PPP Working Group James Carlson Internet Draft IronBridge Networks Expires December 1999 Enrique J. Hernandez-Valencia Lucent Technologies Bell Laboratories Nevin Jones Lucent Technologies Microelectronics Group Paul Langner Lucent Technologies Microelectronics Group June 1999 PPP over Simple Data Link (SDL) using raw lightwave channels with ATM-like framing Status of this Memo This document is an Internet-Draft and is in full conformance with all provisions of Section 10 of RFC2026. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that 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 list Internet-Draft Shadow Directories, see http://www.ietf.org/shadow.html. This document is the product of the Point-to-Point Protocol Extensions Working Group of the Internet Engineering Task Force (IETF). Comments should be submitted to the ietf-ppp@merit.edu mailing list. Distribution of this memo is unlimited. Carlson, et al. expires December 1999 [Page 1] INTERNET DRAFT PPP SDL on Packet-over-Lightwave June 1999 Abstract The Point-to-Point Protocol (PPP) in RFC-1661 [1] provides a standard method for transporting multi-protocol datagrams over point-to-point links, and RFCs 1662 [2] and 1619 [3] provide a means to carry PPP over Synchronous Optical Network (SONET) [5] and Synchronous Digital Hierarchy (SDH) [6] circuits. PPP over Simple Data Link (SDL) using SONET/SDH with ATM-like framing (PPPEXT WG work in Progress) extended these standards to include a new encapsulation for PPP called Simple Data Link (SDL) [8]. This document extends the use of SDL over raw lightwave channels without an intervening SONET/SDH layer, which are also referred to as "dark fiber" or Packet-over-Lightwave" (POL) links, This document is the product of the Point-to-Point Protocol Working Group of the Internet Engineering Task Force (IETF). Comments should be submitted to the ietf-ppp@merit.edu mailing list. Applicability This specification is intended for those implementations which desire to use PPP encapsulation over high speed point-to-point circuits with the so-called "dark fiber" or raw lightwave channels. This enhanced framing mechanisms for PPP encapsulation method has very low overhead, good hardware scaling properties and is resilient to payload expansion. It is anticipated that significantly higher throughput can be attained with SDL when compared to other transport and encapsulation mechanisms for high-speed packet data over lightwave channels, and at a significantly lower cost for line termination equipment. SDL is defined over other media types and for other data link protocols, but this specification covers only the use of PPP over SDL raw lightwave channels. Systems requiring typical public network functions such as transmission quality assessment, protection switching/restoration, and OAM&P at the transmission level will may require either an additional transmission layer (e.g., SONET/SDH or OTN [7]) with or equivalent OAM&P functionality not defined in this document. The use of SDL requires the presentation of packet length information in the SDL header. Thus, hardware implementing SDL must have access to the packet length when generating the header, and where a router's input link does not readily have this information (that is, for non- SDL input links), the router may be required to buffer the entire packet before transmission. "Worm-hole" routing is thus at least problematic with SDL, unless the input links are also SDL. This, however, does not appear to be a great disadvantage on modern routers Carlson, et al. expires December 1999 [Page 2] INTERNET DRAFT PPP SDL on Packet-over-Lightwave June 1999 due to the general requirement of length information in other parts of the system, notably in queueing and congestion control strategies such as Weighted Fair Queuing [12] and Random Early Detection [13]. Carlson, et al. expires December 1999 [Page 3] INTERNET DRAFT PPP SDL on Packet-over-Lightwave June 1999 Table of Contents 1. Introduction ............................................... 5 2. Physical Layer Requirements ................................ 6 2.1. Payload Types ............................................ 6 2.2. Control Signals .......................................... 6 2.3. Synchronization Modes .................................... 6 2.4. Framing .................................................. 6 2.5. Synchronization Procedure ................................ 9 2.6. Scrambler Operation ...................................... 9 2.7. CRC Generation ........................................... 10 2.8. Error Correction ......................................... 10 3. Performance Analysis ....................................... 11 3.1. Mean Time To Frame (MTTF) ................................ 12 3.2. Mean Time To Synchronization (MTTS) ...................... 13 3.3. Probability of False Frame (PFF) ......................... 13 3.4. Probability of False Synchronization (PFS) ............... 13 3.5. Probability of Loss of Frame (PLF) ....................... 14 4. The Special Messages ....................................... 14 4.1. Scrambler State .......................................... 14 4.2. A/B Message .............................................. 14 5. The Set-Reset Scrambler Option ............................. 14 5.1. The Killer Packet Problem ................................ 15 5.2. SDL Set-Reset Scrambler .................................. 15 5.3. SDL Scrambler Synchronization ............................ 15 5.4. SDL Scrambler Operation .................................. 16 6. Configuration Details ...................................... 18 Appendix A: CRC Generation .................................... 19 Appendix B: Error Correction Tables ........................... 21 7. Security Considerations .................................... 23 8. References ................................................. 23 9. Acknowledgments ............................................ 24 10. Intellectual Properties Considerations .................... 24 11. Authors' Addresses ........................................ 25 Carlson, et al. expires December 1999 [Page 4] INTERNET DRAFT PPP SDL on Packet-over-Lightwave June 1999 1. Introduction The term packet-over-lightwave (POL) has been used to refer to the capability of transmitting packet data directly over a raw lightwave channel, also referred to as "dark fiber", without an intervening SONET/SDH or optical transport network (OTN) layer. POL solutions are attractive in data networking scenarios were neither multi- segment/multi-path transport nor the OAM&P capabilities of optical transport networking or SONET/SDH is required. SDL on POL does not rely on SONET/SDH or OTN overheads to enable networking features such as transmission quality assessment, protection switching/restoration, and OAM&P. Performance assessment, switching and OAM&P capabilities for SDL on POL are not defined in this document This document describes a method to enable the use of SDL framing for PPP over such raw lightwave channels and describes the framing and encapsulation requirements for PPP.. The protocol stack is illus- trated in Figure 1. While bit-synchronous HDLC-like framing has a worst-case octet overhead of 20% for some specific data patterns, SDL uses no payload encoding, and hence, has zero payload overhead. +--------------------------+ | | | Higher-layer Protocol | | | +--------------------------+ | | | PPP | | | +--------------------------+ | | | SDL | | | +--------------------------+ | | | Raw Lightwave Channel | | | +--------------------------+ Figure 1: Protocol stack for PPP over SDL over a raw lightwave channel. Carlson, et al. expires December 1999 [Page 5] INTERNET DRAFT PPP SDL on Packet-over-Lightwave June 1999 2. Physical Layer Requirements The transport mode for SDL on POL is packet-oriented. The raw lightwave links are intrinsically bit-synchronous even though PPP treats the lower transport layer as a full-duplex octet-oriented syn- chronous interface. No provision is made to support sending or receiving bare octets over lightwaves (as is the case with SONET/SDH). 2.1. Payload Types Only bit-synchronous payloads at STS-1 and higher line rates are currently considered in this document. Operations at lower bit rates is feasible but not considered at present. Mappings of plesiochronous payloads, such as T1 and T3, on to SDL are not considered in this document. 2.2. Control Signals A prior-arrangement method is required to enable SDL framing for POL. No LCP-negotiated method is currently proposed. LCP may be used to negotiated other PPP-related parameters (see sections 2.4 and 6). 2.3. Synchronization Modes Unlike non-SDL O-S encapsulations, SDL provides a positive indication that it has achieved synchronization with the peer. An SDL PPP implementation MUST provide a means to temporarily suspend PPP data transmission (both user data and negotiation traffic) if synchroniza- tion loss is detected. An SDL PPP implementation SHOULD also provide a configurable timer that is started when SDL is initialized and res- tarted on the loss of synchronization, and is terminated when link synchronization is achieved. If this timer expires, implementation- dependent action should be taken to report the hardware failure. 2.4. Framing PPP over SDL over raw lightwave channels uses the same data link frame format as for PPP over SDL over SONET/SDH [4]. When SDL framing for PPP is employed, the SDL "Datagram Offset" is fixed at 4, and the "A" and "B" messages are not used. Additional information on these optional features of SDL can be found in Lucent's SDL specification [8]. Carlson, et al. expires December 1999 [Page 6] INTERNET DRAFT PPP SDL on Packet-over-Lightwave June 1999 Fixing the Datagram Offset to 4 allows a PPP MRU/MTU of 65536 using SDL. SDL framing is in general accomplished by the use of a four octet header on the packet. This fixed-length header allows the use of a simple framer to detect synchronization as described in section 2.6. For use with PPP, this header precedes each raw PPP packet as fol- lows: 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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Packet Length | Header CRC | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | PPP packet (beginning with address and control fields) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | ..... | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Packet CRC | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ The four octet length header is DC balanced by exclusive-OR (also known as "modulo 2 addition") with the hex value B6AB31E0. This is the maximum transition, minimum sidelobe, Barker-like sequence of length 32. No other scrambling is done on the header itself. Packet Length is an unsigned 16 bit number in network byte order. Unlike the standard PPP FCS, the Header CRC is a CRC-16 generated with initial value zero and transmitted in standard network byte order. The PPP packet is scrambled, and begins with the standard address and control fields, and may be any integral octet length (i.e., it is not padded unless the Self Describing Padding option is used). The Packet CRC is also scrambled, and has a mode-dependent length (described below), and is located only on an octet boundary; no alignment of this field may be assumed. When the Packet Length value is 4 or greater, the distance in octets between one message header and the next in SDL is the sum of Packet Length field, Datagram Offset value, and the fixed size of the Packet CRC field. The Datagram Offset is a configurable SDL parameter, which is set to the fixed value 4 for PPP. When the Packet Length is 0, the distance to the next header is 4 octets. This is the idle fill header. When the Packet Length is 1 to 3, the distance to the next header is 12 octets. These headers are used for special SDL messages used only with optional scrambling and management modes. See section 5 for details of the messages. General SDL, like PPP, allows the use of no CRC, ITU-T CRC-16, or Carlson, et al. expires December 1999 [Page 7] INTERNET DRAFT PPP SDL on Packet-over-Lightwave June 1999 ITU-T CRC-32 for the packet data. However, because the Packet Length field does not include the CRC length, synchronization cannot be maintained if the CRC type is changed per RFC 1570, because frame- to-frame distance is, as described above, calculated including the CRC length. Although synchronization can be regained by readjusting the receiver's frame-to-frame distance after a CRC negotiation, this PPP over SDL specification fixes the CRC type to CRC-32 (four octets), and all SDL implementations MUST reject any LCP FCS Alterna- tives Option [9] requested by the peer when in SDL mode. PPP over SDL implementations MAY allow a configuration option to set different CRC types for use by prior arrangement. Any such configur- able option MUST default to CRC-32, and MUST NOT be include LCP nego- tiation of FCS Alternatives. With the SDL Datagram Offset set to 4, the value placed in the Packet Length field is exactly the length in octets of the PPP frame itself, including the address and control fields but not including the FCS field. Because Packet Lengths below 4 are reserved, the Packet Length MUST be 4 or greater for any legal PPP packet. PPP packets with fewer octets, which are not possible without address/control or protocol field compression, MUST be padded to length 4 for SDL. Inter-packet time fill is accomplished by sending the four octet length header with the Packet Length set to zero. No provision is made for intra-packet time fill. By default, an independent, self-synchronous x^43+1 scrambler is used on the data portion of the message including the 32 bit CRC. This is done in exactly the same manner as with the ATM x^43+1 scrambler on a SONET/SDH link. The scrambler is not clocked when SDL header bits are transmitted. Thus, the data scrambling can be implemented in an entirely independent manner from the SDL framing. Optionally, by prior arrangement, SDL links MAY use a set-reset scrambler as described in section 2.9. If this option is provided, it MUST be configurable by the administrator, and the option MUST default to the self-synchronous scrambler. Once the link enters SYNCH state, the SDL header single bit error correction logic is enabled (see section 2.9). Any unrecoverable header CRC error returns the link to HUNT state, disables PPP transmission, and disables the error correction logic. Carlson, et al. expires December 1999 [Page 8] INTERNET DRAFT PPP SDL on Packet-over-Lightwave June 1999 2.5. Synchronization Procedure The link synchronization procedure is similar to the I.432 section 4.5.1.1 ATM HEC delineation procedure [10], except that the SDL mes- sages are variable length. The machine starts in HUNT state until a four octet sequence in the data stream with a valid CRC-16 is found. (Note that the CRC-16 single-bit error correction technique described in section 2.9 is not employed until the machine is in in SYNCH state. The header must have no bit errors in order to leave HUNT state.) Such a valid sequence is a candidate SDL header. On finding the valid sequence, the machine enters PRESYNCH state. Any one invalid SDL header in PRESYNCH state returns the link to HUNT state. If a second valid SDL header is seen after entering PRESYNCH state, then the link enters SYNCH state and PPP transmission is enabled. If an invalid SDL header is detected, then the link is returned to HUNT state without enabling PPP transmission. 2.6. Scrambler Operation The transmit and receive scramblers are shift registers with 43 stages that MAY be initialized to all-ones when the link is initial- ized. Synchronization is maintained by the data itself. Transmit Receive DATA-STREAM (FROM PPP) IN (FROM SDL FRAMER) | | v | XOR<-------------------------+ +->D0-+->D1-> ... ->D41->D42-+ | | | | +->D0-+->D1-> ... ->D41->D42-+ XOR<-------------------------+ | | v v OUT (TO SDL FRAMER) DATA-STREAM (TO PPP) Each XOR is an exclusive-or gate; also known as a modulo-2 adder. Each Dn block is a D-type flip-flop clocked on the appropriate data clock. The scrambler is clocked once after transmission or reception of each bit of payload and before the next bit is applied as input. Bits within an octet are, per SONET/SDH standard practice, transmitted and received MSB-first. Carlson, et al. expires December 1999 [Page 9] INTERNET DRAFT PPP SDL on Packet-over-Lightwave June 1999 2.7. CRC Generation The CRC-16 and CRC-32 generator polynomials used by SDL are the ITU-T standard polynomials [11]. These are: x^16+x^12+x^5+1 x^32+x^26+x^23+x^22+x^16+x^12+x^11+x^10+x^8+x^7+x^5+x^4+x^2+x+1 The SDL Header CRC and the CRC-16 used for each of the three special messages (scrambler state, message A, and message B; see section 5) are all generated using an initial remainder value of 0000 hex. The optional CRC-16 on the payload data (this mode is not used with PPP over SDL except by prior arrangement) uses the standard initial remainder value of FFFF hex for calculation and the bits are comple- mented before transmission. The final CRC remainder, however, is transmitted in network byte order, unlike the regular PPP FCS. If the CRC-16 algorithm is run over all of the octets including the appended CRC itself, then the remainder value on intact packets will always be E2F0 hex. Alternatively, an implementation may stop CRC calculation before processing the appended CRC itself, and do a direct comparison. The standard CRC-32 on the payload data (used for standard PPP over SDL) uses the initial remainder value of FFFFFFFF hex for calculation and the bits are complemented before transmission. The CRC, however, is transmitted in network byte order, most significant bit first, unlike the optional PPP 32 bit FCS, which is transmitted in reverse order. The remainder value on intact packets when the appended CRC value is included in the calculation is 38FB2284. C code to generate these CRCs is found in Appendix A. 2.8. Error Correction The error correction technique is based on the use of a Galois number field, as with the ATM HEC correction. In a Galois number field, f(a+b) = f(a) + f(b). Since the CRC-16 used for SDL forms such a field, we can state that CRC(message+error) = CRC(message) + CRC(error). Since the CRC-16 remainder of a properly formed message is always zero, this means that, for the N distinct "error" strings corresponding to a single bit error, there are N distinct CRC(error) values, where N is the number of bits in the message. A table look-up is thus applied to the CRC-16 residue after calcula- tion over the four octet SDL header to correct bit errors in the Carlson, et al. expires December 1999 [Page 10] INTERNET DRAFT PPP SDL on Packet-over-Lightwave June 1999 header and to detect multiple bit errors. For the optional set-reset scrambler, a table look-up is similarly applied to the CRC-16 residue after calculation over the eight octet scrambler state message to correct bit errors and to detect multiple bit errors. (This second correction is also used for the special SDL A and B messages, which are not used for standard PPP over SDL.) Note: No error correction is performed for the payload. Note: This error correction technique is used only when the link has entered SYNCH state. While in HUNT or PRESYNCH state, error correc- tion should not be performed, and only messages with syndrome 0000 are accepted. If the calculated syndrome does not appear in this table, then an unrecoverable error has occurred. Any such error in the SDL header will return the link to HUNT state. Since the CRC calculation is started with zero, the two tables can be merged. The four octet table is merely the last 32 entries of the eight octet table. Eight octet (64 bit) single bit error syndrome table (in hex): FD81 F6D0 7B68 3DB4 1EDA 0F6D 8FA6 47D3 ABF9 DDEC 6EF6 377B 93AD C1C6 60E3 B861 D420 6A10 3508 1A84 0D42 06A1 8B40 45A0 22D0 1168 08B4 045A 022D 8906 4483 AA51 DD38 6E9C 374E 1BA7 85C3 CAF1 ED68 76B4 3B5A 1DAD 86C6 4363 A9A1 DCC0 6E60 3730 1B98 0DCC 06E6 0373 89A9 CCC4 6662 3331 9188 48C4 2462 1231 8108 4084 2042 1021 Thus, if the syndrome 6EF6 is seen on an eight octet message, then the third bit (hex 20) of the second octet is in error. Similarly, if 48C4 is seen on an eight octet message, then the second bit (hex 40) in the eighth octet is in error. For a four octet message, the same two syndromes would indicate a multiple bit error for 6EF6, and a single bit error in the second bit of the fourth octet for 48C4. Note that eight octet messages are used only for the optional set- reset scrambling mode, described in section 6. Corresponding C code to generate this table is found in Appendix B. 3. Performance Analysis There are five general statistics that are important for framing algorithms. These are: Carlson, et al. expires December 1999 [Page 11] INTERNET DRAFT PPP SDL on Packet-over-Lightwave June 1999 MTTF Mean time to frame MTTS Mean time to synchronization PFF Probability of false frame PFS Probability of false synchronization PLF Probability of loss of frame The following sections summarize each of these statistics for SDL. Details and mathematic development can be found in the Lucent SDL documentation [8]. 3.1. Mean Time To Frame (MTTF) This metric measures the amount of time required to establish correct framing in the input data. This may be measured in any convenient units, such as seconds or bytes. For SDL, the relevant measurement is in packets, since fragments of packets are not useful. In order to calculate MTTF, we must first determine how often the frame detection state machine is "unavailable" because it failed to detect the next incoming SDL frame within the user data. Since the probability of a false header detection using CRC-16 in random data is 2^-16 and this rate is large compared to the allowable packet size, it is worthwhile to run multiple parallel frame- detection state machines. Each machine starts with a different can- didate framing point in order to reduce the probability of falsely detecting user data as a valid frame header. The results for this calculation for 64KB, 8KB and 384B packets are: Number of Unavailability Unavailability Unavailability Framers 64Kb Packets 8KB packets 384 byte pkts 1 8.75E-1 3.68E-1 2.31E-2 2 7.50E-1 1.04E-1 3.57E-4 3 6.27E-1 2.32E-2 4.17E-6 4 5.08E-1 4.35E-3 3.89E-8 Using these values, MTTF can be calculated as a function of the Bit Error Rate (BER). These plots show a characteristically flat region for all BERs up to a knee, beyond which the begins to rise sharply. In all cases, this knee point has been found to occur at a BER of approximately 1E-4, which is several orders of magnitude above that observed on existing SONET/SDH links. The flat rate values are sum- marized as: Number of Flat Region Flat region Flat region Framers 64KB Packets 8KB packets 384B packets Carlson, et al. expires December 1999 [Page 12] INTERNET DRAFT PPP SDL on Packet-over-Lightwave June 1999 1 8.50 2.08 1.52 2 4.57 1.62 1.50 3 3.18 1.52 1.50 4 2.53 1.50 1.50 Thus, for common packet sizes in an implementation with two parallel framers using links with a BER of 1E-4 or better, the MTTF is approx- imately 1.5 packets. This is also the optimal time, since it represents initiating framing at an average point half-way into one packet, and achieving good framing after seeing exactly one correctly framed packet. Note that the numbers in both tables apply only after the link loses synchronization, which is by itself a very rare event as per the estimate PLF in seccion 3.5. 3.2. Mean Time To Synchronization (MTTS) The MTTS for standard SDL with a self-synchronous scrambler is the same as the MTTF, or 1.5 packets. The MTTS for SDL using the optional set-reset scrambler is one half of the scrambling state transmission interval (in packets) plus the MTTF. For insertion at the default rate of one per eight packets, the MTTS is 5.5 packets. (The probability of receiving a bad scrambling state transmission should also be included in this calculation. The probability of ran- dom corruption of this short message is shown in the SDL document [8] to be small enough that it can be neglected for this calculation.) 3.3. Probability of False Frame (PFF) The PFF is 232.8E-12 (2^-32), since false framing requires two con- secutive headers with falsely correct CRC-16. 3.4. Probability of False Synchronization (PFS) The PFS for the standard self-synchronous scrambler is the same as the PFF, or 232.8E-21 (2^-32). The PFS for the set-reset scrambler is 54.21E-21 (2^-64), and is cal- culated as the PFF above multiplied by the probability of a falsely detected scrambler state message, which itself contains two indepen- dent CRC-16 calculations. Carlson, et al. expires December 1999 [Page 13] INTERNET DRAFT PPP SDL on Packet-over-Lightwave June 1999 3.5. Probability of Loss of Frame (PLF) The PLF is a function of the BER, and for SDL is approximately the square of the BER multiplied by 500, which is the probability of two or more bit errors occurring within the 32 bit SDL header. Thus, at a BER of 1E-5, the PLF is 5E-8. For the typical fiber BER, between 1E- 10 to 1E-12, frame delineation loss would occur less than once every year, on the average, even at OC-768 rates. 4. The Special Messages When the SDL Packet Length field has any value between 0000 and 0003, the message following the header has a special, pre-defined length. The 0 value is a time-fill on an idle link, and no other data fol- lows. The next octet on the link is the first octet of the next SDL header. The values 1 through 3 are defined in the following subsections. These special messages each consist of a six octet data portion fol- lowed by another CRC-16 over that data portion, as with the SDL header, and this CRC is used for single bit error correction. 4.1. Scrambler State The special value of 1 for Packet Length is reserved to transfer the scrambler state from the transmitter to the receiver for the optional set-reset scrambler. In this case, the SDL header is followed by six octets (48 bits) of scrambler state. Neither the scrambler state nor the CRC are scrambled. 4.2. A/B Message The special values of 2 and 3 for Packet Length are reserved for "A" and "B" messages, which are also six octets in length followed by two octets of CRC-16. Each of these eight octets are scrambled. No use for these messages with PPP SDL is defined. These messages are reserved for use by link maintenance protocols, in a manner analogous to ATM's OAM cells. 5. The Set-Reset Scrambler Option Standard PPP over SDL uses a self-synchronous scrambler. SDL imple- mentations MAY also employ a set-reset scrambler to avoid some of the possible inherent problems with self-synchronous scramblers. Carlson, et al. expires December 1999 [Page 14] INTERNET DRAFT PPP SDL on Packet-over-Lightwave June 1999 5.1. The Killer Packet Problem Scrambling in general solves two problems. First, most line inter- faces (e.g., SONET/SDH) require a minimum density of bit transitions in order to maintain hardware clock recovery. Since data streams frequently contain long runs of all zeros or all ones, scrambling the bits using a pseudo-random number sequence breaks up these patters. Second, all link-layer synchronization mechanisms rely on detecting long-range patterns in the received data to detect framing. Self-synchronous scramblers are an easy way to partially avoid these problems. One problem that is inherent with self-synchronous, how- ever, is that long user packets from malicious sites can make use of the known properties of these scramblers to inject either long strings of zeros or other synchronization-destroying patterns into the link. Such carefully constructed packets are called "killer packets." 5.2. SDL Set-Reset Scrambler An alternative to the self-synchronous scrambler is the externally synchronized or "set-reset" scrambler. This is a free-running scram- bler that is not affected by the patterns in the user data, and therefore minimizes the possibility that a malicious user could present data to the network that mimics an undesirable data pattern. The option set-reset scrambler defined for SDL is an x^48+x^28+x^27+x+1 independent scrambler initialized to all ones when the link enters PRESYNCH state and reinitialized if the value ever becomes all zero bits. As with the self-synchronous scrambler, all octets in the PPP packet data following the SDL header through the final packet CRC are scrambled. 5.3. SDL Scrambler Synchronization As described in the previous section, the special value of 1 for Packet Length is reserved to transfer the scrambler state from the transmitter to the receiver. In this case, the SDL header is fol- lowed by six octets (48 bits) of scrambler state plus two octets of CRC-16 over the scrambler state. None of these eight octets are scrambled. SDL synchronization consists of two components, link and scrambler synchronization. Both must be completed before PPP data flows on the link. Carlson, et al. expires December 1999 [Page 15] INTERNET DRAFT PPP SDL on Packet-over-Lightwave June 1999 If a valid SDL header is seen in PRESYNCH state, then the link enters SYNCH state, and the scrambler synchronization sequence is started. If an invalid SDL header is detected, then the link is returned to HUNT state, and PPP transmission is suspended. When scrambler synchronization is started, a scrambler state message is sent (Packet Length set to 1 and six octets of scrambler state in network byte order follow the SDL header). This message is sent once. At this point, PPP transmission is enabled. Scrambler state messages are periodically transmitted to keep the peers in synchronization. A period of once per eight transmitted packets is suggested, and it SHOULD be configurable. Excessive packet CRC errors detected indicates an extended loss of synchroniza- tion and should trigger link resynchronization. On reception of a scrambler state message, an SDL implementation MUST compare the received 48 bits of state with the receiver's scrambler state. If any of these bits differ, then a synchronization slip error is declared. After such an error, the next valid scrambler state message received MUST be loaded into the receiver's scrambler, and the error condition is then cleared. 5.4. SDL Scrambler Operation The transmit and receive scramblers are shift registers with 48 stages that are initialized to all-ones when the link is initialized. Each is refilled with all one bits if the value in the shift register ever becomes all zeros. This scrambler is not reset at the beginning of each frame, as is the SONET/SDH X^7+X^6+1 scrambler, nor is it modified by the transmitted data, as is the ATM self-synchronous scrambler. Instead it is kept in synchronization using special SDL messages. +----XOR<--------------XOR<---XOR<----------------+ | ^ ^ ^ | | | | | | +->D0-+->D1-> ... ->D26-+->D27-+->D28-> ... ->D47-+ | v OUT Each XOR is an exclusive-or gate; also known as a modulo-2 adder. Each Dn block is a D-type flip-flop clocked on the appropriate data clock. The scrambler is clocked once after transmission of each bit of SDL Carlson, et al. expires December 1999 [Page 16] INTERNET DRAFT PPP SDL on Packet-over-Lightwave June 1999 data, whether or not the transmitted bit is scrambled. When scram- bling is enabled for a given octet, the OUT bit is exclusive-ored with the raw data bit to produce the transmitted bit. Bits within an octet are transmitted MSB-first. Reception of scrambled data is identical to transmission. Each received bit is exclusive-ored with the output of the separate receive data scrambler. To generate a scrambler state message, the contents of D47 through D0 are snapshot at the point where the first scrambler state bit is sent. D47 is transmitted as the first bit of the output. The first octet transmitted contains D47 through D40, the second octet D39 through D32, and the sixth octet D7 through D0. The receiver of a scrambler state message MUST first run the CRC-16 check and correct algorithm over this message. If the CRC-16 message check detects multiple bit errors, then the message is dropped and is not processed further. Otherwise, it then should compare the contents of the entire receive scrambler state D47:D0 with the corrected message. (By pipelining the receiver with multiple clock stages between SDL Header error- correction block and the descrambling block, the receive descrambler will be on the correct clock boundary when the message arrives at the descrambler. This means that the decoded scrambler state can be treated as immediately available at the beginning of the D47 clock cycle into the receive scrambler.) If any of the received scrambler state bits is different from the corresponding shift register bit, then a soft error flag is set. If the flag was already set when this occurs, then a synchronization slip error is declared. This error SHOULD be counted and reported through implementation-defined network management procedures. When the receiver has this soft error flag set, any scrambler state mes- sage that passes the CRC-16 message check without multiple bit errors is clocked directly into the receiver's state register after the com- parison is done, and the soft error flag is then cleared. Otherwise, while uncorrectable scrambler state messages are received, the soft error flag state is maintained. (The intent of this mechanism is to reduce the likelihood that a falsely corrected scrambler state message with multiple bit errors can corrupt the running scrambler state.) Carlson, et al. expires December 1999 [Page 17] INTERNET DRAFT PPP SDL on Packet-over-Lightwave June 1999 6. Configuration Details The following PPP Configuration Options are recommended: Magic Number No Address and Control Field Compression No Protocol Field Compression No FCS alternatives (32-bit FCS default) This configuration means that standard PPP over SDL on POL generally presents a 32-bit aligned datagram to the network layer. With the address, control, and protocol field intact, the PPP overhead on each packet is four octets. If the SDL framer presents the SDL packet header to the PPP input handling in order to communicate the packet length , this header is also four octets, and word-alignment is preserved. Since SDL does take the place of HDLC as a transport for PPP, it is at least tempting to remove the HDLC-derived overhead. This is not done for standard PPP over SDL in order to preserve the message alignment and the future possibility of Frame Relay internetworking. By prior external arrangement, any two SDL implementations MAY omit the address and control fields or implement protocol field compres- sion. Such use is not standardized and MUST NOT be the default on any SDL implementation. Carlson, et al. expires December 1999 [Page 18] INTERNET DRAFT PPP SDL on Packet-over-Lightwave June 1999 Appendix A: CRC Generation The following unoptimized code generates proper CRC-16 and CRC-32 values for SDL messages. Note that the polynomial bits are numbered in big-endian order for SDL CRCs; bit 0 is the MSB. typedef unsigned char u8; typedef unsigned short u16; typedef unsigned long u32; #define POLY16 0x1021 #define POLY32 0x04C11DB7 u16 crc16(u16 crcval, u8 cval) { int i; crcval ^= cval << 8; for (i = 8; i--; ) crcval = crcval & 0x8000 ? (crcval << 1) ^ POLY16 : crcval << 1; return crcval; } u32 crc32(u32 crcval, u8 cval) { int i; crcval ^= cval << 24; for (i = 8; i--; ) crcval = crcval & 0x80000000 ? (crcval << 1) ^ POLY32 : crcval << 1; return crcval; } u16 crc16_special(u8 *buffer, int len) { u16 crc; crc = 0; while (--len >= 0) crc = crc16(crc,*buffer++); return crc; } Carlson, et al. expires December 1999 [Page 19] INTERNET DRAFT PPP SDL on Packet-over-Lightwave June 1999 u16 crc16_payload(u8 *buffer, int len) { u16 crc; crc = 0xFFFF; while (--len >= 0) crc = crc16(crc,*buffer++); return crc ^ 0xFFFF; } u32 crc32_payload(u8 *buffer, int len) { u32 crc; crc = 0xFFFFFFFFul; while (--len >= 0) crc = crc32(crc,*buffer++); return crc ^ 0xFFFFFFFFul; } void make_sdl_header(int packet_length, u8 *buffer) { u16 crc; buffer[0] = (packet_length >> 8) & 0xFF; buffer[1] = packet_length & 0xFF; crc = crc16_special(buffer,2); buffer[0] ^= 0xB6; buffer[1] ^= 0xAB; buffer[2] = ((crc >> 8) & 0xFF) ^ 0x31; buffer[3] = (crc & 0xFF) ^ 0xE0; } Carlson, et al. expires December 1999 [Page 20] INTERNET DRAFT PPP SDL on Packet-over-Lightwave June 1999 Appendix B: Error Correction Tables To generate the error correction table, the following implementation may be used. It creates a table called sdl_error_position, which is indexed on CRC residue value. The tables can be used to determine if no error exists (table entry is equal to FE hex), one correctable error exists (table entry is zero-based index to errored bit with MSB of first octet being 0), or more than one error exists, and error is uncorrectable (table entry is FF hex). To use for eight octet mes- sages, the bit index from this table is used directly. To use for four octet messages, the index is treated as an unrecoverable error if it is below 32, and as bit index plus 32 if it is above 32. The program also prints out the error syndrome table shown in section 2.9. This may be used as part of a "switch" statement in a hardware implementation. u8 sdl_error_position[65536]; /* Calculate new CRC from old^(byte<<8) */ u16 crc16_t8(u16 crcval) { u16 f1,f2,f3; f1 = (crcval>>8) | (crcval<<8); f2 = (crcval>>12) | (crcval&0xF000) | ((crcval>>7)&0x01E0); f3 = ((crcval>>3) & 0x1FE0) ^ ((crcval<<4) & 0xF000); return f1^f2^f3; } void generate_error_table(u8 *bptab, int nbytes) { u16 crc; int i, j, k; /* Marker for no error */ bptab[0] = 0xFE; /* Marker for >1 error */ for (i = 1; i < 65536; i++ ) bptab[i] = 0xFF; /* Mark all single bit error cases. */ printf("Error syndrome table:\n"); for (i = 0; i < nbytes; i++) { putchar(' '); Carlson, et al. expires December 1999 [Page 21] INTERNET DRAFT PPP SDL on Packet-over-Lightwave June 1999 for (j = 0; j < 8; j++) { crc = 0; for (k = 0; k < i; k++) crc = crc16_t8(crc); crc = crc16_t8(crc ^ (0x8000>>j)); for (k++; k < nbytes; k++) crc = crc16_t8(crc); bptab[crc] = (i * 8) + j; printf(" %04X",crc); } putchar('\n'); } } int main(int argc, char **argv) { u8 buffer[8] = { 0x01,0x55,0x02,0xaa, 0x99,0x72,0x18,0x56 }; u16 crc; int i; generate_error_table(sdl_error_position,8); /* Run sample message through check routine. */ crc = 0; for (i = 0; i < 8; i++) crc = crc16_t8(crc ^ (buffer[i]<<8)); /* Output is 0000 64 -- no error encountered. */ printf("\nError test: CRC %04X, bit position %d\n", crc,sdl_message_error_position[crc]); } Carlson, et al. expires December 1999 [Page 22] INTERNET DRAFT PPP SDL on Packet-over-Lightwave June 1999 7. Security Considerations The reliability of communication networks places special requirements in the handling of data payloads as appropriate to the specific line encoding schemes. This document describes framing and scrambling options for SDL over raw lightwave channels that enable the use of current typical design (non burst mode) optical transceiver and tim- ing subsystem. In particular, this proposal is compatible with DWDM regenerator networks. No other security concerns have been identi- fied. 8. References [1] Simpson, W., Editor, "The Point-to-Point Protocol (PPP)," RFC 1661, Daydreamer, July 1994. [2] Simpson, W., Editor, "PPP in HDLC-like Framing," RFC 1662, Daydreamer, July 1994. [3] Simpson, W., Editor, "PPP over SONET/SDH," RFC 1619, Daydreamer, May 1994. [4] Carlson, Langner, Hernandez-Valencia, Manchester, "PPP over Simple Data Link (SDL) using SONET/SDH with ATM-like framing." PPPEXT WG work in progress. [5] "American National Standard for Telecommunications - Synchronous Optical Network (SONET) Payload Mappings," ANSI T1.105.02-1995. [6] ITU-T Recommendation G.707, "Network Node Interface for the Synchronous Digital Hierarchy (SDH)," March 1996. [7] ITU-T Recommendation G.872, "Architecture of Optical Transport Networks," February 1999. [8] Doshi,B., Dravida, S., Hernandez-Valencia, E., Matragi, W., Qureshi, M., Anderson, J., Manchester, J.,"A Simple Data Link Protocol for High Speed Packet Networks", Bell Labs Technical Journal, pp. 85-104, Vol.4 No.1, January-March 1999. [9] Simpson, W., Editor, "PPP LCP Extensions," RFC 1570, Daydreamer, January 1994. [10] ITU-T Recommendation I.432.1, "B-ISDN User-Network Interface - Physical Layer Specification: General Characteristics," Carlson, et al. expires December 1999 [Page 23] INTERNET DRAFT PPP SDL on Packet-over-Lightwave June 1999 February 1999. [11] ITU-T Recommendation V.41, "Code-independent error-control system," November 1989. [12] Demers, A., S. Keshav, and S. Shenker, "Analysis and simulation of a fair queueing algorithm," ACM SIGCOMM volume 19 number 4, pp. 1-12, September 1989. [13] Floyd, S. and V. Jacobson, "Random Early Detection Gateways for Congestion Avoidance," IEEE/ACM Transactions on Networking, August 1993. 9. Acknowledgments The authors recognize Jon Anderson, Bharat Doshi, Subra Dravida and James Manchester from Lucent Technologies for their various contribu- tions to this work. 10. Intellectual Properties Considerations Lucent Technologies Inc. may own intellectual property on some of the technologies disclosed in this document. The patent licensing policy of Lucent Technologies Inc. with respect to any patents or patent applications relating to this submission is stated in the March 1, 1999, letter to the IETF from Dr. Roger E. Stricker, Intellectual Property Vice President, Lucent Technologies Inc. This letter is on file in the offices of the IETF Secretariat. IronBridge Networks has no claim on any of this material. Carlson, et al. expires December 1999 [Page 24] INTERNET DRAFT PPP SDL on Packet-over-Lightwave June 1999 11. Authors' Addresses James Carlson IronBridge Networks 55 Hayden Avenue Lexington MA 02421-7996 Phone: +1 781 372 8132 Fax: +1 781 372 8090 Email: carlson@ibnets.com Enrique J. Hernandez-Valencia Lucent Technologies Bell Laboratories 101 Crawford Corners Rd. Holmdel NJ 07733-3030 Email: enrique@lucent.com Nevin Jones Lucent Technologies Microelectronics Group 555 Union Boulevard Allentown PA 18103-1286 Email: nrjones@lucent.com Paul Langner Lucent Technologies Microelectronics Group 555 Union Boulevard Allentown PA 18103-1286 Email: plangner@lucent.com Carlson, et al. expires December 1999 [Page 25]