idnits 2.17.1 draft-ietf-ipngwg-bsd-api-05.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** Cannot find the required boilerplate sections (Copyright, IPR, etc.) in this document. Expected boilerplate is as follows today (2024-04-20) according to https://trustee.ietf.org/license-info : IETF Trust Legal Provisions of 28-dec-2009, Section 6.a: This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79. IETF Trust Legal Provisions of 28-dec-2009, Section 6.b(i), paragraph 2: Copyright (c) 2024 IETF Trust and the persons identified as the document authors. All rights reserved. IETF Trust Legal Provisions of 28-dec-2009, Section 6.b(i), paragraph 3: This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- ** Missing expiration date. The document expiration date should appear on the first and last page. ** The document seems to lack a 1id_guidelines paragraph about Internet-Drafts being working documents. ** The document seems to lack a 1id_guidelines paragraph about 6 months document validity. ** The document seems to lack a 1id_guidelines paragraph about the list of current Internet-Drafts. ** The document seems to lack a 1id_guidelines paragraph about the list of Shadow Directories. ** The document is more than 15 pages and seems to lack a Table of Contents. == No 'Intended status' indicated for this document; assuming Proposed Standard Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- ** The document seems to lack an IANA Considerations section. (See Section 2.2 of https://www.ietf.org/id-info/checklist for how to handle the case when there are no actions for IANA.) ** The document seems to lack separate sections for Informative/Normative References. All references will be assumed normative when checking for downward references. ** There are 11 instances of too long lines in the document, the longest one being 5 characters in excess of 72. ** The abstract seems to contain references ([1]), which it shouldn't. Please replace those with straight textual mentions of the documents in question. == There are 1 instance of lines with non-RFC2606-compliant FQDNs in the document. == There are 1 instance of lines with non-RFC6890-compliant IPv4 addresses in the document. If these are example addresses, they should be changed. Miscellaneous warnings: ---------------------------------------------------------------------------- == Line 187 has weird spacing: '... u_char s6_ad...' == Line 695 has weird spacing: '...imit to use ...' == Line 696 has weird spacing: '... (Note a sep...' == Line 703 has weird spacing: '...packets sent ...' == Line 704 has weird spacing: '...livered back ...' == (4 more instances...) -- The document seems to lack a disclaimer for pre-RFC5378 work, but may have content which was first submitted before 10 November 2008. If you have contacted all the original authors and they are all willing to grant the BCP78 rights to the IETF Trust, then this is fine, and you can ignore this comment. If not, you may need to add the pre-RFC5378 disclaimer. (See the Legal Provisions document at https://trustee.ietf.org/license-info for more information.) -- The document date (April 18, 1996) is 10229 days in the past. Is this intentional? -- Found something which looks like a code comment -- if you have code sections in the document, please surround them with '' and '' lines. Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Missing Reference: '16' is mentioned on line 187, but not defined ** Obsolete normative reference: RFC 1883 (ref. '1') (Obsoleted by RFC 2460) ** Obsolete normative reference: RFC 1884 (ref. '2') (Obsoleted by RFC 2373) -- Possible downref: Non-RFC (?) normative reference: ref. '3' -- Possible downref: Non-RFC (?) normative reference: ref. '4' ** Obsolete normative reference: RFC 1933 (ref. '5') (Obsoleted by RFC 2893) Summary: 14 errors (**), 0 flaws (~~), 10 warnings (==), 5 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 1 Internet Engineering Task Force R. E. Gilligan (Sun) 2 INTERNET-DRAFT S. Thomson (Bellcore) 3 J. Bound (Digital) 4 April 18, 1996 6 Basic Socket Interface Extensions for IPv6 7 9 Abstract 11 In order to implement the version 6 Internet Protocol (IPv6) [1] in 12 an operating system based on Berkeley Unix (4.x BSD), changes must be 13 made to the application program interface (API). TCP/IP applications 14 written for BSD-based operating systems have in the past enjoyed a 15 high degree of portability because most of the systems derived from 16 BSD provide the same API, known informally as "the socket interface". 17 We would like the same portability with IPv6. This memo presents a 18 basic set of extensions to the BSD socket API to support IPv6. The 19 changes include a new data structure to carry IPv6 addresses, new 20 address conversion functions, and some new setsockopt() options. The 21 extensions are designed to provide access to IPv6 features, while 22 introducing a minimum of change into the system and providing 23 complete compatibility for existing IPv4 applications. Additional 24 extensions for new IPv6 features may be added at a later time. 26 Status of this Memo 28 This document is an Internet Draft. Internet Drafts are working 29 documents of the Internet Engineering Task Force (IETF), its Areas, 30 and its Working Groups. Note that other groups may also distribute 31 working documents as Internet Drafts. 33 Internet Drafts are draft documents valid for a maximum of six 34 months. This Internet Draft expires on October 18, 1996. Internet 35 Drafts may be updated, replaced, or obsoleted by other documents at 36 any time. It is not appropriate to use Internet Drafts as reference 37 material or to cite them other than as a "working draft" or "work in 38 progress." 40 To learn the current status of any Internet-Draft, please check the 41 1id-abstracts.txt listing contained in the Internet-Drafts Shadow 42 Directories on ds.internic.net, nic.nordu.net, ftp.isi.edu, or 43 munnari.oz.au. 45 Distribution of this memo is unlimited. 47 1. Introduction 48 While IPv4 addresses are 32-bits long, IPv6 nodes are identified by 49 128-bit addresses. The socket interface make the size of an IP 50 address quite visible to an application; virtually all TCP/IP 51 applications for BSD-based systems have knowledge of the size of an 52 IP address. Those parts of the API that expose the addresses need to 53 be extended to accommodate the larger IPv6 address size. IPv6 also 54 introduces new features, some of which must be made visible to 55 applications via the API. This paper defines a set of extensions to 56 the socket interface to support the larger address size and new 57 features of IPv6. 59 This specification is preliminary. These API extensions are expected 60 to evolve as we gain more implementation experience. 62 2. Design Considerations 64 There are a number of important considerations in designing changes 65 to this well-worn API: 67 - The extended API should provide both source and binary 68 compatibility for programs written to the original API. That is, 69 existing program binaries should continue to operate when run on 70 a system supporting the new API. In addition, existing 71 applications that are re-compiled and run on a system supporting 72 the new API should continue to operate. Simply put, the API 73 changes for IPv6 should not break existing programs. 75 - The changes to the API should be as small as possible in order to 76 simplify the task of converting existing IPv4 applications to 77 IPv6. 79 - Where possible, applications should be able to use the extended 80 API to interoperate with both IPv6 and IPv4 hosts. Applications 81 should not need to know which type of host they are communicating 82 with. 84 - IPv6 addresses carried in data structures should be 64-bit 85 aligned. This is necessary in order to obtain optimum 86 performance on 64-bit machine architectures. 88 Because of the importance of providing IPv4 compatibility in the API, 89 these extensions are explicitly designed to operate on machines that 90 provide complete support for both IPv4 and IPv6. A subset of this 91 API could probably be designed for operation on systems that support 92 only IPv6. However, this is not addressed in this document. 94 2.1. What Needs to be Changed 95 The socket interface API consists of a few distinct components: 97 - Core socket functions. 99 - Address data structures. 101 - Name-to-address translation functions. 103 - Address conversion functions. 105 The core socket functions -- those functions that deal with such 106 things as setting up and tearing down TCP connections, and sending 107 and receiving UDP packets -- were designed to be transport 108 independent. Where protocol addresses are passed as function 109 arguments, they are carried via opaque pointers. A protocol specific 110 address data structure is defined for each protocol that the socket 111 functions support. Applications must cast these protocol specific 112 address structures into the generic "sockaddr" data type when using 113 the socket functions. These functions need not change for IPv6, but 114 a new IPv6 specific address data structure is needed. 116 The "sockaddr_in" structure is the protocol specific data structure 117 for IPv4. This data structure actually includes 8-octets of unused 118 space, and it is tempting to try to use this space to adapt the 119 sockaddr_in structure to IPv6. Unfortunately, the sockaddr_in 120 structure is not large enough to hold the 16-octet IPv6 address as 121 well as the other information (2-octet address family and 2-octet 122 port number) that is needed. So a new address data structure must be 123 defined for IPv6. 125 The name-to-address translation functions in the socket interface are 126 gethostbyname() and gethostbyaddr(). Gethostbyname() does not 127 provide enough flexibility to accommodate protocols other than IPv4. 128 POSIX, in its 1003.g draft specification, has proposed a new hostname 129 to address translation function which is protocol independent. This 130 function can be used with IPv6, so no new function is defined here. 132 The address conversion functions -- inet_ntoa() and inet_addr() -- 133 convert IPv4 addresses between binary and printable form. These 134 functions are quite specific to 32-bit IPv4 addresses. We have 135 designed two analogous functions which convert both IPv4 and IPv6 136 addresses, and carry an address type parameter so that they can be 137 extended to other protocol families as well. 139 Finally, a few miscellaneous features are needed to support IPv6. A 140 new interface is needed in order to support the IPv6 flow label and 141 priority header fields. New interfaces are needed in order to 142 receive IPv6 multicast packets and control the sending of multicast 143 packets. 145 The socket interface may be further extended in the future to provide 146 access to other IPv6 features. These extensions will be made in 147 separate documents. 149 3. Socket Interface 151 This section specifies the socket interface changes for IPv6. 153 The data types of the structure elements given in the following 154 section are intended to be examples, not absolute requirements. 155 System implementations may use other types if they are appropriate. 156 In some cases, such as when a field of a data structure holds a 157 protocol value, the structure field must be of some minimum size. 158 These size requirements are noted in the text. For example, since 159 the UDP and TCP port values are 16-bit quantities, the sin6_port 160 field must be at least a 16-bit data types. The sin6_port field is 161 specified as a u_int16m_t type, but an implementation may use any 162 data type that is at least 16-bits long. 164 3.1. New Address Family 166 A new address family macro, named AF_INET6, is defined in 167 . The AF_INET6 definition is used to distinguish 168 between the original sockaddr_in address data structure, and the new 169 sockaddr_in6 data structure. 171 A new protocol family macro, named PF_INET6, is defined in 172 . Like most of the other protocol family macros, this 173 will usually be defined to have the same value as the corresponding 174 address family macro: 176 #define PF_INET6 AF_INET6 178 The PF_INET6 is used in the first argument to the socket() function 179 to indicate that an IPv6 socket is being created. 181 3.2. IPv6 Address Data Structure 183 A new data structure to hold a single IPv6 address is defined as 184 follows: 186 struct in6_addr { 187 u_char s6_addr[16]; /* IPv6 address */ 188 } 190 This data structure contains an array of sixteen 8-bit elements, 191 which make up one 128-bit IPv6 address. The IPv6 address is stored 192 in network byte order. 194 Applications obtain the declaration for this structure by including 195 the system header file . 197 3.3. Socket Address Structure for 4.3 BSD-Based Systems 199 In the socket interface, a different protocol-specific data structure 200 is defined to carry the addresses for each of the protocol suite. 201 Each protocol-specific data structure is designed so it can be cast 202 into a protocol-independent data structure -- the "sockaddr" 203 structure. Each has a "family" field which overlays the "sa_family" 204 of the sockaddr data structure. This field can be used to identify 205 the type of the data structure. 207 The sockaddr_in structure is the protocol-specific address data 208 structure for IPv4. It is used to pass addresses between 209 applications and the system in the socket functions. The following 210 structure is defined to carry IPv6 addresses: 212 struct sockaddr_in6 { 213 u_int16m_t sin6_family; /* AF_INET6 */ 214 u_int16m_t sin6_port; /* Transport layer port # */ 215 u_int32m_t sin6_flowinfo; /* IPv6 flow information */ 216 struct in6_addr sin6_addr; /* IPv6 address */ 217 }; 219 This structure is designed to be compatible with the sockaddr data 220 structure used in the 4.3 BSD release. 222 The sin6_family field is used to identify this as a sockaddr_in6 223 structure. This field is designed to overlay the sa_family field 224 when the buffer is cast to a sockaddr data structure. The value of 225 this field must be AF_INET6. 227 The sin6_port field is used to store the 16-bit UDP or TCP port 228 number. This field is used in the same way as the sin_port field of 229 the sockaddr_in structure. The port number is stored in network byte 230 order. 232 The sin6_flowinfo field is a 32-bit field that is used to store two 233 pieces of information: the 24-bit IPv6 flow label and the 4-bit 234 priority field. The IPv6 flow label is represented as the low-order 235 24-bits of the 32-bit field. The priority is represented in the next 236 4-bits above this. The high-order 4 bits of this field are reserved. 237 The sin6_flowinfo field is stored in network byte order. The use of 238 the flow label and priority fields are explained in sec 4.9. 240 The sin6_addr field is a single in6_addr structure (defined in the 241 previous section). This field holds one 128-bit IPv6 address. The 242 address is stored in network byte order. 244 The ordering of elements in this structure is specifically designed 245 so that the sin6_addr field will be aligned on a 64-bit boundary. 246 This is done for optimum performance on 64-bit architectures. 248 Applications obtain the declaration of the sockaddr_in6 structure by 249 including the system header file . 251 3.4. Socket Address Structure for 4.4 BSD-Based Systems 253 The 4.4 BSD release includes a small, but incompatible change to the 254 socket interface. The "sa_family" field of the sockaddr data 255 structure was changed from a 16-bit value to an 8-bit value, and the 256 space saved used to hold a length field, named "sa_len". The 257 sockaddr_in6 data structure given in the previous section can not be 258 correctly cast into the newer sockaddr data structure. For this 259 reason, following alternative IPv6 address data structure is provided 260 to be used on systems based on 4.4 BSD: 262 #define SIN6_LEN 264 struct sockaddr_in6 { 265 u_char sin6_len; /* length of this struct */ 266 u_char sin6_family; /* AF_INET6 */ 267 u_int16m_t sin6_port; /* Transport layer port # */ 268 u_int32m_t sin6_flowinfo; /* IPv6 flow information */ 269 struct in6_addr sin6_addr; /* IPv6 address */ 270 }; 272 The only differences between this data structure and the 4.3 BSD 273 variant are the inclusion of the length field, and the change of the 274 family field to a 8-bit data type. The definitions of all the other 275 fields are identical to the 4.3 BSD variant defined in the previous 276 section. 278 Systems that provide this version of the sockaddr_in6 data structure 279 must also declare the SIN6_LEN as a result of including the 280 header file. This macro allows applications to 281 determine whether they are being built on a system that supports the 282 4.3 BSD or 4.4 BSD variants of the data structure. Applications can 283 be written to run on both systems by simply making their assignments 284 and use of the sin6_len field conditional on the SIN6_LEN field. For 285 example, to fill in an IPv6 address structure in an application, one 286 might write: 288 struct sockaddr_in6 sin6; 290 bzero((char *) &sin6, sizeof(struct sockaddr_in6)); 291 #ifdef SIN6_LEN 292 sin6.sin6_len = sizeof(struct sockaddr_in6); 293 #endif 294 sin6.sin6_family = AF_INET6; 295 sin6.sin6_port = htons(23); 297 Note that the size of the sockaddr_in6 structure is larger than the 298 size of the sockaddr structure. Applications that use the 299 sockaddr_in6 structure need to be aware that they can not use 300 sizeof(sockaddr) to allocate a buffer to hold a sockaddr_in6 301 structure. They should use sizeof(sockaddr_in6) instead. 303 3.5. The Socket Functions 305 Applications use the socket() function to create a socket descriptor 306 that represents a communication endpoint. The arguments to the 307 socket() function tell the system which protocol to use, and what 308 format address structure will be used in subsequent functions. For 309 example, to create an IPv4/TCP socket, applications make the call: 311 s = socket (PF_INET, SOCK_STREAM, 0); 313 To create an IPv4/UDP socket, applications make the call: 315 s = socket (PF_INET, SOCK_DGRAM, 0); 317 Applications may create IPv6/TCP and IPv6/UDP sockets by simply using 318 the constant PF_INET6 instead of PF_INET in the first argument. For 319 example, to create an IPv6/TCP socket, applications make the call: 321 s = socket (PF_INET6, SOCK_STREAM, 0); 323 To create an IPv6/UDP socket, applications make the call: 325 s = socket (PF_INET6, SOCK_DGRAM, 0); 327 Once the application has created a PF_INET6 socket, it must use the 328 sockaddr_in6 address structure when passing addresses in to the 329 system. The functions which the application uses to pass addresses 330 into the system are: 332 bind() 333 connect() 334 sendmsg() 335 sendto() 337 The system will use the sockaddr_in6 address structure to return 338 addresses to applications that are using PF_INET6 sockets. The 339 functions that return an address from the system to an application 340 are: 342 accept() 343 recvfrom() 344 recvmsg() 345 getpeername() 346 getsockname() 348 No changes to the syntax of the socket functions are needed to 349 support IPv6, since the all of the "address carrying" functions use 350 an opaque address pointer, and carry an address length as a function 351 argument. 353 3.6. Compatibility with IPv4 Applications 355 In order to support the large base of applications using the original 356 API, system implementations must provide complete source and binary 357 compatibility with the original API. This means that systems must 358 continue to support PF_INET sockets and the sockaddr_in addresses 359 structure. Applications must be able to create IPv4/TCP and IPv4/UDP 360 sockets using the PF_INET constant in the socket() function, as 361 described in the previous section. Applications should be able to 362 hold a combination of IPv4/TCP, IPv4/UDP, IPv6/TCP and IPv6/UDP 363 sockets simultaneously within the same process. 365 Applications using the original API should continue to operate as 366 they did on systems supporting only IPv4. That is, they should 367 continue to interoperate with IPv4 nodes. It is not clear, though, 368 how, or even if, those IPv4 applications should interoperate with 369 IPv6 nodes. The open issues section (section 9) discusses some of 370 the alternatives. 372 3.7. Compatibility with IPv4 Nodes 374 The API also provides a different type of compatibility: the ability 375 for applications using the extended API to interoperate with IPv4 376 nodes. This feature uses the IPv4-mapped IPv6 address format defined 377 in the IPv6 addressing architecture specification [2]. This address 378 format allows the IPv4 address of an IPv4 node to be represented as 379 an IPv6 address. The IPv4 address is encoded into the low-order 32- 380 bits of the IPv6 address, and the high-order 96-bits hold the fixed 381 prefix 0:0:0:0:0:FFFF. IPv4-mapped addresses are written as follows: 383 ::FFFF: 385 Applications may use PF_INET6 sockets to open TCP connections to IPv4 386 nodes, or send UDP packets to IPv4 nodes, by simply encoding the 387 destination's IPv4 address as an IPv4-mapped IPv6 address, and 388 passing that address, within a sockaddr_in6 structure, in the 389 connect() or sendto() call. When applications use PF_INET6 sockets 390 to accept TCP connections from IPv4 nodes, or receive UDP packets 391 from IPv4 nodes, the system returns the peer's address to the 392 application in the accept(), recvfrom(), or getpeername() call using 393 a sockaddr_in6 structure encoded this way. 395 Few applications will likely need to know which type of node they are 396 interoperating with. However, for those applications that do need to 397 know, the inet6_isipv4addr() function, defined in section 6.3, is 398 provided. 400 3.8. Flow Information 402 The IPv6 header has a 24-bit field to hold a "flow label", and a 4- 403 bit field to hold a "priority" value. Applications have control over 404 what values for these fields are used in packets that they originate, 405 and have access to the field values of packets that they receive. 407 The sin6_flowinfo field of the sockaddr_in6 structure encodes two 408 pieces of information: IPv6 flow label and IPv6 priority. 409 Applications use this field to set the flow label and priority in 410 IPv6 headers of packets they generate, and to retrieve the flow label 411 and priority from the packets they receive. The header fields of an 412 actively opened TCP connection are set by assigning in the 413 sin6_flowinfo field of the destination address sockaddr_in6 structure 414 passed in the connect() function. The same technique can be used 415 with the sockaddr_in6 structure passed in to the sendto() or 416 sendmsg() function to set the flow label and priority fields of UDP 417 packets. Similarly, the flow label and priority values of received 418 UDP packets and accepted TCP connections are reflected in the 419 sin6_flowinfo field of the sockaddr_in6 structure returned to the 420 application by the recvfrom(), recvmsg(), and accept() functions. 421 And an application may specify the flow label and priority to use in 422 transmitted packets of a passively accepted TCP connection, by 423 setting the sin6_flowinfo field of the address passed in the bind() 424 function. 426 Implementations provide two bitmask constant declarations to help 427 applications select out the flow label and priority fields. These 428 constants are: 430 IPV6_FLOWINFO_FLOWLABEL 431 IPV6_FLOWINFO_PRIORITY 433 These constants can be applied to the sin6_flowinfo field of 434 addresses returned to the application, for example: 436 struct sockaddr_in6 sin6; 437 . . . 438 recvfrom(s, buf, buflen, flags, (struct sockaddr *) &sin6, &fromlen); 439 . . . 440 received_flowlabel = sin6.sin6_flowinfo & IPV6_FLOWINFO_FLOWLABEL; 441 received_priority = sin6.sin6_flowinfo & IPV6_FLOWINFO_PRIORITY; 443 On the sending side, applications are responsible for selecting the 444 flow label value. The system provides constant declarations for the 445 IPv6 priority values defined in the IPv6 specification [1]. These 446 constants are: 448 IPV6_PRIORITY_UNCHARACTERIZED 449 IPV6_PRIORITY_FILLER 450 IPV6_PRIORITY_UNATTENDED 451 IPV6_PRIORITY_RESERVED1 452 IPV6_PRIORITY_BULK 453 IPV6_PRIORITY_RESERVED2 454 IPV6_PRIORITY_INTERACTIVE 455 IPV6_PRIORITY_CONTROL 456 IPV6_PRIORITY_8 457 IPV6_PRIORITY_9 458 IPV6_PRIORITY_10 459 IPV6_PRIORITY_11 460 IPV6_PRIORITY_12 461 IPV6_PRIORITY_13 462 IPV6_PRIORITY_14 463 IPV6_PRIORITY_15 465 Applications can use these constants along with the flow label they 466 selected to assign the sin6_flowinfo field, for example: 468 struct sockaddr_in6 sin6; 469 . . . 470 send_flowlabel = . . . ; 471 . . . 472 sin6.sin6_flowinfo = IPV6_PRIORITY_UNATTENDED | 473 (IPV6_FLOWINFO_FLOWLABEL & send_flowlabel); 475 The macro declarations for these constants are obtained by including 476 the header file . 478 3.9. Binding to System-Selected Address 480 While the bind() function allows applications to select the source IP 481 address of UDP packets and TCP connections, applications often wish 482 to let the system select the source address for them. In IPv4, this 483 is done by specifying the IPv4 address represented by the symbolic 484 constant INADDR_ANY in the bind() call, or by simply by skipping the 485 bind() entirely. 487 Since the IPv6 address type is a structure (struct in6_addr), a 488 symbolic constant can be used to initialize an IPv6 address variable, 489 but can not be used in an assignment. Therefore systems provide the 490 IPv6 address value that can be used to instruct the system to select 491 the source IPv6 address in two forms. 493 The first version is a global variable named "in6addr_any" which is 494 an in6_addr type structure. The extern declaration for this variable 495 is: 497 extern const struct in6_addr in6addr_any; 499 Applications use in6addr_any similarly to the way they use INADDR_ANY 500 in IPv4. For example, to bind a socket to port number 23, but let 501 the system select the source address, an application could use the 502 following code: 504 struct sockaddr_in6 sin6; 505 . . . 506 sin6.sin6_family = AF_INET6; 507 sin6.sin6_flowinfo = 0; 508 sin6.sin6_port = htons(23); 509 sin6.sin6_addr = in6addr_any; 510 . . . 511 if (bind(s, (struct sockaddr *) &sin6, sizeof(sin6)) == -1) 512 . . . 514 The other version is a symbolic constant named IN6ADDR_ANY_INIT. 515 This constant can be used to initialize an in6_addr structure: 517 struct in6_addr anyaddr = IN6ADDR_ANY_INIT; 519 Note that this constant can be used ONLY at declaration type. It can 520 not be used assign a previously declared in6_addr structure. For 521 example, the following code will not work: 523 /* This is the WRONG way to assign an unspecified address */ 524 struct sockaddr_in6 sin6; 525 . . . 526 sin6.sin6_addr = IN6ADDR_ANY_INIT; /* Will NOT compile */ 528 The extern declaration for in6addr_any and the macro declaration for 529 IN6ADDR_ANY_INIT are obtained by including . 531 3.10. Communicating with Local Services 533 Applications may need to send UDP packets to, or originate TCP 534 connections to, services residing on the local node. In IPv4, they 535 can do this by using the constant IPv4 address INADDR_LOOPBACK in 536 their connect(), sendto(), or sendmsg() call. 538 IPv6 also provides a loopback address which can be used to contact 539 local TCP and UDP services. Like the unspecified address, the IPv6 540 loopback address is provided in two forms -- a global variable and a 541 symbolic constant. 543 The global variable is an in6_addr type structure named 544 "in6addr_loopback." The extern declaration for this variable is: 546 extern const struct in6_addr in6addr_loopback; 548 Applications use in6addr_loopback as they would use INADDR_LOOPBACK 549 in IPv4 applications. For example, to open a TCP connection to the 550 local telnet server, an application could use the following code: 552 struct sockaddr_in6 sin6; 553 . . . 554 sin6.sin6_family = AF_INET6; 555 sin6.sin6_flowinfo = 0; 556 sin6.sin6_port = htons(23); 557 sin6.sin6_addr = in6addr_loopback; 558 . . . 559 if (connect(s, (struct sockaddr *) &sin6, sizeof(sin6)) == -1) 560 . . . 562 The symbolic constant is named IN6ADDR_LOOPBACK_INIT. It can be used 563 at declaration time ONLY; for example: 565 struct in6_addr loopbackaddr = IN6ADDR_LOOPBACK_INIT; 567 Like IN6ADDR_ANY_INIT, this constant can not be used in an assignment 568 to a previously declared IPv6 address variable. 570 The extern declaration for in6addr_loopback and the macro declaration 571 for IN6ADDR_LOOPBACK_INIT are obtained by including . 573 4. Socket Options 575 A number of new socket options are defined for IPv6. All of these 576 new options are at the IPPROTO_IPV6 level. That is, the "level" 577 parameter in the getsockopt() and setsockopt() call is IPPROTO_IPV6 578 when using these options. The constant name prefix IPV6_ is used in 579 all of the new socket options. This serves to clearly identify these 580 options as applying to IPv6. 582 The macro declaration for IPPROTO_IPV6, the new IPv6 socket options, 583 and related constants defined in this section are obtained by 584 including the header file 586 4.1 Changing Socket Type 588 Unix allows open sockets to be passed between processes via the 589 exec() call and other means. It is a relatively common application 590 practice to pass open sockets across exec() calls. Thus it is 591 possible for an application using the original API to pass an open 592 PF_INET socket to an application that is expecting to receive a 593 PF_INET6 socket. Similarly, it is possible for an application using 594 the extended API to pass an open PF_INET6 socket to an application 595 using the original API, which would be equipped only to deal with 596 PF_INET sockets. Either of these cases could cause problems, because 597 the application which is passed the open socket might not know how to 598 decode the address structures returned in subsequent socket 599 functions. 601 To remedy this problem, a new setsockopt() option is defined that 602 allows an application to "transform" a PF_INET6 socket into a PF_INET 603 socket and vice-versa. 605 An IPv6 application that is passed an open socket from an unknown 606 process may use the IPV6_ADDRFORM setsockopt() option to "convert" 607 the socket to PF_INET6. Once that has been done, the system will 608 return sockaddr_in6 address structures in subsequent socket 609 functions. Similarly, an IPv6 application that is about to pass an 610 open PF_INET6 socket to a program that may not be IPv6 capable may 611 "downgrade" the socket to PF_INET before calling exec(). After that, 612 the system will return sockaddr_in address structures to the 613 application that was exec()'ed. 615 The IPV6_ADDRFORM option is valid at both the IPPROTO_IP and 616 IPPROTO_IPV6 levels. The only valid option values are PF_INET6 and 617 PF_INET. For example, to convert a PF_INET6 socket to PF_INET, a 618 program would call: 620 int addrform = PF_INET; 622 if (setsockopt(s, IPPROTO_IPV6, IPV6_ADDRFORM, (char *) &addrform, 623 sizeof(addrform)) == -1) 624 perror("setsockopt IPV6_ADDRFORM"); 626 An application may use IPV6_ADDRFORM in the getsockopt() function to 627 learn whether an open socket is a PF_INET of PF_INET6 socket. For 628 example: 630 int addrform; 631 size_t len = sizeof(int); 633 if (getsockopt(s, IPPROTO_IPV6, IPV6_ADDRFORM, (char *) &addrform, 634 &len) == -1) 635 perror("getsockopt IPV6_ADDRFORM"); 636 if (addrform == PF_INET) 637 printf("This is an IPv4 socket.\n"); 638 else if (addrform == PF_INET6) 639 printf("This is an IPv6 socket.\n"); 640 else 641 printf("This system is broken.\n"); 643 4.2. Unicast Hop Limit 645 A new setsockopt() option is used to control the hop limit used in 646 outgoing unicast IPv6 packets. The name of this option is 647 IPV6_UNICAST_HOPS, and it is used at the IPPROTO_IPV6 layer. The 648 following example illustrates how it is used: 650 int hoplimit = 10; 652 if (setsockopt(s, IPPROTO_IPV6, IPV6_UNICAST_HOPS, (char *) &hoplimit, 653 sizeof(hoplimit)) == -1) 654 perror("setsockopt IPV6_UNICAST_HOPS"); 656 When the IPV6_UNICAST_HOPS option is set with setsockopt(), the 657 option value given is used as the hop limit for all subsequent 658 unicast packets sent via that socket. If the option is not set, the 659 system selects a default value. 661 The IPV6_UNICAST_HOPS option may be used in the getsockopt() function 662 to determine the hop limit value that the system will use for 663 subsequent unicast packets sent via that socket. For example: 665 int hoplimit; 666 int len = sizeof(hoplimit); 668 if (getsockopt(s, IPPROTO_IPV6, IPV6_UNICAST_HOPS, (char *) &hoplimit, 669 &len) == -1) 670 perror("getsockopt IPV6_UNICAST_HOPS"); 671 else 672 printf("Using %d for hop limit.\n", hoplimit); 674 4.3. Sending and Receiving Multicast Packets 676 IPv6 applications may send UDP multicast packets by simply specifying 677 an IPv6 multicast address in the address argument of the sendto() 678 function. 680 A few setsockopt options at the IPPROTO_IPV6 layer are used to 681 control some of the parameters of sending multicast packets. These 682 options are optional: applications may send multicast packets without 683 using these options. The setsockopt() options for controlling the 684 sending of multicast packets are summarized below: 686 IPV6_MULTICAST_IF 688 Set the interface to use for outgoing multicast packets. The 689 argument is an IPv6 address of the interface to use. 691 Argument type: struct in6_addr 693 IPV6_MULTICAST_HOPS 695 Set the hop limit to use for outgoing multicast packets. 696 (Note a separate option - IPV6_UNICAST_HOPS - is provided to 697 set the hop limit to use for outgoing unicast packets.) 699 Argument type: unsigned int 701 IPV6_MULTICAST_LOOP 703 Controls whether outgoing multicast packets sent should be 704 delivered back to the local application. A toggle. If the 705 option is set to 1, multicast packets are looped back. If it 706 is set to 0, they are not. 708 Argument type: unsigned int 710 The reception of multicast packets is controlled by the two 711 setsockopt() options summarized below: 713 IPV6_ADD_MEMBERSHIP 715 Join a multicast group. Requests that multicast packets sent 716 to a particular multicast address be delivered to this 717 socket. The argument is the IPv6 multicast address of the 718 group to join. 720 Argument type: struct ipv6_mreq 722 IPV6_DROP_MEMBERSHIP 724 Leave a multicast group. Requests that multicast packets 725 sent to a particular multicast address no longer be delivered 726 to this socket. The argument is the IPv6 multicast address 727 of the group to join. 729 Argument type: struct ipv6_mreq 731 The argument type of both of these options is the ipv6_mreq 732 structure, which is defined as follows: 734 struct ipv6_mreq { 735 /* IPv6 multicast address of group */ 736 struct in6_addr ipv6mr_multiaddr; 738 /* local IPv6 address of interface */ 739 struct in6_addr ipv6mr_interface; 740 }; 742 5. Library Functions 744 New library functions are needed to perform a variety of operations 745 with IPv6 addresses. Functions are needed to lookup IPv6 addresses 746 in the Domain Name System (DNS). Both forward lookup (hostname to 747 address translation) and reverse lookup (address to hostname 748 translation) need to be supported. Functions are also needed to 749 convert IPv6 addresses between their binary and textual form. 751 5.1. Hostname to Address Translation 753 A new hostname to address translation function is being defined by 754 the Institute of Electrical and Electronic Engineers (IEEE) as part 755 of the POSIX 1003.1g (Protocol Independent Interfaces) draft 756 specification [4]. This function, named getaddrinfo(), has been 757 designed to be protocol independent, so it can be used without change 758 to lookup IPv6 addresses. 760 As discussed in the "Transition Mechanisms for IPv6 Hosts and 761 Routers" specification [5], systems may provide the ability to 762 transparently query for IPv4 address records when the application 763 requests an IPv6 lookup. The getaddrinfo() function can implement 764 this by automatically performing a query for IPv4 records if its 765 initial query for IPv6 records finds none. Or it may elect to always 766 query for both IPv6 and IPv4 records on all lookups. (Many DNS 767 implementations do not support querying for multiple record types in 768 a single request, so the IPv6 and IPv4 lookups can not be performed 769 simultaneously.) If IPv4 records are found, the addresses can be 770 returned to the application as IPv4-mapped IPv6 addresses. Systems 771 that support transparent querying for IPv4 address records should 772 provide a system-wide configuration switch to allow the system 773 administrator to enable or disable that feature. 775 5.2. Address to Hostname Translation 777 The POSIX 1003.1g specification includes no function to perform a 778 reverse DNS lookup (query based on IPv6 address). Therefore, we have 779 defined the following function: 781 int getnameinfo( 782 const struct sockaddr *sa, 783 size_t addrlen, 784 char *host, 785 size_t hostlen, 786 char *serv, 787 size_t servlen); 789 This function looks up an IP address and port number provided by the 790 caller in the DNS and system-specific database, and returns text 791 strings for both in buffers provided by the caller. The first 792 argument, sa, points to either a sockaddr_in structure (for IPv4) or 793 a sockaddr_in6 structure (for IPv6) which holds the IP address and 794 port number. The addrlen argument gives the length of the 795 sockaddr_in or sockaddr_in6 structure. The function returns the 796 hostname associated with the IP address in the buffer pointed to by 797 the host argument. The caller provides the size of this buffer via 798 the hostlen argument. The service name associated with the port 799 number is returned in the buffer pointed to by serv, and the servlen 800 argument gives the length of this buffer. The caller may instruct 801 the function not to return either string by providing a zero value 802 for the hostlen or servlen arguments. Otherwise, the caller must 803 provide buffers large enough to hold the fully qualified domain 804 hostname, and the full service name, including the terminating null 805 character. The function indicates successful completion by a zero 806 return value; a non-zero return value indicates failure. 808 Applications obtain the function prototype declaration for 809 getnameinfo() by including the header file . 811 5.3. Address Conversion Functions 813 BSD Unix provides two functions, inet_addr() and inet_ntoa(), to 814 convert an IPv4 address between binary and text form. IPv6 815 applications need similar functions. The following two functions 816 convert both IPv6 and IPv4 addresses: 818 ssize_t inet_pton( 819 int af, 820 const char *cp, 821 void *ap); 823 and 825 char *inet_ntop( 826 int af, 827 const void *ap, 828 size_t len, 829 char *cp); 831 The first function converts an address in its standard text 832 presentation form into its numeric binary form. The af argument 833 specifies the family of the address. Currently AF_INET and AF_INET6 834 address families are supported. The cp argument points to the string 835 being passed in. The ap argument points to a buffer into which the 836 function stores the numeric address. The address is returned in 837 network byte order. Inet_pton() returns the length of the address in 838 octets if the conversion succeeds, and -1 otherwise. The function 839 does not modify the buffer pointed to by ap if the conversion fails. 840 The calling application must ensure that the buffer referred to by ap 841 is large enough to hold the converted address. 843 If the af argument is AF_INET, the function accepts a string in the 844 standard IPv4 dotted decimal form: 846 ddd.ddd.ddd.ddd 848 where ddd is a one to three digit decimal number between 0 and 255. 850 If the af argument is AF_INET6, then the function accepts a string in 851 one of the standard IPv6 text forms defined in the addressing 852 architecture specification [2]. 854 The second function converts a numeric address into a text string 855 suitable for presentation. The af argument specifies the family of 856 the address. This can be AF_INET or AF_INET6. The ap argument 857 points to a buffer holding an IPv4 address if the af argument is 858 AF_INET, or an IPv6 address if the af argument is AF_INET6. The len 859 field specifies the length in octets of the address pointed to by ap. 860 This must be 4 if af is AF_INET, or 16 if af is AF_INET6. The cp 861 argument points to a buffer that the function can use to store the 862 text string. If the cp argument is NULL, the function uses its own 863 private static buffer. If the application specifies a non-NULL cp 864 argument, the buffer must be large enough to hold the text 865 representation of the address, including the terminating null octet. 866 For IPv6 addresses, the buffer must be at least 46-octets. For IPv4 867 addresses, the buffer must be at least 16-octets. In order to allow 868 applications to easily declare buffers of the proper size to store 869 IPv4 and IPv6 addresses in string form, implementations should 870 provide the following constants, made available to applications that 871 include : 873 #define INET_ADDRSTRLEN 16 874 #define INET6_ADDRSTRLEN 46 876 The inet_ntop() function returns a pointer to the buffer containing 877 the text string if the conversion succeeds, and NULL otherwise. The 878 function does not modify the storage pointed to by cp if the 879 conversion fails. 881 Applications obtain the prototype declarations for inet_ntop() and 882 inet_pton() by including the header file . 884 5.4. Embedded IPv4 Addresses 886 The IPv4-mapped IPv6 address format is used to represent IPv4 887 addresses as IPv6 addresses. Most applications should be able to to 888 manipulate IPv6 addresses as opaque 16-octet quantities, without 889 needing to know whether they represent IPv4 addresses. However, a 890 few applications may need to determine whether an IPv6 address is an 891 IPv4-mapped address or not. The following function is provided for 892 those applications: 894 int inet6_isipv4addr (const struct in6_addr *addr); 896 The "addr" argument to this function points to a buffer holding an 897 IPv6 address in network byte order. The function returns true (non- 898 zero) if that address is an IPv4-mapped address, and returns 0 899 otherwise. 901 This function could be used by server applications to determine 902 whether the peer is an IPv4 node or an IPv6 node. After accepting a 903 TCP connection via accept(), or receiving a UDP packet via 904 recvfrom(), the application can apply the inet6_isipv4addr() function 905 to the returned address. 907 Applications obtain the prototype for this function by including the 908 header file . 910 6. Security Considerations 912 IPv6 provides a number of new security mechanisms, many of which need 913 to be accessible to applications. A companion document detailing the 914 extensions to the socket interfaces to support IPv6 security is being 915 written [3]. At some point in the future, that document and this one 916 may be merged into a single API specification. 918 7. Change History 920 Changes from the January 1996 Edition 922 - Eliminated source routing and interface identification features 923 in order to simplify the spec. API features to provide this 924 functionallity can be defined at a later time. 926 - Eliminated definitions of hostname2addr() and addr2hostname(). 927 Added reference to POSIX getaddrinfo() function to provide 928 functionallity previously provided by hostname2addr(). Added 929 definition of getnameinfo() function to provide functionallity of 930 addr2hostname(). 932 - Changed name of addr2ascii() and ascii2addr() functions to 933 inet_ntop() and inet_pton() to be more consistent with BSD 934 function naming conventions. 936 - Changed some type definitions to align with POSIX. 938 Changes from the November 1995 Edition 940 - Added the symbolic constants IPV6ADDR_ANY_INIT and 941 IPV6ADDR_LOOPBACK_INIT for applications to use for 942 initializations. 944 - Eliminated restrictions on the value of ipv6addr_any. Systems 945 may now choose any value, including all-zeros. 947 - Added a mechanism for returning time to live with the address in 948 the name-to-address translation functions. 950 - Added a mechanism for applications to specify the interface in 951 the setsockopt() options to join and leave a multicast group. 953 Changes from the July 1995 Edition 955 - Changed u_long and u_short types in structures to u_int32_t and 956 u_int16_t for consistency and clarity. 958 - Added implementation-provided constants for IPv4 and IPv6 text 959 address buffer length. 961 - Defined a set of constants for subfields of sin6_flowid and for 962 priority values. 964 - Defined constants for getting and setting the source route flag. 966 - Define where ansi prototypes for hostname2addr(), 967 addr2hostname(), addr2ascii(), ascii2addr(), and 968 ipv6_isipv4addr() reside. 970 - Clarified the include file requirements. Say that the structure 971 definitions are defined as a result of including the header file 972 , not that the structures are necessarily defined 973 there. 975 - Removed underscore chars from is_ipv4_addr() function name for 976 BSD compatibility. 978 - Added inet6_ prefix to is_ipv4_addr() function name to avoid name 979 space conflicts. 981 - Changes setsockopt option naming convention to use IPV6_ prefix 982 instead of IP_ so that there is clearly no ambiguity with IPv4 983 options. Also, use level IPPROTO_IPV6 for these options. 985 - Made hostname2addr() and addr2hostname() functions thread-safe. 987 - Added support for sendmsg() and recvmsg() in source routing 988 section. 990 - Changed in_addr6 to in6_addr for consistency. 992 - Re-structured document into sub-sections. 994 - Deleted the implementation experience section. It was too wordy. 996 - Added argument types to multicast socket options. 998 - Added constant for largest source route array buffer. 1000 - Added the freehostent() function. 1002 - Added receiving interface determination and sending interface 1003 selection options. 1005 - Added definitions of ipv6addr_any and ipv6addr_loopback. 1007 - Added text making the lookup of IPv4 addresses by hostname2addr() 1008 optional. 1010 Changes from the June 1995 Edition 1012 - Added capability for application to select loose or strict source 1013 routing. 1015 Changes from the March 1995 Edition 1017 - Changed the definition of the ipv6_addr structure to be an array 1018 of sixteen chars instead of four longs. This change is necessary 1019 to support machines which implement the socket interface, but do 1020 not have a 32-bit addressable word. Virtually all machines which 1021 provide the socket interface do support an 8-bit addressable data 1022 type. 1024 - Added a more detailed explanation that the data types defined in 1025 this documented are not intended to be hard and fast 1026 requirements. Systems may use other data types if they wish. 1028 - Added a note flagging the fact that the sockaddr_in6 structure is 1029 not the same size as the sockaddr structure. 1031 - Changed the sin6_flowlabel field to sin6_flowinfo to accommodate 1032 the addition of the priority field to the IPv6 header. 1034 Changes from the October 1994 Edition 1036 - Added variant of sockaddr_in6 for 4.4 BSD-based systems (sa_len 1037 compatibility). 1039 - Removed references to SIT transition specification, and added 1040 reference to addressing architecture document, for definition of 1041 IPv4-mapped addresses. 1043 - Added a solution to the problem of the application not providing 1044 enough buffer space to hold a received source route. 1046 - Moved discussion of IPv4 applications interoperating with IPv6 1047 nodes to open issues section. 1049 - Added length parameter to addr2ascii() function to be consistent 1050 with addr2hostname(). 1052 - Changed IP_MULTICAST_TTL to IP_MULTICAST_HOPS to match IPv6 1053 terminology, and added IP_UNICAST_HOPS option to match 1054 IP_MULTICAST_HOPS. 1056 - Removed specification of numeric values for AF_INET6, 1057 IP_ADDRFORM, and IP_RCVSRCRT, since they need not be the same on 1058 different implementations. 1060 - Added a definition for the in_addr6 IPv6 address data structure. 1061 Added this so that applications could use sizeof(struct in_addr6) 1062 to get the size of an IPv6 address, and so that a structured type 1063 could be used in the is_ipv4_addr(). 1064 8. Open Issues 1066 A few open issues for IPv6 socket interface API specification remain, 1067 including: 1069 - An API should be provided to allocate and free a flow label that 1070 meets the uniqueness and randomness requirements outlined in the 1071 IPv6 protocol spec. 1073 - Should we add a timeout parameter to the hostname/address 1074 translation functions? DNS lookups need to be given some finite 1075 timeout interval, so it might be nice to let the application 1076 specify that interval. 1078 - Can the IPV6_ADDRFORM option really be implemented? 1080 - Can existing IPv4 applications interoperate with IPv6 nodes? 1081 This issue is discussed in more detail in the following section. 1083 8.1. IPv4 Applications Interoperating with IPv6 Nodes 1085 This problem primarily has to do with the how IPv4 applications 1086 represent addresses of IPv6 nodes. What address should be returned 1087 to the application when an IPv6/UDP packet is received, or an 1088 IPv6/TCP connection is accepted? The peer's address could be any 1089 arbitrary 128-bit IPv6 address. But the application is only equipped 1090 to deal with 32-bit IPv4 addresses encoded in sockaddr_in data 1091 structures. 1093 We have not discovered any solution that provides complete 1094 transparent interoperability with IPv6 nodes for applications using 1095 the original IPv4 API. However, two techniques that partially solve 1096 the problem are: 1098 1) Prohibit communication between IPv4 applications and IPv6 nodes. 1099 Only UDP packets received from IPv4 nodes would be passed up to 1100 the application, and only TCP connections received from IPv4 1101 nodes would be accepted. UDP packets from IPv6 nodes would be 1102 dropped, and TCP connections from IPv6 nodes would be refused. 1104 2) The system could generate a local 32-bit cookie to represent the 1105 full 128-bit IPv6 address, and pass this value to the 1106 application. The system would maintain a mapping from cookie 1107 value into the 128-bit IPv6 address that it represents. When the 1108 application passed a cookie back into the system (for example, in 1109 a sendto() or connect() call) the system would use the 128-bit 1110 IPv6 address that the cookie represents. 1112 The cookie would have to be chosen so as to be an invalid IPv4 1113 address (e.g. an address on net 127.0.0.0), and the system would 1114 have to make sure that these cookie values did not escape into 1115 the Internet as the source or destination addresses of IPv4 1116 packets. 1118 Both of these techniques have drawbacks. This is an area for further 1119 study. System implementors may use one of these techniques or 1120 implement another solution. 1122 Acknowledgments 1124 Thanks to the many people who made suggestions and provided feedback 1125 to to the numerous revisions of this document, including: Werner 1126 Almesberger, Ran Atkinson, Fred Baker, Dave Borman, Andrew Cherenson, 1127 Alex Conta, Alan Cox, Steve Deering, Francis Dupont, Robert Elz, Marc 1128 Hasson, Tom Herbert, Christian Huitema, Wan-Yen Hsu, Alan Lloyd, 1129 Charles Lynn, Dan McDonald, Craig Metz, Erik Nordmark, Josh Osborne, 1130 Craig Partridge, Richard Stevens, Matt Thomas, Dean D. Throop, Glenn 1131 Trewitt, Paul Vixie, David Waitzman, and Carl Williams. The 1132 getnameinfo() function was based on the getinfobysockaddr() function 1133 defined by Keith Sklower. 1135 Ramesh Govindan made a number of contributions and co-authored an 1136 earlier version of this paper. 1138 References 1140 [1] S. Deering, R. Hinden. "Internet Protocol, Version 6 (IPv6) 1141 Specification". RFC 1883. December 1995. 1143 [2] R. Hinden, S. Deering. "IP Version 6 Addressing Architecture". 1144 RFC 1884. December 1995. 1146 [3] D. McDonald. "IPv6 Security API for BSD Sockets". Internet 1147 Draft. January 1995. 1149 [4] IEEE, "Protocol Independent Interfaces", IEEE Std 1003.1g, DRAFT 1150 6.3. November 1995. 1152 [5] R. Gilligan, E. Nordmark. "Transition Mechanisms for IPv6 Hosts 1153 and Routers". RFC 1933. April 1996. 1155 Authors' Address 1157 Jim Bound 1158 Digital Equipment Corporation 1159 110 Spitbrook Road ZK3-3/U14 1160 Nashua, NH 03062-2698 1161 Phone: +1 603 881 0400 1162 Email: bound@zk3.dec.com 1164 Susan Thomson 1165 Bell Communications Research 1166 MRE 2P-343, 445 South Street 1167 Morristown, NJ 07960 1168 Telephone: +1 201 829 4514 1169 Email: set@thumper.bellcore.com 1171 Robert E. Gilligan 1172 Mailstop MPK 17-202 1173 Sun Microsystems, Inc. 1174 2550 Garcia Avenue 1175 Mountain View, CA 94043-1100 1176 Phone: +1 415 786 5151 1177 Email: gilligan@eng.sun.com