Internet Draft Danny Cohen Myricom Expires in six months Craig Lund Mercury Computers Tony Skjellum, Thom McMahon, Robert George Mississippi State University June 1998 The Router-to-Router (RRP) PacketWay Protocol for High-Performance Interconnection of Computer Clusters Status of this Memo This document is an Internet-Draft. 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." To view the entire list of current Internet-Drafts, please check the "1id-abstracts.txt" listing contained in the Internet-Drafts Shadow Directories on ftp.is.co.za (Africa), ftp.nordu.net (Northern Europe), ftp.nis.garr.it (Southern Europe), munnari.oz.au (Pacific Rim), ftp.ietf.org (US East Coast), or ftp.isi.edu (US West Coast). Table of Contents Introduction ..................................................... 2 Notations ........................................................ 2 PacketWay and IP ................................................. 3 Node Attributes .................................................. 3 RRP Messages ..................................................... 4 Structure of RRP Messages ........................................ 5 RRP Records ...................................................... 8 Example .......................................................... 12 Glossary ......................................................... 16 Acronyms and Abbreviations ....................................... 18 Editor's Address ................................................. 19 Cohen et al. Experimental [Page 1] Internet-Draft PacketWay RRP May 1998 Introduction The PacketWay family of protocols is introduced in the "The End-to- End (EEP) PacketWay Protocol for High-Performance Interconnection of Computer Clusters". This document defines the Router-to-Router Protocol (RRP), the basic messages used by routers to exchange routing information with endpoints and each other. In the PacketWay model a router is a set of cooperating hosts on two (or more) networks. These hosts, each a full-fledged host on its SAN, are called "half-routers" (HRs). RRP defines, via message structure and behavior, the interactions between HRs as well as the interactions between HRs and nodes. RRP does not define the lower level protocols that deliver its messages. RRP also does not define the connection between the HRs within the router-- these are left for mutual agreements among the implementors of each HR. However, the intra-router communication among these hosts is a "public" issue, handled according to the RRP which defines only the Network-level [Level-3], and not the lower levels of this communication. All RRP messages are carried via EEP packets with the "Packet-Type" field of the EEP header set to "RRP". This document does not define how source routes are initially constructed. It is expected that static tables may be manually maintained for simple or very stable systems. Dynamic table- maintenance protocols will likely be outlined in a future document. Notations 8B means "8-byte" (64 bits). 0x indicates hexadecimal values, e.g., 0x0100 is 2^8=256(decimal). 0b indicates binary values, e.g., 0b0100 is 4(decimal). xxxx indicate a field that is discarded without any checking (e.g., padding). [exp] in equations, is the integral part, rounded down, of `exp` (e.g., [23/8]=2). All length fields do not include themselves, and therefore may be 0. Lengths are specified either (a) by byte count, implying that some padding bytes may follow to fill 8B-words, or (b) by 8B-word count and PL, the number of trailing padding bytes (with PL between 0 and 7). Cohen et al. Experimental [Page 2] Internet-Draft PacketWay RRP May 1998 PacketWay and IP The architecture of PacketWay is very similar to the IP family (in fact it heavily borrows from IP), with emphasis on performance not generality and scaleability as was selected for IP. Like IP, PacketWay is based on an End-to-End protocol (EEP) that assumes that if an address (or equivalent specification of the desti- nation) is placed in the appropriate field in the packet header, then the packet will arrive to that destination. Neither IP nor EEP specify how this happens. Routers are responsible for transferring packets from their source networks to their destination networks (possibly via other networks). The communication among the routers (such the entire family of the GGPs [Gateway/Gateway Protocols] as they were originally called) is NOT a part of IP (as defined originally in RFC-791 and MIL-STD-1777). Similarly, it is not a part of EEP. Like the IP family, PacketWay defines separately its Router-to-Router Protocol (RRP), in a device- and network-independent manner. However, the model of routers in PacketWay is slightly different from the original model in the IP family. IP routers (or gateways as they were called then) are monolithic devices, provided by their vendors. Each IP-router is a bona-fide host on two (or more) networks. The communication among these intra-router hosts is an internal "private" issue, handled by each vendor as it sees fit, not subject to pub- lished standards. Node Attributes Each node must have a Physical Address. Optionally it may also have Name, Capabilities, and Logical-Addresses: Physical Address 23 bits, flat, unique in this PacketWay. Name flat, globally unique (e.g., IP address), arbi- trary length Capabilities regular GP node, router, PacketWay-server, NFS, paging server, M/C server, SRVLOC-server, DSP, printer, etc. Some capabilities may need additional parameters (e.g., SAN-ID for routers, and resolution+colors for printers). Their parameters are capability- specific. The capabilities are defined in the PacketWay Enumeration document. Cohen et al. Experimental [Page 3] Internet-Draft PacketWay RRP May 1998 Logical-Addresses a set of (logical) addresses to which this node requests to listen. Logical addresses designate multicast and broadcast groups. The control of the Logical-Addresses (a la IGMP) is not defined in this document. This will be designed by the applications that use it (e.g., PacketWay-Multicast). The management of logical addresses (e.g., JOIN and LEAVE) is not defined here. RRP Messages RRP messages are PacketWay messages with PT="RRP" and TE="RRP-Type" in their EEP-header, followed by zero or more RRP-records according to their RRP-type and completed by the TAIL which is the EI field of the EEP packet. The RRP-records are defined in the next sections. The RRP-records constitute the Data Block (DB) of the PacketWay- message. They must be in Big-Endian order, with e=0 in the EEP- header. We use "[XXX]" to indicate the RRP-message XXX, and to indicate the RRP-record YYY. XXX is the RRP-Type, carried in the Type Extension (TE) field of the EEP header (with Packet-Type of "RRP"), and YYY is the RTyp field, carried in the first byte of that RRP-record. Following are the 7 RRP messages, with their RRP-type, and the related error messages. The column S->D (Source to Destination) shows who sends such messages to whom, where N is for Node, H is for HR, and A is for Any. RRP- Type S->D Description -------- ------ ----------------------------------------------- [GVL2] N->H Please give me L2-routes to node (address) Replies to [GVL2]: [L2SR], [RDRC], or [ERR/UNK]. [L2SR] H->N Here are L2-routes to node (address) [HRTO] N->H Which HR should I use for node (address)? Replies to [HRTO]: [RDRC] or [ERR/UNK]. [RDRC] H->N Re-direct to node (address) via an HR on same SAN [TELL] N->H Please tell me about node (address, name, capa's) The reply to [TELL] is [INFO], or [ERR/UNK]. Cohen et al. Experimental [Page 4] Internet-Draft PacketWay RRP May 1998 [INFO] A->A Info about node (address, name, capabilities, LAs) [WRU?] A->A Who/what-Are-You? (Tell me all about yourself) The reply to [WRU?] is [INFO] about the replier. RRP also uses the following error messages: [ERR/UNK] Destination Unknown (address) [ERR/HRDOWN] HR Down [ERR/LKDOWN] Link Down [ERR/GENERAL] General error message [GVL2] Please give me L2-routes from you to node (address) PH (with [PT/TE]=[RRP/GVL2]) (address of the node for which [L2SR] is requested) Structure of RRP Messages [L2SR] Here are L2-routes from me to node (address) PH (with [PT/TE]=[RRP/L2SR]) (address of the node for which the following is provided) (Source Route/Quality record) (optional) MTU records for the above This message may have several (, ) pairs, one such pair for each source route. [HRTO] Which HR should I use for node (address) PH (with [PT/TE]=[RRP/HRTO]) (address of the node for which initial HR is requested) [RDRC] Re-direct to destination node (address) via a HR (address), on the same SAN. PH (with [PT/TE]=[RRP/RDRC]) (address of the destination node) (address of the HR to be used for that destination) The above addresses are expected to be physical (but they be otherwise). Cohen et al. Experimental [Page 5] Internet-Draft PacketWay RRP May 1998 [TELL] Please tell me about node (address | name | capabilities) PH (with [PT/TE]=[RRP/TELL]) (address of that node) or PH (with [PT/TE]=[RRP/TELL]) (name of that node) or PH (with [PT/TE]=[RRP/TELL]) (capabilities for which nodes are requested) This message may have several 's, one for each capability. [TELL] identifies a node by an address and/or a name and/or capabilities. If more than one attribute is specified (e.g., an address and name(s)) any nodes that meets any of them should be considered (like an implied OR). [INFO] Info about node(s) (address, name, capabilities) PH (with [PT/TE]=[RRP/INFO]) (address of that node) (name of that node) (capabilities for which nodes are requested) (Logical-Addresses for the requested node) This message may have several 's, one for each capability. For nodes without , , or any , these records are omitted. [INFO] provides all the known information about all the nodes that match the [TELL]. The -records are the separators between the nodes. [WRU?] Who/what-Are-You? PH (with [PT/TE]=[RRP/WRU?] and [DD]=0x7FFFFE) Cohen et al. Experimental [Page 6] Internet-Draft PacketWay RRP May 1998 [ERR/UNK] Destination Unknown (address) PH (with [PT/TE]=ERROR/UNK) (XXXX of the Destination node for which the requested information is not available), where >XXXX> is the and/or of the node(s) about which this message is sent [ERR/HRDOWN] HR Down (or Router-Down) PH (with [PT/TE]=[ERROR/HRDOWN]) (address of the HR that is down) (the other address of the router that is down) [ERR/LINKDOWN] Link Down PH (with [PT/TE]=[ERROR/LINKDOWN]) (address of one end of the link that is down) (address of the other end of the link that is down) [ERR/GENERAL] General Error (i.e., none of the above) PH (with [PT/TE]=[ERROR/GENERAL]) XX (The entire message that caused the error : PH+OH+DB+TAIL) Cohen et al. Experimental [Page 7] Internet-Draft PacketWay RRP May 1998 RRP Records Each RRP-record starts with an 8B-word header as shown below. Its first byte identifies the record type (RTyp). The second byte is the Pad-Count byte (PL) indicating the number of padding bytes. The third and the fourth bytes (RL) are the length (in 8B-words) of the record, excluding the record header, hence it may be zero. The rest of the header bytes depend on the record type (RTyp). +--------+--------+--------+--------+--------+--------+--------+--------+ | RTyp | PL | RL |........|........|........|........| +--------+--------+--------+--------+--------+--------+--------+--------+ Some records that have an arbitrary length are "right justified" by having PL padding bytes before the data (Padding Before Data [PBD]). Some records that have an arbitrary length are "left justified" by having PL bytes after the data (Padding After Data [PAD]). In either case the total number of data bytes is: (8*RL+4-PL). Following are the RRP-records. These records are the building blocks used to construct RRP-messages. In the following, "xxxx" indicate bytes that are discarded, such as for padding. It is recommended to set them to all-0. ===> Node-Address Record [PAD] This record specifies either a single address (with AT=1) or a range of addresses (with AT=2 followed by AT=3, or by AT=4 followed by AT=5). AT is the "Address-Type". 0 1 2 3 4 5 6 7 +--------+--------+--------+--------+--------+--------+--------+--------+ | | PL=0 | RL=0 | AT=1 | PacketWay-Address | +--------+--------+--------+--------+--------+--------+--------+--------+ or: 0 1 2 3 4 5 6 7 +--------+--------+--------+--------+--------+--------+--------+--------+ | | PL=4 | RL=1 | AT=2 | Min-PacketWay-Address | +--------+--------+--------+--------+--------+--------+--------+--------+ | AT=3 | Max-PacketWay-Address | xxxx | xxxx | xxxx | xxxx | +--------+--------+--------+--------+--------+--------+--------+--------+ or: 0 1 2 3 4 5 6 7 +--------+--------+--------+--------+--------+--------+--------+--------+ | | PL=4 | RL=1 | AT=4 | PacketWay-Address-Value | +--------+--------+--------+--------+--------+--------+--------+--------+ | AT=5 | PacketWay-Address-Mask | xxxx | xxxx | xxxx | xxxx | +--------+--------+--------+--------+--------+--------+--------+--------+ Cohen et al. Experimental [Page 8] Internet-Draft PacketWay RRP May 1998 The address-mask follows the address-value. The above addresses may be physical or logical. The address X is specified by an -record if: if AT=1: X == PacketWay-Address if AT=2,3: Min-PacketWay-Address <= X <= Max-PacketWay-Address if AT=4,5: (PacketWay-Address-Mask & X) == PacketWay-Address-Value An -record defines only one PacketWay-address (or one range), unlike an -record (see below) that may specify multiple addresses and multiple address-ranges. If the -record is followed by other records that describe the same node (such as , , , , and ) then the RL of the -records also covers all these records. All these records apply to all the addresses specified in this -record. Needless to say that is not expected to appear within a record that specifies more than one address. Hence, if an -record with AT=1 has RL>1, or if an -record with AT>1 has RL>2, then this -record includes additional records (such as , , , and/or ) about the specified address(es). The enumeration is guaranteed not to have overlap between the AT and the RTyp codes. ===> Node-Name Record [PAD] (e.g., a name with 7 bytes B1..B7) 0 1 2 3 4 5 6 7 +--------+--------+--------+--------+--------+--------+--------+--------+ | | PL=3 | RL=1 | B1 | B2 | B3 | B4 | +--------+--------+--------+--------+--------+--------+--------+--------+ | B5 | B6 | B7 | xxxx | xxxx | xxxx | xxxx | xxxx | +--------+--------+--------+--------+--------+--------+--------+--------+ The number of bytes in the name is 8*RL+4-PL. Cohen et al. Experimental [Page 9] Internet-Draft PacketWay RRP May 1998 ===> Node-Capability Record [PAD] (e.g., 9 parameter bytes) 0 1 2 3 4 5 6 7 +--------+--------+--------+--------+--------+--------+--------+--------+ | | PL=2 | RL=1 | CC=Cx | P1 | P2 | P3 | +--------+--------+--------+--------+--------+--------+--------+--------+ | P4 | P5 | P6 | P7 | P8 | P9 | xxxx | xxxx | +--------+--------+--------+--------+--------+--------+--------+--------+ Byte#4 is the Capability Code, CC, followed by as many parameter bytes as needed (9 in the above example). The capability codes are listed in the PacketWay Enumeration docu- ment. The number of bytes used by the parameters is 8*RL+3-PL. ===> Logical-Addresses Record [PAD] (e.g., 2 logical addresses and a range of logical addresses) 0 1 2 3 4 5 6 7 +--------+--------+--------+--------+--------+--------+--------+--------+ | | PL=4 | RL=2 | AT=1 |1110 Logical-Address-#1 | +--------+--------+--------+--------+--------+--------+--------+--------+ | AT=2 |1110 Min-Logical-Address | AT=3 |1110 Max-Logical-Address | +--------+--------+--------+--------+--------+--------+--------+--------+ | AT=1 |1110 Logical-Address-#2 | xxxx | xxxx | xxxx | xxxx | +--------+--------+--------+--------+--------+--------+--------+--------+ Whereas an -record defines only one PacketWay-address (or one range), an -record may specify multiple addresses (each with AT=1) and multiple ranges (each with a pair of AT=2,3 or AT=4,5). ===> Source-Route Record [PBD], with Q for that route. (e.g., an SR combined of 2 L2RHs, one with 13 bytes and one with 4 bytes) This record carries one, or more, L2RHs (2 in the following example, one with SR of 13B, followed by an SR of 5B). Cohen et al. Experimental [Page 10] Internet-Draft PacketWay RRP May 1998 0 1 2 3 4 5 6 7 +--------+--------+--------+--------+--------+--------+--------+--------+ | | PL=2 | RL=3 | xxxx | xxxx | Q | +--------+--------+--------+--------+--------+--------+--------+--------+ |vv000000|10 L=13B| SR01 | SR02 | SR03 | SR04 | SR05 | SR06 | +--------+--------+--------+--------+--------+--------+--------+--------+ | SR07 | SR08 | SR09 | SR10 | SR11 | SR12 | SR13 | xxxx | +--------+--------+--------+--------+--------+--------+--------+--------+ |vv000000|10 L=4B | SR01 | SR02 | SR03 | SR04 | xxxx | xxxx | +--------+--------+--------+--------+--------+--------+--------+--------+ Q (the Route Quality) is an unsigned 16-bit integer. The units are not defined here. It is assumed that it is monotonic with all-0 being the best and all-1 the worst. If there is an (MTU-record) for that SR it should follow this -record. However, the RL of the does not include the RL of the . ===> MTU record [PBD] 0 1 2 3 4 5 6 7 +--------+--------+--------+--------+--------+--------+--------+--------+ | | PL=0 | RL=0 | MTU (in 8B-words) | +--------+--------+--------+--------+--------+--------+--------+--------+ The MTU record provides the MTU for the SR defined before (by an ). The value of 0 means indefinite MTU (i.e., any length is OK). Cohen et al. Experimental [Page 11] Internet-Draft PacketWay RRP May 1998 Example In the following PacketWay network used for this example, 3 SANs are interconnected via 2 routers, Router-A (RTRA) between SAN1 and SAN3, and RTRB between SAN1 and SAN2. +-------+ +--0--+ SAN1 +--0--+ +--0--+ | Node1 +----------3 SW0 1----------3 SW1 1----------3 SW2 1 MTU=16KB +-------+ +--2--+ +--2--+ +--2--+ | | RTRA1 *********** +---+---+ *********** RTRB1 * RouterA * | Node2 | * RouterB * RTRA3 *********** +---+---+ *********** RTRB2 | | | +-------+ SAN3 +--0--+ +--0--+ SAN2 +--0--+ | Node3 +----------3 SW3 1 3 SW4 1----------3 SW5 1 MTU=8KB +-------+ +--2--+ +--2--+ +--2--+ In this example Node1 on SAN1 (with MTU=16KB) is looking for Node2 which is on SAN2 (with MTU=8KB). It first asks its default router (RTRA1) for an L2RH to Node2. RTRA1 redirects Node1 to RTRB1 regarding Node2. Node1 asks RTRA1 (by [HRTO], in message M1) which router to use for Node2. RTRA1 suggests (using [RDRC], M2) to use RouterB. Node1 uses L3-forwarding ([WRU?], M3), via Router-B, to verify that RTRB can indeed get to Node2, by asking Node2 for information about itself. Node2 provides this information ([TELL], M4) which Node1 likes. Node1 asks RouterB ([GVL2], M5) for L2RH(s) to Node2. RouterB pro- vides ([L2SR], M6) the requested L2RH with its MTU of 1,024 8B-words (8KB). Finally, Node1 sends data (by M7) to Node2 using L2-forwarding. Similarly, Node2 may ask its default router which HR to use for Node1 and for L2RH(s) to Node1. The sequence of messages (M1 thru M7) is shown below. (M1) Node1 sends [HRTO] to its default router RTRA1 asking which HR to use for node2. Cohen et al. Experimental [Page 12] Internet-Draft PacketWay RRP May 1998 0 1 2 3 4 5 6 7 +-----------------------------------------------------------------------+ | <---- The L2-header needed to get from Node1 to RouterA1 ----> | | It may be any number of bytes. In this example it's 9 bytes:230000000| +--------+--------+--------+--------+--------+--------+--------+--------+ |00 P |0 RTRA1 | "HRTO" | "R R P" | +---+----+--------+--------+--------+-+------+--------+--------+--------+ |E=0|PL=0| Data-Length=1 (8B-words) |0| RZ |0 Node1 | +---+----+--------+--------+--------+-+------+--------+--------+--------+ | | PL=0 | RL=0 | AT=1 |0 Node2 | +--------+--------+--------+--------+--------+--------+--------+--------+ | 64 zero bits, unless any error was indicated along the path | +--------+--------+--------+--------+--------+--------+--------+--------+ (M2) RTRA1 uses [RDRC] to re-direct to Node2 via RouterB. 0 1 2 3 4 5 6 7 +-----------------------------------------------------------------------+ | <---- The L2-header needed to get from RouterA1 to Node1 ----> | | It may be any number of bytes. In this example it's 9 bytes:330000000| +--------+--------+--------+--------+--------+--------+--------+--------+ |00 P |0 Node1 | "RDRC" | "R R P" | +---+----+--------+--------+--------+-+------+--------+--------+--------+ |E=0|PL=0| Data-Length=2 (8B-words) |0| RZ |0 RTRA1 | +---+----+--------+--------+--------+-+------+--------+--------+--------+ | | PL=0 | RL=0 | AT=1 |0 Node2 | +--------+--------+--------+--------+--------+--------+--------+--------+ | | PL=0 | RL=0 | AT=1 |0 RTRB1 | +--------+--------+--------+--------+--------+--------+--------+--------+ | 64 zero bits, unless any error was indicated along the path | +--------+--------+--------+--------+--------+--------+--------+--------+ Node1 knows how to get to RouterB over its SAN. (M3) Node1 uses [WRU?] (still using L3-forwarding via RouterB) to verify the capabilities of Node-2, and that RTRB can indeed get to it. This is done by asking Node2 for information about itself. 0 1 2 3 4 5 6 7 +-----------------------------------------------------------------------+ | <---- The L2-header needed to get from Node1 to RouterB1 ----> | | It may be any number of bytes. Here it is 11 bytes: 11230000000 | +--------+--------+--------+--------+--------+--------+--------+--------+ |00 P |0 Node2 | "WRU?" | "R R P" | +---+----+--------+--------+--------+-+------+--------+--------+--------+ |E=0|PL=0| Data-Length=0 (8B-words) |0| RZ |0 Node1 | +---+----+--------+--------+--------+-+------+--------+--------+--------+ | 64 zero bits, unless any error was indicated along the path | +--------+--------+--------+--------+--------+--------+--------+--------+ Cohen et al. Experimental [Page 13] Internet-Draft PacketWay RRP May 1998 (M4) Node2 uses [INFO] (via RouterB2, also using L3-forwarding) to provide information about itself to Node1. This info includes its PacketWay-address and its name ("Super"). If Node2 had implemented also Level-C of the RRP it would also provide a record about its capabilities (as shown in this example with 2 capabilities (with codes of 5 and 7). 0 1 2 3 4 5 6 7 +-----------------------------------------------------------------------+ | <---- The L2-header needed to get from Node2 to RouterB2 ----> | | It may be any number of bytes. Here it is 10 bytes: 1030000000 | +--------+--------+--------+--------+--------+--------+--------+--------+ |00 P |0 Node1 | "INFO" | "R R P" | +---+----+--------+--------+--------+-+------+--------+--------+--------+ |E=0|PL=0| Data-Length=5 (8B-words) |0| RZ |0 Node2 | +---+----+--------+--------+--------+-+------+--------+--------+--------+ | | PL=0 | RL=4 | AT=1 |0 Node2 | +--------+--------+--------+--------+--------+--------+--------+--------+ | | PL=7 | RL=1 | "S" | "u" | "p" | "e" | +--------+--------+--------+--------+--------+--------+--------+--------+ | "r" | xxxx | xxxx | xxxx | xxxx | xxxx | xxxx | xxxx | +--------+--------+--------+--------+--------+--------+--------+--------+ | | PL=1 | RL=0 | CC=7 | 4 | 8 | xxxx | +--------+--------+--------+--------+--------+--------+--------+--------+ | | PL=3 | RL=0 | CC=5 | xxxx | xxxx | xxxx | +--------+--------+--------+--------+--------+--------+--------+--------+ | 64 zero bits, unless any error was indicated along the path | +--------+--------+--------+--------+--------+--------+--------+--------+ By receiving this message Node1 knows that RTRB could indeed be used for communication with Node2. (M5) Node1 uses [GVL2] to ask RouterB for L2RH(s) from RouterB to Node2. 0 1 2 3 4 5 6 7 +-----------------------------------------------------------------------+ | <---- The L2-header needed to get from Node1 to RouterB1 ----> | | It may be any number of bytes. Here it is 11 bytes: 11230000000 | +--------+--------+--------+--------+--------+--------+--------+--------+ |00 P |0 RTRB1 | "GVL2" | "R R P" | +---+----+--------+--------+--------+-+------+--------+--------+--------+ |E=0|PL=0| Data-Length=1 (8B-words) |0| RZ |0 Node1 | +---+----+--------+--------+--------+-+------+--------+--------+--------+ | | PL=0 | RL=0 | AT=1 |0 Node2 | +--------+--------+--------+--------+--------+--------+--------+--------+ | 64 zero bits, unless any error was indicated along the path | +--------+--------+--------+--------+--------+--------+--------+--------+ Cohen et al. Experimental [Page 14] Internet-Draft PacketWay RRP May 1998 (M6) RouterB uses [L2SR] to provide Node1 with an L2RH from RTRB2 to Node2, with its Q and MTU. This L2RH is {3,0,3,0,0,0,0,0,0,0} from RouterB to Node2, and the MTU is 1,024 (meaning 8KB). 0 1 2 3 4 5 6 7 +-----------------------------------------------------------------------+ | <---- The L2-header needed to get from RouterB1 to Node1 ----> | | It may be any number of bytes. Here it is 11 bytes: 33330000000 | +--------+--------+--------+--------+--------+--------+--------+--------+ |00 P |0 Node1 | "L2SR" | "R R P" | +---+----+--------+--------+--------+-+------+--------+--------+--------+ |E=0|PL=0| Data-Length=4 (8B-words) |0| RZ |0 RTRA1 | +---+----+--------+--------+--------+-+------+--------+--------+--------+ | | PL=0 | RL=3 | AT=1 |0 Node2 | +--------+--------+--------+--------+--------+--------+--------+--------+ | | PL=2 | RL=1 | xxxx | xxxx | Q | +--------+--------+--------+--------+--------+--------+--------+--------+ |vv000000|10 L=4B | 3 | 0 | 3 | 0 | xxxx | xxxx | +--------+--------+--------+--------+--------+--------+--------+--------+ | | PL=1 | RL=0 | MTU=1,024 (in 8B-words) | +--------+--------+--------+--------+--------+--------+--------+--------+ | 64 zero bits, unless any error was indicated along the path | +--------+--------+--------+--------+--------+--------+--------+--------+ The MTU in the above is the lessor of the MTUs of both networks. The RL (record-length) of the last -record is NOT included in the RL of the preceding -record, but is included in the RL of the preceding -record (since the RL of the is included in the RL of the ). The RL=3 of the includes 2 words of and 1 word of . (M7) Finally, Node1 sends data to Node2 using L2-forwarding. 0 1 2 3 4 5 6 7 +-----------------------------------------------------------------------+ | <---- The L2-header needed to get from Node1 to RouterB1 ----> | | It may be any number of bytes. Here it is 11 bytes: 11230000000 | +--------+--------+--------+--------+--------+--------+--------+--------+ |vv000000|10 L=4B | 3 | 0 | 3 | 0 | xxxx | xxxx | +--------+--------+--------+--------+--------+--------+--------+--------+ |00 P |0 Node2 |Sensor.SubType=? | "Sensor" | +---+----+--------+--------+--------+-+------+--------+--------+--------+ |E=3|PL=0| Data-Length=? (8B-words) |0| RZ |0 Node1 | +---+----+--------+--------+--------+-+------+--------+--------+--------+ | | | <------------------- The sensor data goes here ---------------------> | | | +--------+--------+--------+--------+--------+--------+--------+--------+ | 64 zero bits, unless any error was indicated along the path | +--------+--------+--------+--------+--------+--------+--------+--------+ Cohen et al. Experimental [Page 15] Internet-Draft PacketWay RRP May 1998 E=3 (0b0011) indicates that all the data is 64-bit, Big Endian order. All the messages shown in this appendix start with local L2 routing bytes needed to get across either SAN1 or SAN2 (indicated with "The L2-header needed to get from ... to ...") which are not L2RHs. The difference is that these bytes are in front of the packet, exposed to the local switches, whereas the L2RHs are only exposed to PacketWay- entities. These local L2 routing bytes are the actual bytes required by the SANs and likely to be consumed as the messages traverses the SAN, unlike the L2RHs that are intact until converted to actual routing bytes. The L2RHs start with 0bvv00000010 followed by the number of routing bytes in that L2RH, and possibly also by several bytes of padding. Glossary Address A unique designation of a node (actually an interface to that node) or a SAN. Buddy-HR HRs are "buddies" if they are on the same SAN. Cut-Through See Wormhole. Destination The node to which a packet is intended. Dynamic-Routing Routing according to dynamic information (i.e., acquired at run time, rather than pre-set). Endianness The property of being Big-Endian or Little-Endian (transmission order, etc.) Ethertype A 16-bit value designating the type of Level-3 packets carried by a Level-2 communication sys- tem. HR Half-Router, the part of a router that handles one network only. L2-Forwarding Forwarding based on Level-2 (i.e., data-link layer of the ISORM) information, e.g., the native technique of each SAN or LAN. Also called "source routing." L3-Forwarding Forwarding based on end-to-end (Level-3 i.e., network layer of the ISORM) addresses. Also called "destination routing." Cohen et al. Experimental [Page 16] Internet-Draft PacketWay RRP May 1998 Map The topology of a network. Mapper A node on a SAN/LAN that has the map and an RT for that network. It is expected that the mapper dynamically updates the map and the RT. Multi-homed A node with more than one network interface, where each interface has another address. Node Whatever can send and receive packets (e.g., a computer, an MPP, a software process, etc.) Node A C-struct (or equivalent) containing values for some attributes of a node. Planned Transfer of information, occurs after an initial phase in which the sender decides which Level-2 route to use for that transfer. RCVF The "Received From" set includes all the physical addresses through which an RT was disseminated, starting with the address of the mapper that created that RT. Redirect A message that tells nodes which HR should be used in order to get to a certain remote address. Router The inter-SAN communication device. Security A relationship between 2 (or more) nodes that defines how the nodes utilize security services to communicate securely. Source The node that created a packet. Source-Route A Level-2 route that is chosen for a packet by its source. Symbol Data preceding the EEP header of a PacketWay mes- sage, interleaving with the L2RHs. Twin-HR Two HRs are twins if they both are parts of the same inter-SAN router. Wormhole-routing (aka cut-thru routing) forwarding packets out of switches as soon as possible, without storing that entire packet in the switch (unlike Stop- and-forward) Zero-copy A TCP system that copies data directly between the user area and the network device, bypassing OS copies Cohen et al. Experimental [Page 17] Internet-Draft PacketWay RRP May 1998 Acronyms and Abbreviations 0bNNNN The binary number NNNN (e.g., 0b0100 is 4-decimal) 0xNNNN The hexadecimal number NNNN (e.g., 0x0100 is 256-decimal) 8B 8 byte (64 bits) entity ADDR The Address-record of RRP AT Address Type BER Bit Error Rate CAPA The CAPAbility-record of RRP CSR Common Source-Route DA Destination Address DB Data Block DL Data Length (in 8B words) DT Destination-Type EEP End-to-End Protocol EI Error Indication GVL2 An RRP message, requesting L2 route to a given destination GVRT An RRP message asking an HR to give its routing tables HR Half Router HRTO An RRP message asking which HR to use for a given destination INFO An RRP message providing information about nodes L2 Level-2 of the ISO Reference Model (Link) L2RH Level-2 Routing Header L2SR Source Route L3 Level-3 of the ISO Reference Model (Network) LADR The Logical-addresses-record of RRP Cohen et al. Experimental [Page 18] Internet-Draft PacketWay RRP May 1998 LSbit Least Significant bit LSbyte Least Significant byte MSbit Most Significant bit MSbyte Most Significant byte MTU Maximum Transmission Unit MTUR The MTU-record of RRP NAME The name-record of RRP OH Optional Header field OH-TYPE The Type of an Optional Header field Q Quality (of a path) RCVF Received-From list, or the Received-From record of RRP RDRC A re-direct message of RRP RRP Router-to-Router Protocol RTBL An RRP message proving a Routing Table SRQR The Source-Route-and-Q-record of RRP TELL RRP message requesting INFO about a partially specified node UNK Unknown WRU? An RRP message asking its recipient to identify itself Editor's Address Anthony Skjellum Computer Science Department Box 9637 Mississippi State University Mississippi State, MS 39762-9637 Phone: 601-325-8435 Fax: 601-325-8997 Email: tony@cs.msstate.edu Cohen et al. Experimental [Page 19] --------