbor has failed, e.g., after attempting a maximum number of retransmissions. If such an indication is provided by the link layer, then node i SHOULD perform the Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 16] INTERNET-DRAFT TBRPF 28 November 2001 following step upon receipt of a link-layer failure indication for the link to node j: If nbr_level(j) = 2-WAY, set nbr_level(j) = LOST and nbr_count(j) = NBR_HOLD_COUNT (a NEIGHBOR LOST will be sent). Call Link_Down(j). 7.7. Configurable Parameters This section lists the parameters used in the description of the neighbor discovery protocol, and their proposed default values. HELLO_INTERVAL (1 second) MAX_JITTER (0.1 second) NBR_HOLD_TIME (3 seconds) NBR_HOLD_COUNT (3) HELLO_ACQUIRE_COUNT (2) HELLO_ACQUIRE_WINDOW (3) 8. Topology Discovery and Route Computation This section describes the module of TBRPF that performs topology discovery and route computation. The interface of this module with the neighbor discovery module (TND) is defined by the two functions Link_Up(j) and Link_Down(j), which are called by TND. Therefore, it is possible to use another neighbor discovery mechanism in place of TND, e.g., one that is provided by the link layer. 8.1. Conceptual Data Structures In addition to the information required by the neighbor discovery protocol, each node running TBRPF contains a topology table TT, which stores information for each known link and node in the network. The following information is stored at node i for each known link (u,v) and node u: T(u,v) - Equal to 1 if (u,v) is in node i's source tree T, and 0 otherwise. The previous source tree is also maintained as old_T. RN(u) - Equal to 1 if u is in node i's reportable node set RN, and 0 otherwise. The previous reportable node set is also maintained as old_RN. Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 17] INTERNET-DRAFT TBRPF 28 November 2001 RT(u,v) - Equal to 1 if (u,v) is in node i's reportable subtree RT, and 0 otherwise. Since RT is defined as the set of links (u,v) in T such that u is in RN, this variable need not be maintained explicitly. TG(u,v) = Equal to 1 if (u,v) is in node i's topology graph TG, and 0 otherwise. N - The set of 2-way neighbors of node i. r(u,v) - The list of neighbors that are reporting link (u,v) in their reportable subtree RT. The set of links (u,v) reported by neighbor j is denoted RT_j. r(u) - The list of neighbors that are reporting source u in their reportable node set RN. The set of nodes u reported by neighbor j is denoted RN_j. p(u) - The current parent for source u, equal to the next node on the min-hop path to u. next(u) - The next node on the min-hop path to destination u, used for forwarding data packets. Equal to p(u) in the current version. pred(u) - The node preceding node u in node i's source tree T. Equal to NULL if node u is not reachable. pred(j,u) - The node preceding node u in the subtree RT_j reported by neighbor j. dist(u) - The number of hops on the shortest path to u. reported(u,v) - Equal to 1 if link (u,v) in TG is reported by p(u). tg_expire(u) - Expiration time for links (u,v) in TG. rt_expire(j,u) - Expiration time for links (u,v) in RT_j. nr_expire(u,v) - Expiration time for a link (u,v) in TG such that reported(u,v) = 0. Such non-reported links can be used temporarily during rerouting and usually have an earlier expiration time than tg_expire(u). 8.2. TOPOLOGY UPDATE Message Format The TOPOLOGY UPDATE message has the following format: Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 18] INTERNET-DRAFT TBRPF 28 November 2001 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | TYPE |0|0| n | NRL | NRNL | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Router ID of u | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Router ID of v_1 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ ~ ... ~ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Router ID of v_n | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ The message body contains the N+1 router IDs for nodes u, v_1,...,v_n, which represent the links (u,v_1),..., (u,v_n). The first NRL of the v_k are "reported leaf nodes", the next NRNL of the v_k are "non-reported leaf nodes", and the last n - (NRL+NRNL) of the v_k are "non-reported non-leaf nodes". (The meanings of these terms are defined in the pseudocode.) The TOPOLOGY UPDATE message has the following three subtypes: FULL (TYPE = 5) A FULL update (FULL, n, NRL, NRNL, u, v_1,..., v_n) reports that the links (u,v_1),..., (u,v_n) belong to the sending router's reportable subtree RT, and that RT contains no other links with tail u. ADD (TYPE = 6) An ADD update (ADD, n, NRL, NRNL, u, v_1,..., v_n) reports that the links (u,v_1),..., (u,v_n) have been added to the sending router's reportable subtree RT. DELETE (TYPE = 7) A DELETE update (DELETE, n, NRL, NRNL, u, v_1,..., v_n) reports that the links (u,v_1),..., (u,v_n) have been deleted from the sending router's reportable subtree RT. 8.3. TBRPF Operation This section describes the operation of TBRPF topology discovery and route computation, with the aid of the pseudocode given in the next section. The procedure Update_All() is called periodically every MIN_UPDATE_INTERVAL seconds, which is typically equal to HELLO_INTERVAL. Update_All() first generates a HELLO message. It then calls Update_Routes(), which updates T, TG, and the routing table. Update_RN() is then called to update the reportable node set RN. (Recall that the reportable subtree RT is defined to be the set Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 19] INTERNET-DRAFT TBRPF 28 November 2001 of links (u,v) in T such that u is in RN.) If it is time for a periodic update, Generate_Periodic_Update() is called; otherwise, Generate_Diff_Update() is called, which generates differential topol- ogy updates if RT has changed. The HELLO and TOPOLOGY UPDATE messages are then broadcast to all neighbors. Whenever possible, these mes- sages are combined into the same packet, in order to minimize the number of control packets transmitted. When a packet containing TOPOLOGY UPDATE messages is received, Process_Updates() is called to update the topology table TT, in par- ticular to update TG and the reporting neighbor lists r(u) and r(u,v). If any link in T has been deleted from TG, then Update_Routes() is called to provide immediate rerouting. Link_Up(j) simply adds node j to the neighbor list N, and adds the link (i,j) to TG, which will result in a change to the source tree T when Update_All() is next called. Link_Down(j) removes node j from N and link (i,j) from TG, and calls Update_Routes() to provide immediate rerouting. If the link loss is the result of a link-layer failure indication, a differential update is generated immediately. TBRPF maintains and updates the routing table, which is used by the operating system to forward data packets. TBRPF does not specify how data packets are to be forwarded, but assumes that the forwarding of data packets is compliant with RFC1812. Each node is required to report its reportable subtree RT to neigh- bors. However, each node (independently of the other nodes) MAY report additional links, e.g., to provide increased robustness in highly mobile networks. More precisely, a node may compute any sub- graph H of TG that contains T. It would then report the "reportable subgraph" RH which consists of links (u,v) of H such that u is in RN. RH is then be reported instead of RT to all neighbors, in periodic and differential updates. Additional details for this option will be described in a future version of this draft. Non-relay nodes are implemented simply by not generating or transmit- ting any TOPOLOGY UPDATE messages (only HELLO messages are transmit- ted). Therefore, the following functions can be omitted in non-relay nodes: Update_RN(), Generate_Periodic_Update(), and Generate_Diff_Update(). 8.4. Pseudocode The following pseudocode describes the procedures of TBRPF for topol- ogy discovery and route computation, as performed by node i. Update_All() { Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 20] INTERNET-DRAFT TBRPF 28 November 2001 // Called periodically every MIN_UPDATE_INTERVAL seconds Set msg_list = empty. Generate_HELLO(msg_list). // Adds HELLO to msg_list. Expire_Links(). Update_Routes(). Update_RN(). If (current_time >= next_periodic) { Generate_Periodic_Update(msg_list). Set next_periodic = current_time + PER_UPDATE_INTERVAL.} Else Generate_Diff_Update(msg_list). Broadcast msg_list to neighbors. Set old_T = T and old_RN = RN. } Update_Routes() { For each node v in TT { Set d(v) = INFINITY, pred(v) = NULL. Set old_p(v) = p(v), p(v) = NULL. } Set d(i) = 0, p(i) = i, pred(i) = i. Set S = {i} (labeled nodes). For each node j in N { // c(u,v) = 1 for min-hop routing Set d(j) = c(i,j), pred(j) = i, p(j) = j. } While there exists a node u in TT - S s.t. d(u) < INFINITY { Let u be a node in TT - S that minimizes d(u). // A heap should be used to find u efficiently. Add u to S. If p(u) != old_p(u) { // parent has changed For each link (u,v) in TG { If (p(u) is not in r(u,v) AND reported(u,v) = 1) { Set reported(u,v) = 0. // (u,v) is not reported by p(u) Set nr_expire(u,v) = current_time + PER_UPDATE_INTERVAL.}} If (p(u) is in r(u)) { Set tg_expire(u) = rt_expire(j,u). For each link (u,v) such that p(u) is in r(u,v) { Add (u,v) to TG. Set reported(u,v) = 1. }} // (u,v) is reported by p(u) } For each node v s.t. (u,v) is in TG { // Penalize links (u,v) not reported by p(u). If (reported(u,v) = 0 OR p(u) is not in r(v)) Set cost = c(u,v) + NON_REPORT_PENALTY. If (d(u) + cost < d(v) OR [d(u) + cost = d(v) AND ID(u) < ID(pred(v))]) { Set d(v) = d(u) + c(u,v). Set pred(v) = u. Set p(v) = p(u). }} } Set T = empty set. For each node u in TT - i { Set dist(u) = d(u). // route metric Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 21] INTERNET-DRAFT TBRPF 28 November 2001 If p(u) != NULL, set next(u) = p(u). // route table entry Add link (pred(u), u) to T. } } Update_RN() { // RN is the set of nodes reported by node i. Set RN = empty. // A neighbor j is in RN if some other neighbor s // would select node i as p(j). For each neighbor s in N s.t. j is in r(s) { // Initialize to run Dijkstra for source s, for 2 hops For each node j in N+{i}, Set d(j) = INFINITY, par(j) = NULL. Set d(s) = 0, par(s) = s. For each link (s,j) s.t. s is in r(s,j) and j is in N+{i} { Set d(j) = 1, par(j) = j. For each link (j,k) s.t. j is in r(j,k) and k in in N { Set cost = 1. If (1 + cost < d(k) OR (1 + cost = d(k) AND ID(j) < ID(par(k)))) { Set d(k) = 1 + cost, par(k) = j. }}} // End of Dijkstra for source s For each neighbor j in N { // Add neighbor j to RN if its parent is i. If par(j) = i, add j to RN. } } // End of selection of neighbors in RN Add i to RN. // Node i is always in RN // A non-neighbor node u is in RN if p(u) is in RN. For each node u in TT - i, If p(u) is in RN, add u to RN. } Generate_Periodic_Update(msg_list) { // Generates updates describing the reportable subtree RT. For each node u in RN that is not a leaf of T, add the update (FULL, n, NRL, NRNL, u, v_1,..., v_n) to msg_list, where: v_1,..., v_n are the nodes v such that (u,v) is in T, the first NRL of these are nodes in RN that are leaves of T, the next NRNL of these are nodes in RN that are not leaves of T, and the last n-(NRL+NRNL) of these are not in RN. } Generate_Diff_Update(msg_list) { // Generates updates reporting changes to the reportable // subtree RT. For each node u in RN { If u is not in old_RN and is not a leaf of T { // u was added to RN Add the update (FULL, n, NRL, NRNL, u, v_1,..., v_n) Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 22] INTERNET-DRAFT TBRPF 28 November 2001 to msg_list, where v_1,..., v_n, NRL, and NRNL are defined as above for periodic updates. } Else if u is in old_RN and is not a leaf of T { Let v_1,..., v_n be the nodes v s.t. (u,v) is in T AND { [(u,v) is not in old_T] OR [v is in RN - old_RN AND v is a leaf] OR [v is in old_RN - RN] } If this set of nodes is nonempty { Add the update (ADD, n, NRL, NRNL, u, v_1,..., v_n) to msg_list, where NRL and NRNL are defined as above.} } If (u is in old_RN) { Let v_1,..., v_n be the nodes v s.t. (u,v) is in old_T - TG AND [pred(v) is NULL or is not in RN]. // If pred(v) is in RN, the delete is implied by an add. If this set of nodes is nonempty, add the update (DELETE, n, u, v_1,..., v_n) to msg_list. } } } Process_Updates(j, msg_list) { // Processes a list of update messages from node j. Set update_routes_flag = 0. // Flag will be set to 1 if a link in T is deleted. For each update = (subtype, n, NRL, NRNL, u, v_1,..., v_n) in msg_list { Create an entry for u in TT if it does not exist. Create an entry for u in TT_j if it does not exist. If (subtype = FULL) Process_Full_Update(j, update). If (subtype = ADD) Process_Add_Update(j, update). If (subtype = DELETE) Process_Delete_Update(j, update). // Set update_routes_flag. If update_routes_flag = 0 { For each link (u,v) in TT(u) { If (u,v) is in T but not in TG { Set update_routes_flag = 1. Break. }}} } // End for each update If (update_routes_flag = 1) Update_Routes(). } Process_Full_Update(j, update) { // update = (FULL, n, NRL, NRNL, u, v_1,..., v_n) Add u to RN. Set rt_expire(j,u) = current_time + TOP_HOLD_TIME. For each link (u,v) s.t. j is in r(u,v) { Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 23] INTERNET-DRAFT TBRPF 28 November 2001 Remove j from r(u,v). If pred(j,v) = u, set pred(j,v) = NULL. } If (j = p(u) OR p(u) = NULL) { Set tg_expire(u) = current_time + TOP_HOLD_TIME. For each v s.t. (u,v) is in TG, // Delete old reported links before adding new ones. If reported(u,v) = 1, remove (u,v) from TG. } Process_Add_Update(j, update). } Process_Add_Update(j, update) { // update = (subtype, ADD, n, NRL, NRNL, u, v_1,..., v_n) For m = 1,..., n { Let v = v_m. Create an entry for v in TT if it does not exist. Create an entry for v in TT_j if it does not exist. Add j to r(u,v). If (j = p(u) OR p(u) = NULL) { Add (u,v) to TG. Set reported(u,v) = 1. } // Process implicit delete Set w = pred(j,v). Set pred(j,v) = u. If (w != NULL AND w != u) { Remove j from r(w,v). If (j = p(w)) remove (w,v) from TG. } If (m <= NRL) { // v is a reported leaf Set leaf_update = (FULL, 0, 0, 0, u) Process_Full_Update(j, leaf_update). } If (m > NRL + NRNL) { // v is not reported by j Remove j from r(v). Set rt_expire(j,v) = 0. For each node w s.t. j is in r(v,w) Remove j from r(v,w). If (j = p(v)) For each node w s.t. (v,w) is in TG { Set reported(v,w) = 0. // (v,w) is not reported by p(v) Set nr_expire(u,v) = current_time + PER_UPDATE_INTERVAL.} } } } Process_Delete_Update(j, update) { // update = (DELETE, n, NRL, NRNL, u, v_1,..., v_n) For m = 1,..., n { Let v = v_m. Remove j from r(u,v). Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 24] INTERNET-DRAFT TBRPF 28 November 2001 If pred(j,v) = u, set pred(j,v) = NULL. If (j = p(u)) remove (u,v) from TG. } } Link_Up(j) { // Called when a link to j is discovered If j is not in N { Add j to N. Add (i,j) to TG. Set report(i,j) = 1. } } Link_Down(j) { // Called when the link to neighbor j is lost If j is in N { Remove j from N. Remove (i,j) from TG. If lost link is due to link-layer failure indication { Update_Routes(i). Update_RN(i). Set msg_list = empty. Generate_Diff_Update(i, msg_list). Broadcast msg_list to neighbors. Set old_T = T and old_RN = RN. } Else Update_Routes(i). } } Expire_Links() { For each node u in TT - i { If (tg_expire(u) < current_time) { For each v s.t. (u,v) is in TG, Remove (u,v) from TG. } Else for each v s.t. (u,v) is in TG { If (report(u,v) = 0 AND nr_expire(u,v) < current_time) Remove (u,v) from TG. } For each node j in r(u) { If (rt_expire(j,u) < current_time) { Remove j from r(u). For each link (u,v) s.t. j is in r(u,v) Remove j from r(u,v). }} } } Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 25] INTERNET-DRAFT TBRPF 28 November 2001 8.5. Configurable Parameters This section lists the values of the parameters used in the descrip- tion of the protocol, and their proposed default values. MIN_UPDATE_INTERVAL (1 second) PER_UPDATE_INTERVAL (5 seconds) NON_REPORT_PENALTY (0.01) 9. Application of TBRPF In Mobile Ad-hoc Networks (MANETs) The TBRPF routing protocols provide efficient proactive topology discovery with dynamic adaptation to link state changes in both fixed and mobile environments whether the topology is relatively static or highly dynamic in nature. TBRPF is particularly well suited to MANETs consisting of mobile nodes with wireless network interfaces operating in peer-to-peer fashion over a multiple access communications channel (e.g. the IEEE 802.11 Distributed Coordination Function (DCF) [13].) Although applicable across a much broader field of use, TBRPF is par- ticularly well suited for supporting the standard DARPA Internet pro- tocols [10] [IPv6] as per current practices advocated by the IETF MANET working group. In the following sections, we discuss practical considerations the operation of TRBPF on MANETs. 9.1. Data Link Layer Assumptions We assume a MANET data link layer that supports broadcast, multicast and unicast addressing with best-effort (not guaranteed) delivery services between neighbors (i.e. a pair of nodes within operational communications range of one another). We further assume that each interface belonging to a node in the MANET is assigned a unicast data link layer address that is unique within that particular MANET. While uniqueness for data link layer address assignments is not strictly guaranteed, the assumption of uniqueness is consistent with current practices for deployment of the Internet protocols on specific link layers, e.g. [5]. Methods for duplicate data link layer address detection and deconfliction are beyond the scope of this document. 9.2. Network Layer Assumptions MANETs are formed as collections of MANET routers [14] and non- routing nodes that use network layer addresses when calculating the MANET topology. We assume that each node has at least one data link layer interface (as described above) and that each such interface is assigned a network layer address that is unique within the MANET. Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 26] INTERNET-DRAFT TBRPF 28 November 2001 (Methods for network layer address assignment and duplicate address detection are beyond the scope of this document.) We further assume that each node will select a unique ID for use in TBRPF protocol mes- sages, which we refer to as the Router ID (RID) whether/not the node acts as a MANET router. Finally, we assume that each MANET router supports the multi-hop relay paradigm at the network layer; i.e. each router provides an inter-node forwarding service via network layer host routes which reflect the current MANET topology as perceived by TBRPF. 9.3. Optional Automatic Address Resolution TBRPF employs a proactive neighbor discovery protocol at the network layer that maintains bi-directional link state for neighboring nodes through the periodic transmission of messages. Since each neighbor discovery message contains both the data link and network layer address of the sender, implementations MAY OPTIONALLY employ automatic network-to-data link layer address resolution for the nodes with which they form links. An implementation may use such a mechan- ism to avoid additional message overhead and potential for packet loss associated with on-demand address resolution mechanisms such as ARP [9] or IPv6 Neighbor Discovery [DISC]. Since this behavior is optional, however, implementations MUST respond to on-demand address resolution requests in the normal manner. 9.4. Support for Multiple Interfaces and/or Alias Addresses MANET nodes may comprise multiple interfaces; each with a unique net- work layer address. Additionally, MANET nodes may wish to publish *alias* addresses such as when multiple network layer addresses are assigned to the same interface or when the MANET node is serving as a Mobile IP [MobileIP] home agent for nodes that send binding updates from foreign networks. Multiple interfaces and network layer alias addresses will be handled in a manner similar to that described in section A.4.8 of [OSPFv6], i.e. bindings between the MANET node's RID and any additional network layer addresses will be established. A future version of this draft will specify a new message format and protocol extension to support this function. 9.5. Support for Hierarchical Network Prefixes MANET routers may wish to advertise hierarchical network prefixes which the router discovered via attached networks, route redistribu- tions from other routing protocols, or other means. Hierarchical net- work prefixes will be handled in a manner similar to that described in section A.4.9 of [OSPFv6], i.e. bindings between the MANET node's RID and any hierarchical network prefixes will be established. A future version of this draft will specify a new message format and Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 27] INTERNET-DRAFT TBRPF 28 November 2001 protocol extension to support this function. 9.6. Support for non-MANET Hosts Non-MANET hosts may establish connections to MANET routers through on-demand mechanisms such as ARP or IPv6 Neighbor Discovery. Such connections do not constitute a MANET link and therefore are not reported in TBRPF topology updates. Future versions of this draft will specify methods for incorporating non-MANET hosts, such as requiring that MANET routers initiate periodic Neighbor Unreachabil- ity Detection to such hosts to detect connection failures. 9.7. Internet Protocol Considerations As with [OSPF], TBRPF packets are sent directly over the Internet Protocol's network layer and are encapsulated solely by IP and local data-link headers. An official IP Protocol number registration [IANA] for TBRPF will be procured in due course. The same IANA regis- tration will be used whether encapsulation is via IPv4 or IPv6. (IPv6 treats the IP protocol number as a "Next Header" designator). While IP encapsulation is specified, implementations MAY OPTIONALLY employ alternate data delivery services (e.g. UDP/IP, local data-link encap- sulation) in private networks. The selection of an alternate data delivery service MUST be consistent among all MANET routers in the private network. The following subsections discuss the operation of TBRPF over IPv4 and IPv6, respectively. 9.7.1. IPv4 Operation When the IPv4 protocol is used, all TBRPF packets are sent to the "All_MANET_Neighbors" multicast address (see: "IANA Considerations") and thus reach all MANET routers within single-hop transmission range of the sender. MANET routers MUST NOT forward packets sent to the "All_MANET_Neighbors" multicast address. Since packet loss due to link failure, interference, etc. can occur, implementations SHOULD split large TBRPF protocol packets into several smaller packets to avoid IPv4 fragmentation/reassembly. Since packets sent to "All_MANET_Neighbors" are delivered only to single-hop neighbors, Path MTU Discovery is not required. Instead, MANET routers can derive the maximum TBRPF packet length(s) directly from the maximum transmission unit(s) (MTUs) of their attached interface(s). Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 28] INTERNET-DRAFT TBRPF 28 November 2001 9.7.2. IPv6 Operation Transition mechanisms described in the IETF NGTRANS working group (e.g. ISATAP) enable IPv6 operation over IPv4 routing infrastruc- tures. ISATAP [19] can be used on TBRPF MANETs to enable automatic IPv6-in-IPv4 operation regardless of route changes due to mobility. Future versions of this draft will specify a native IPv6 capability for TBRPF using mechanisms similar to those specified in [OSPFv6]. Packet formats which implement such mechanisms will use 4-octet router ID's instead of 16-octet IPv6 addresses for greater effi- ciency. 10. IANA Considerations An official IANA IP protocol number assignment is required for TBRPF protocol messages. Additionally, an IANA IPv4 multicast address assignment for "All_MANET_Neighbors" is required from the range of addresses between 224.0.0.0 and 224.0.0.255 inclusive. This range of addresses is reserved for the use of routing protocols and other low-level topology discovery or maintenance protocols, such as gate- way discovery and group membership reporting. Multicast routers should not forward any multicast datagram with destination addresses in this range, regardless of its TTL [12]. Note that such an IPv4 multicast address assignment may have general application for MANET beyond its anticipated use for TBRPF. We there- fore propose that MANET consider the procurement of an "All_MANET_Neighbors" IPv4 multicast address from the range of addresses between 224.0.0.0 and 224.0.0.255. 11. Implementation Status The current version of TBRPF has been implemented in Linux and ns-2. TBRPF version 1 (as described in version 0 of the draft) has been implemented in the FreeBSD V3.2 operating system with the Merit Multi-Threaded Routing Toolkit daemon (mrtd). The initial FreeBSD V3.2 implementation has been in use for laboratory and fielded exper- iments since June 1999. TBRPF version 1 has been ported to Linux. The current port runs on RedHat Version 7.0 and has been tested under multiple Linux kernel releases including 2.2.16 and a 2.4 pre-release. Additionally, the Linux and FreeBSD ports are fully interoperable under TBRPF Version 1. Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 29] INTERNET-DRAFT TBRPF 28 November 2001 12. Patent Rights Statement SRI International has filed one or more patents protecting the inven- tions described in this submission. If SRI's submission (or portions thereof) is accepted and becomes part of the IETF standard, then SRI will grant royalty-free permission under such patents for both com- mercial and non-commercial uses, to the extent necessary for compli- ance with the standard. Any aspects or variants of SRI's submission that are not included in the IETF standard and/or are not necessary for compliance with the standard may require an additional license from SRI under terms to be negotiated by the parties. 13. Acknowledgments The authors would like to thank ASEO for funding this work, and the following people for helping in the development of TBRPF: Yonael Gorfu and Julie S. Wong. 14. References [1] Bhargav Bellur and Richard G. Ogier. A Reliable, Efficient Topology Broadcast Protocol for Dynamic Networks. Proc. IEEE INFOCOM '99, New York, March 1999. [2] Scott Bradner. Key words for use in RFCs to Indicate Require- ment Levels. RFC 2119, March 1997. [3] M. Scott Corson and Joe Macker. Mobile Ad hoc Networking (MANET): Routing Protocol Performance Issues and Evaluation Con- sideration. RFC 2501, 1999. [4] J.J. Garcia-Luna-Aceves and M. Spohn. Efficient Routing in Packet-Radio Networks Using Link-State Information. Proc. IEEE WCNC '99, September 1999. [5] C. Hornig. Standard for the Transmission of IP Datagrams over Ethernet Networks. STD0041, April 1984. [6] David B. Johnson and David A. Maltz. Protocols for Adaptive Wireless and Mobile Networking. IEEE Personal Communications, Vol. 3, No. 1, pp 34-42, February 1996. [7] Richard G. Ogier. Efficient Routing Protocols for Packet-Radio Networks Based on Tree Sharing. Proc. Sixth IEEE Intl. Workshop on Mobile Multimedia Communications (MOMUC'99), November 1999. [8] Charles E. Perkins, Elizabeth M Royer, and Samir R. Das. Ad Hoc On-Demand Distance Vector (AODV) Routing. draft-ietf-manet- aodv-05.txt, March 2000 (work in progress). Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 30] INTERNET-DRAFT TBRPF 28 November 2001 [9] D.C. Plummer. An Ethernet address resolution protocol: Or con- verting network protocol addresses to 48.bit Ethernet addresses for transmission on Ethernet hardware. RFC 826, November 1982. [10] J. Postel. Internet Protocol. RFC 791, September 1981. [11] J. Postel. User Datagram Protocol. RFC 768, August 1980. [12] J. Reynolds and J. Postel. Assigned Numbers. RFC 1700, October 1994. [13] Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) Specifications, ISO/IEC Std. 8802-11, ANSI/IEEE Std 802.11, 1999. [14] An Internet MANET Encapsulation Protocol (IMEP) Specification, draft-ietf-manet-imep-spec-01.txt [15] S. Deering and R. Hinden. Internet Protocol Version 6 (IPv6) Specification. RFC 2460, December 1998. [16] S. Corson. MANET Routing Protocol Applicability Statement, draft-ietf-manet-appl-00.txt, November 1998 (work in progress). [17] J. Moy. OSPF Version 2. RFC 1583, March 1994. [18] P. Jacquet et al. Optimized Link State Routing Protocol, draft-ietf-manet-olsr-04.txt, March 2001 (work in progress). [19] F. Templin. "Intra-Site Automatic Tunnel Addressing Protocol (ISATAP)", draft-ietf-ngtrans-isatap-02.txt (work in progress) [20] D. Bertsekas and R. Gallager. Data Networks, Prentice-Hall, 1987. Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 31] INTERNET-DRAFT TBRPF 28 November 2001 Authors' Addresses Bhargav Bellur SRI International 333 Ravenswood Ave. Menlo Park, CA 94025 USA Phone: +1 650 859-6335 Fax: +1 650 859-4812 Email: bhargav@erg.sri.com Richard Ogier SRI International 333 Ravenswood Ave. Menlo Park, CA 94025 USA Phone: +1 650 859-4216 Fax: +1 650 859-4812 Email: ogier@erg.sri.com Fred L. Templin SRI International 333 Ravenswood Ave. Menlo Park, CA 94025 USA Phone: +1 650 859-3144 Fax: +1 650 859-4812 Email: templin@erg.sri.com Ogier, Templin, Bellur, Lewis Expires 28 May 2002 [Page 32]