net/ipv4/Kconfig

   1 #
   2 # IP configuration
   3 #
   4 config IP_MULTICAST
   5         bool "IP: multicasting"
   6         help
   7           This is code for addressing several networked computers at once,
   8           enlarging your kernel by about 2 KB. You need multicasting if you
   9           intend to participate in the MBONE, a high bandwidth network on top
  10           of the Internet which carries audio and video broadcasts. More
  11           information about the MBONE is on the WWW at
  12           <http://www.savetz.com/mbone/>. Information about the multicast
  13           capabilities of the various network cards is contained in
  14           <file:Documentation/networking/multicast.txt>. For most people, it's
  15           safe to say N.
  16
  17 config IP_ADVANCED_ROUTER
  18         bool "IP: advanced router"
  19         ---help---
  20           If you intend to run your Linux box mostly as a router, i.e. as a
  21           computer that forwards and redistributes network packets, say Y; you
  22           will then be presented with several options that allow more precise
  23           control about the routing process.
  24
  25           The answer to this question won't directly affect the kernel:
  26           answering N will just cause the configurator to skip all the
  27           questions about advanced routing.
  28
  29           Note that your box can only act as a router if you enable IP
  30           forwarding in your kernel; you can do that by saying Y to "/proc
  31           file system support" and "Sysctl support" below and executing the
  32           line
  33
  34           echo "1" > /proc/sys/net/ipv4/ip_forward
  35
  36           at boot time after the /proc file system has been mounted.
  37
  38           If you turn on IP forwarding, you should consider the rp_filter, which
  39           automatically rejects incoming packets if the routing table entry
  40           for their source address doesn't match the network interface they're
  41           arriving on. This has security advantages because it prevents the
  42           so-called IP spoofing, however it can pose problems if you use
  43           asymmetric routing (packets from you to a host take a different path
  44           than packets from that host to you) or if you operate a non-routing
  45           host which has several IP addresses on different interfaces. To turn
  46           rp_filter on use:
  47
  48           echo 1 > /proc/sys/net/ipv4/conf/<device>/rp_filter
  49            and
  50           echo 1 > /proc/sys/net/ipv4/conf/all/rp_filter
  51
  52           Note that some distributions enable it in startup scripts.
  53           For details about rp_filter strict and loose mode read
  54           <file:Documentation/networking/ip-sysctl.txt>.
  55
  56           If unsure, say N here.
  57
  58 choice
  59         prompt "Choose IP: FIB lookup algorithm (choose FIB_HASH if unsure)"
  60         depends on IP_ADVANCED_ROUTER
  61         default ASK_IP_FIB_HASH
  62
  63 config ASK_IP_FIB_HASH
  64         bool "FIB_HASH"
  65         ---help---
  66           Current FIB is very proven and good enough for most users.
  67
  68 config IP_FIB_TRIE
  69         bool "FIB_TRIE"
  70         ---help---
  71           Use new experimental LC-trie as FIB lookup algorithm.
  72           This improves lookup performance if you have a large
  73           number of routes.
  74
  75           LC-trie is a longest matching prefix lookup algorithm which
  76           performs better than FIB_HASH for large routing tables.
  77           But, it consumes more memory and is more complex.
  78
  79           LC-trie is described in:
  80
  81           IP-address lookup using LC-tries. Stefan Nilsson and Gunnar Karlsson
  82           IEEE Journal on Selected Areas in Communications, 17(6):1083-1092,
  83           June 1999
  84
  85           An experimental study of compression methods for dynamic tries
  86           Stefan Nilsson and Matti Tikkanen. Algorithmica, 33(1):19-33, 2002.
  87           http://www.nada.kth.se/~snilsson/public/papers/dyntrie2/
  88
  89 endchoice
  90
  91 config IP_FIB_HASH
  92         def_bool ASK_IP_FIB_HASH || !IP_ADVANCED_ROUTER
  93
  94 config IP_FIB_TRIE_STATS
  95         bool "FIB TRIE statistics"
  96         depends on IP_FIB_TRIE
  97         ---help---
  98           Keep track of statistics on structure of FIB TRIE table.
  99           Useful for testing and measuring TRIE performance.
 100
 101 config IP_MULTIPLE_TABLES
 102         bool "IP: policy routing"
 103         depends on IP_ADVANCED_ROUTER
 104         select FIB_RULES
 105         ---help---
 106           Normally, a router decides what to do with a received packet based
 107           solely on the packet's final destination address. If you say Y here,
 108           the Linux router will also be able to take the packet's source
 109           address into account. Furthermore, the TOS (Type-Of-Service) field
 110           of the packet can be used for routing decisions as well.
 111
 112           If you are interested in this, please see the preliminary
 113           documentation at <http://www.compendium.com.ar/policy-routing.txt>
 114           and <ftp://post.tepkom.ru/pub/vol2/Linux/docs/advanced-routing.tex>.
 115           You will need supporting software from
 116           <ftp://ftp.tux.org/pub/net/ip-routing/>.
 117
 118           If unsure, say N.
 119
 120 config IP_ROUTE_MULTIPATH
 121         bool "IP: equal cost multipath"
 122         depends on IP_ADVANCED_ROUTER
 123         help
 124           Normally, the routing tables specify a single action to be taken in
 125           a deterministic manner for a given packet. If you say Y here
 126           however, it becomes possible to attach several actions to a packet
 127           pattern, in effect specifying several alternative paths to travel
 128           for those packets. The router considers all these paths to be of
 129           equal "cost" and chooses one of them in a non-deterministic fashion
 130           if a matching packet arrives.
 131
 132 config IP_ROUTE_VERBOSE
 133         bool "IP: verbose route monitoring"
 134         depends on IP_ADVANCED_ROUTER
 135         help
 136           If you say Y here, which is recommended, then the kernel will print
 137           verbose messages regarding the routing, for example warnings about
 138           received packets which look strange and could be evidence of an
 139           attack or a misconfigured system somewhere. The information is
 140           handled by the klogd daemon which is responsible for kernel messages
 141           ("man klogd").
 142
 143 config IP_PNP
 144         bool "IP: kernel level autoconfiguration"
 145         help
 146           This enables automatic configuration of IP addresses of devices and
 147           of the routing table during kernel boot, based on either information
 148           supplied on the kernel command line or by BOOTP or RARP protocols.
 149           You need to say Y only for diskless machines requiring network
 150           access to boot (in which case you want to say Y to "Root file system
 151           on NFS" as well), because all other machines configure the network
 152           in their startup scripts.
 153
 154 config IP_PNP_DHCP
 155         bool "IP: DHCP support"
 156         depends on IP_PNP
 157         ---help---
 158           If you want your Linux box to mount its whole root file system (the
 159           one containing the directory /) from some other computer over the
 160           net via NFS and you want the IP address of your computer to be
 161           discovered automatically at boot time using the DHCP protocol (a
 162           special protocol designed for doing this job), say Y here. In case
 163           the boot ROM of your network card was designed for booting Linux and
 164           does DHCP itself, providing all necessary information on the kernel
 165           command line, you can say N here.
 166
 167           If unsure, say Y. Note that if you want to use DHCP, a DHCP server
 168           must be operating on your network.  Read
 169           <file:Documentation/filesystems/nfsroot.txt> for details.
 170
 171 config IP_PNP_BOOTP
 172         bool "IP: BOOTP support"
 173         depends on IP_PNP
 174         ---help---
 175           If you want your Linux box to mount its whole root file system (the
 176           one containing the directory /) from some other computer over the
 177           net via NFS and you want the IP address of your computer to be
 178           discovered automatically at boot time using the BOOTP protocol (a
 179           special protocol designed for doing this job), say Y here. In case
 180           the boot ROM of your network card was designed for booting Linux and
 181           does BOOTP itself, providing all necessary information on the kernel
 182           command line, you can say N here. If unsure, say Y. Note that if you
 183           want to use BOOTP, a BOOTP server must be operating on your network.
 184           Read <file:Documentation/filesystems/nfsroot.txt> for details.
 185
 186 config IP_PNP_RARP
 187         bool "IP: RARP support"
 188         depends on IP_PNP
 189         help
 190           If you want your Linux box to mount its whole root file system (the
 191           one containing the directory /) from some other computer over the
 192           net via NFS and you want the IP address of your computer to be
 193           discovered automatically at boot time using the RARP protocol (an
 194           older protocol which is being obsoleted by BOOTP and DHCP), say Y
 195           here. Note that if you want to use RARP, a RARP server must be
 196           operating on your network. Read
 197           <file:Documentation/filesystems/nfsroot.txt> for details.
 198
 199 # not yet ready..
 200 #   bool '    IP: ARP support' CONFIG_IP_PNP_ARP
 201 config NET_IPIP
 202         tristate "IP: tunneling"
 203         select INET_TUNNEL
 204         ---help---
 205           Tunneling means encapsulating data of one protocol type within
 206           another protocol and sending it over a channel that understands the
 207           encapsulating protocol. This particular tunneling driver implements
 208           encapsulation of IP within IP, which sounds kind of pointless, but
 209           can be useful if you want to make your (or some other) machine
 210           appear on a different network than it physically is, or to use
 211           mobile-IP facilities (allowing laptops to seamlessly move between
 212           networks without changing their IP addresses).
 213
 214           Saying Y to this option will produce two modules ( = code which can
 215           be inserted in and removed from the running kernel whenever you
 216           want). Most people won't need this and can say N.
 217
 218 config NET_IPGRE
 219         tristate "IP: GRE tunnels over IP"
 220         help
 221           Tunneling means encapsulating data of one protocol type within
 222           another protocol and sending it over a channel that understands the
 223           encapsulating protocol. This particular tunneling driver implements
 224           GRE (Generic Routing Encapsulation) and at this time allows
 225           encapsulating of IPv4 or IPv6 over existing IPv4 infrastructure.
 226           This driver is useful if the other endpoint is a Cisco router: Cisco
 227           likes GRE much better than the other Linux tunneling driver ("IP
 228           tunneling" above). In addition, GRE allows multicast redistribution
 229           through the tunnel.
 230
 231 config NET_IPGRE_BROADCAST
 232         bool "IP: broadcast GRE over IP"
 233         depends on IP_MULTICAST && NET_IPGRE
 234         help
 235           One application of GRE/IP is to construct a broadcast WAN (Wide Area
 236           Network), which looks like a normal Ethernet LAN (Local Area
 237           Network), but can be distributed all over the Internet. If you want
 238           to do that, say Y here and to "IP multicast routing" below.
 239
 240 config IP_MROUTE
 241         bool "IP: multicast routing"
 242         depends on IP_MULTICAST
 243         help
 244           This is used if you want your machine to act as a router for IP
 245           packets that have several destination addresses. It is needed on the
 246           MBONE, a high bandwidth network on top of the Internet which carries
 247           audio and video broadcasts. In order to do that, you would most
 248           likely run the program mrouted. Information about the multicast
 249           capabilities of the various network cards is contained in
 250           <file:Documentation/networking/multicast.txt>. If you haven't heard
 251           about it, you don't need it.
 252
 253 config IP_PIMSM_V1
 254         bool "IP: PIM-SM version 1 support"
 255         depends on IP_MROUTE
 256         help
 257           Kernel side support for Sparse Mode PIM (Protocol Independent
 258           Multicast) version 1. This multicast routing protocol is used widely
 259           because Cisco supports it. You need special software to use it
 260           (pimd-v1). Please see <http://netweb.usc.edu/pim/> for more
 261           information about PIM.
 262
 263           Say Y if you want to use PIM-SM v1. Note that you can say N here if
 264           you just want to use Dense Mode PIM.
 265
 266 config IP_PIMSM_V2
 267         bool "IP: PIM-SM version 2 support"
 268         depends on IP_MROUTE
 269         help
 270           Kernel side support for Sparse Mode PIM version 2. In order to use
 271           this, you need an experimental routing daemon supporting it (pimd or
 272           gated-5). This routing protocol is not used widely, so say N unless
 273           you want to play with it.
 274
 275 config ARPD
 276         bool "IP: ARP daemon support"
 277         ---help---
 278           The kernel maintains an internal cache which maps IP addresses to
 279           hardware addresses on the local network, so that Ethernet/Token Ring/
 280           etc. frames are sent to the proper address on the physical networking
 281           layer. Normally, kernel uses the ARP protocol to resolve these
 282           mappings.
 283
 284           Saying Y here adds support to have an user space daemon to do this
 285           resolution instead. This is useful for implementing an alternate
 286           address resolution protocol (e.g. NHRP on mGRE tunnels) and also for
 287           testing purposes.
 288
 289           If unsure, say N.
 290
 291 config SYN_COOKIES
 292         bool "IP: TCP syncookie support (disabled per default)"
 293         ---help---
 294           Normal TCP/IP networking is open to an attack known as "SYN
 295           flooding". This denial-of-service attack prevents legitimate remote
 296           users from being able to connect to your computer during an ongoing
 297           attack and requires very little work from the attacker, who can
 298           operate from anywhere on the Internet.
 299
 300           SYN cookies provide protection against this type of attack. If you
 301           say Y here, the TCP/IP stack will use a cryptographic challenge
 302           protocol known as "SYN cookies" to enable legitimate users to
 303           continue to connect, even when your machine is under attack. There
 304           is no need for the legitimate users to change their TCP/IP software;
 305           SYN cookies work transparently to them. For technical information
 306           about SYN cookies, check out <http://cr.yp.to/syncookies.html>.
 307
 308           If you are SYN flooded, the source address reported by the kernel is
 309           likely to have been forged by the attacker; it is only reported as
 310           an aid in tracing the packets to their actual source and should not
 311           be taken as absolute truth.
 312
 313           SYN cookies may prevent correct error reporting on clients when the
 314           server is really overloaded. If this happens frequently better turn
 315           them off.
 316
 317           If you say Y here, note that SYN cookies aren't enabled by default;
 318           you can enable them by saying Y to "/proc file system support" and
 319           "Sysctl support" below and executing the command
 320
 321           echo 1 >/proc/sys/net/ipv4/tcp_syncookies
 322
 323           at boot time after the /proc file system has been mounted.
 324
 325           If unsure, say N.
 326
 327 config INET_AH
 328         tristate "IP: AH transformation"
 329         select XFRM
 330         select CRYPTO
 331         select CRYPTO_HMAC
 332         select CRYPTO_MD5
 333         select CRYPTO_SHA1
 334         ---help---
 335           Support for IPsec AH.
 336
 337           If unsure, say Y.
 338
 339 config INET_ESP
 340         tristate "IP: ESP transformation"
 341         select XFRM
 342         select CRYPTO
 343         select CRYPTO_AUTHENC
 344         select CRYPTO_HMAC
 345         select CRYPTO_MD5
 346         select CRYPTO_CBC
 347         select CRYPTO_SHA1
 348         select CRYPTO_DES
 349         ---help---
 350           Support for IPsec ESP.
 351
 352           If unsure, say Y.
 353
 354 config INET_IPCOMP
 355         tristate "IP: IPComp transformation"
 356         select INET_XFRM_TUNNEL
 357         select XFRM_IPCOMP
 358         ---help---
 359           Support for IP Payload Compression Protocol (IPComp) (RFC3173),
 360           typically needed for IPsec.
 361
 362           If unsure, say Y.
 363
 364 config INET_XFRM_TUNNEL
 365         tristate
 366         select INET_TUNNEL
 367         default n
 368
 369 config INET_TUNNEL
 370         tristate
 371         default n
 372
 373 config INET_XFRM_MODE_TRANSPORT
 374         tristate "IP: IPsec transport mode"
 375         default y
 376         select XFRM
 377         ---help---
 378           Support for IPsec transport mode.
 379
 380           If unsure, say Y.
 381
 382 config INET_XFRM_MODE_TUNNEL
 383         tristate "IP: IPsec tunnel mode"
 384         default y
 385         select XFRM
 386         ---help---
 387           Support for IPsec tunnel mode.
 388
 389           If unsure, say Y.
 390
 391 config INET_XFRM_MODE_BEET
 392         tristate "IP: IPsec BEET mode"
 393         default y
 394         select XFRM
 395         ---help---
 396           Support for IPsec BEET mode.
 397
 398           If unsure, say Y.
 399
 400 config INET_LRO
 401         bool "Large Receive Offload (ipv4/tcp)"
 402         default y
 403         ---help---
 404           Support for Large Receive Offload (ipv4/tcp).
 405
 406           If unsure, say Y.
 407
 408 config INET_DIAG
 409         tristate "INET: socket monitoring interface"
 410         default y
 411         ---help---
 412           Support for INET (TCP, DCCP, etc) socket monitoring interface used by
 413           native Linux tools such as ss. ss is included in iproute2, currently
 414           downloadable at <http://linux-net.osdl.org/index.php/Iproute2>.
 415
 416           If unsure, say Y.
 417
 418 config INET_TCP_DIAG
 419         depends on INET_DIAG
 420         def_tristate INET_DIAG
 421
 422 menuconfig TCP_CONG_ADVANCED
 423         bool "TCP: advanced congestion control"
 424         ---help---
 425           Support for selection of various TCP congestion control
 426           modules.
 427
 428           Nearly all users can safely say no here, and a safe default
 429           selection will be made (CUBIC with new Reno as a fallback).
 430
 431           If unsure, say N.
 432
 433 if TCP_CONG_ADVANCED
 434
 435 config TCP_CONG_BIC
 436         tristate "Binary Increase Congestion (BIC) control"
 437         default m
 438         ---help---
 439         BIC-TCP is a sender-side only change that ensures a linear RTT
 440         fairness under large windows while offering both scalability and
 441         bounded TCP-friendliness. The protocol combines two schemes
 442         called additive increase and binary search increase. When the
 443         congestion window is large, additive increase with a large
 444         increment ensures linear RTT fairness as well as good
 445         scalability. Under small congestion windows, binary search
 446         increase provides TCP friendliness.
 447         See http://www.csc.ncsu.edu/faculty/rhee/export/bitcp/
 448
 449 config TCP_CONG_CUBIC
 450         tristate "CUBIC TCP"
 451         default y
 452         ---help---
 453         This is version 2.0 of BIC-TCP which uses a cubic growth function
 454         among other techniques.
 455         See http://www.csc.ncsu.edu/faculty/rhee/export/bitcp/cubic-paper.pdf
 456
 457 config TCP_CONG_WESTWOOD
 458         tristate "TCP Westwood+"
 459         default m
 460         ---help---
 461         TCP Westwood+ is a sender-side only modification of the TCP Reno
 462         protocol stack that optimizes the performance of TCP congestion
 463         control. It is based on end-to-end bandwidth estimation to set
 464         congestion window and slow start threshold after a congestion
 465         episode. Using this estimation, TCP Westwood+ adaptively sets a
 466         slow start threshold and a congestion window which takes into
 467         account the bandwidth used  at the time congestion is experienced.
 468         TCP Westwood+ significantly increases fairness wrt TCP Reno in
 469         wired networks and throughput over wireless links.
 470
 471 config TCP_CONG_HTCP
 472         tristate "H-TCP"
 473         default m
 474         ---help---
 475         H-TCP is a send-side only modifications of the TCP Reno
 476         protocol stack that optimizes the performance of TCP
 477         congestion control for high speed network links. It uses a
 478         modeswitch to change the alpha and beta parameters of TCP Reno
 479         based on network conditions and in a way so as to be fair with
 480         other Reno and H-TCP flows.
 481
 482 config TCP_CONG_HSTCP
 483         tristate "High Speed TCP"
 484         depends on EXPERIMENTAL
 485         default n
 486         ---help---
 487         Sally Floyd's High Speed TCP (RFC 3649) congestion control.
 488         A modification to TCP's congestion control mechanism for use
 489         with large congestion windows. A table indicates how much to
 490         increase the congestion window by when an ACK is received.
 491         For more detail see http://www.icir.org/floyd/hstcp.html
 492
 493 config TCP_CONG_HYBLA
 494         tristate "TCP-Hybla congestion control algorithm"
 495         depends on EXPERIMENTAL
 496         default n
 497         ---help---
 498         TCP-Hybla is a sender-side only change that eliminates penalization of
 499         long-RTT, large-bandwidth connections, like when satellite legs are
 500         involved, especially when sharing a common bottleneck with normal
 501         terrestrial connections.
 502
 503 config TCP_CONG_VEGAS
 504         tristate "TCP Vegas"
 505         depends on EXPERIMENTAL
 506         default n
 507         ---help---
 508         TCP Vegas is a sender-side only change to TCP that anticipates
 509         the onset of congestion by estimating the bandwidth. TCP Vegas
 510         adjusts the sending rate by modifying the congestion
 511         window. TCP Vegas should provide less packet loss, but it is
 512         not as aggressive as TCP Reno.
 513
 514 config TCP_CONG_SCALABLE
 515         tristate "Scalable TCP"
 516         depends on EXPERIMENTAL
 517         default n
 518         ---help---
 519         Scalable TCP is a sender-side only change to TCP which uses a
 520         MIMD congestion control algorithm which has some nice scaling
 521         properties, though is known to have fairness issues.
 522         See http://www.deneholme.net/tom/scalable/
 523
 524 config TCP_CONG_LP
 525         tristate "TCP Low Priority"
 526         depends on EXPERIMENTAL
 527         default n
 528         ---help---
 529         TCP Low Priority (TCP-LP), a distributed algorithm whose goal is
 530         to utilize only the excess network bandwidth as compared to the
 531         ``fair share`` of bandwidth as targeted by TCP.
 532         See http://www-ece.rice.edu/networks/TCP-LP/
 533
 534 config TCP_CONG_VENO
 535         tristate "TCP Veno"
 536         depends on EXPERIMENTAL
 537         default n
 538         ---help---
 539         TCP Veno is a sender-side only enhancement of TCP to obtain better
 540         throughput over wireless networks. TCP Veno makes use of state
 541         distinguishing to circumvent the difficult judgment of the packet loss
 542         type. TCP Veno cuts down less congestion window in response to random
 543         loss packets.
 544         See http://www.ntu.edu.sg/home5/ZHOU0022/papers/CPFu03a.pdf
 545
 546 config TCP_CONG_YEAH
 547         tristate "YeAH TCP"
 548         depends on EXPERIMENTAL
 549         select TCP_CONG_VEGAS
 550         default n
 551         ---help---
 552         YeAH-TCP is a sender-side high-speed enabled TCP congestion control
 553         algorithm, which uses a mixed loss/delay approach to compute the
 554         congestion window. It's design goals target high efficiency,
 555         internal, RTT and Reno fairness, resilience to link loss while
 556         keeping network elements load as low as possible.
 557
 558         For further details look here:
 559           http://wil.cs.caltech.edu/pfldnet2007/paper/YeAH_TCP.pdf
 560
 561 config TCP_CONG_ILLINOIS
 562         tristate "TCP Illinois"
 563         depends on EXPERIMENTAL
 564         default n
 565         ---help---
 566         TCP-Illinois is a sender-side modification of TCP Reno for
 567         high speed long delay links. It uses round-trip-time to
 568         adjust the alpha and beta parameters to achieve a higher average
 569         throughput and maintain fairness.
 570
 571         For further details see:
 572           http://www.ews.uiuc.edu/~shaoliu/tcpillinois/index.html
 573
 574 choice
 575         prompt "Default TCP congestion control"
 576         default DEFAULT_CUBIC
 577         help
 578           Select the TCP congestion control that will be used by default
 579           for all connections.
 580
 581         config DEFAULT_BIC
 582                 bool "Bic" if TCP_CONG_BIC=y
 583
 584         config DEFAULT_CUBIC
 585                 bool "Cubic" if TCP_CONG_CUBIC=y
 586
 587         config DEFAULT_HTCP
 588                 bool "Htcp" if TCP_CONG_HTCP=y
 589
 590         config DEFAULT_VEGAS
 591                 bool "Vegas" if TCP_CONG_VEGAS=y
 592
 593         config DEFAULT_WESTWOOD
 594                 bool "Westwood" if TCP_CONG_WESTWOOD=y
 595
 596         config DEFAULT_RENO
 597                 bool "Reno"
 598
 599 endchoice
 600
 601 endif
 602
 603 config TCP_CONG_CUBIC
 604         tristate
 605         depends on !TCP_CONG_ADVANCED
 606         default y
 607
 608 config DEFAULT_TCP_CONG
 609         string
 610         default "bic" if DEFAULT_BIC
 611         default "cubic" if DEFAULT_CUBIC
 612         default "htcp" if DEFAULT_HTCP
 613         default "vegas" if DEFAULT_VEGAS
 614         default "westwood" if DEFAULT_WESTWOOD
 615         default "reno" if DEFAULT_RENO
 616         default "cubic"
 617
 618 config TCP_MD5SIG
 619         bool "TCP: MD5 Signature Option support (RFC2385) (EXPERIMENTAL)"
 620         depends on EXPERIMENTAL
 621         select CRYPTO
 622         select CRYPTO_MD5
 623         ---help---
 624           RFC2385 specifies a method of giving MD5 protection to TCP sessions.
 625           Its main (only?) use is to protect BGP sessions between core routers
 626           on the Internet.
 627
 628           If unsure, say N.
 629