Home > Support > Knowledge Base > Knowledge Base Article

Digi TransPort BGP configuration file syntax

DESCRIPTION
     BGP on TransPort routers implements the Border Gateway Protocol version 4 as
     described in RFC 4271. The following text is from the OpenBSD man page for
     bgpd.conf.  The SarOS operating system implementation of BGP is almost
     identical to the OpenBSD version but some of the features described may not
     be supported.

SECTIONS
     The BGP config file is divided into five main sections.

     Macros
           User-defined variables may be defined and used later, simplifying
           the configuration file.

     Global Configuration
           Global settings for BGP.

     Routing Domain Configuration
           The definition and properties for BGP MPLS VPNs are set in this
           section.

     Neighbors and Groups
           BGP establishes sessions with neighbors.  The neighbor
           definition and properties are set in this section, as well as
           grouping neighbors for the ease of configuration.

     Filter
           Filter rules for incoming and outgoing UPDATES.

     With the exception of macros, the sections should be grouped and appear
     in the BGP config file in the order shown above.

     The current line can be extended over multiple lines using a backslash
     (`\').  Comments can be put anywhere in the file using a hash mark (`#'),
     and extend to the end of the current line.  Care should be taken when
     commenting out multi-line text: the comment is effective until the end of
     the entire block.

     Argument names not beginning with a letter, digit, or underscore must be
     quoted.

MACROS
     Macros can be defined that will later be expanded in context.  Macro
     names must start with a letter, digit, or underscore, and may contain any
     of those characters.  Macro names may not be reserved words (for example,
     AS, neighbor, or group).  Macros are not expanded inside quotes.

     For example:

           peer1="1.2.3.4"
           neighbor $peer1 {
                   remote-as 65001
           }

GLOBAL CONFIGURATION
     There are quite a few settings that affect the global operation of BGP.

     AS as-number [as-number]
             Set the local autonomous system number to [as-number].  If the
             first AS number is a 4-byte AS it is possible to specify a
             secondary 2-byte AS number which is used for neighbors which do
             not support 4-byte AS numbers.  The default for the secondary AS
             is 23456.

             The AS numbers are assigned by local RIRs, such as:

             AfriNIC   for Africa
             APNIC     for Asia Pacific
             ARIN      for North America and parts of the Caribbean
             LACNIC    for Latin America and the Caribbean
             RIPE NCC  for Europe, the Middle East, and parts of Asia

             For example:

                   AS 65001

             sets the local AS to 65001.

             The AS numbers 64512 - 65534 are designated for private use.  The
             AS number 23456 is a specially designated Autonomous System
             Number and should not be used.  4-byte AS numbers are specified
             as two numbers separated by a dot (ASDOT format), for example:

                   AS 3.10

             or as a large number (ASPLAIN format), for example:

                   AS 196618

     connect-retry seconds
             Set the number of seconds before retrying to open a connection.
             This timer should be sufficiently large in EBGP configurations.
             The default is 120 seconds.

     fib-update (yes|no)
             If set to no, do not update the Forwarding Information Base,
             a.k.a. the kernel routing table.  The default is yes.

     holdtime seconds
             Set the holdtime in seconds.  The holdtime is reset to its
             initial value every time either a KEEPALIVE or an UPDATE message
             is received from the neighbor.  If the holdtime expires the
             session is dropped.  The default is 90 seconds.  Neighboring
             systems negotiate the holdtime used when the connection is
             established in the OPEN messages.  Each neighbor announces its
             configured holdtime; the smaller one is then agreed upon.

     holdtime min seconds
             The minimal accepted holdtime in seconds.  This value must be
             greater than or equal to 3.

     listen on address
             Specify the local IP address BGP should listen on.

                   listen on 127.0.0.1

     log updates
             Log received and sent updates.

     network address/prefix [set ...]
     network (inet|inet6) static [set ...]
     network (inet|inet6) connected [set ...]
             Announce the specified network as belonging to our AS.  If set to
             connected, routes to directly attached networks will be
             announced.  If set to static, all static routes will be
             announced.

                   network 192.168.7.0/24

             It is possible to set default AS path attributes per network
             statement:

                   network 192.168.7.0/24 set localpref 220

             See also the ATTRIBUTE SET section.

     nexthop qualify via (bgp|default)
             If set to bgp, BGP may use BGP routes to verify nexthops.  If
             set to default, BGP may use the default route to verify
             nexthops.  By default BGP will only use static routes or routes
             added by other routing protocols like OSPF.

     rde med compare (always|strict)
             If set to always, the MULTI_EXIT_DISC attributes will always be
             compared.  The default is strict, where the metric is only
             compared between peers belonging to the same AS.

     rde rib name [no evaluate]
     rde rib name [rtable number]
             Create an additional RIB named name.  It is possible to disable
             the decision process per RIB with the no evaluate flag.  If a
             rtable is specified, routes will be exported to the given kernel
             routing table.  Currently the routing table must belong to the
             default routing domain and nexthop verification happens on table
             0.  Routes in the specified table will not be considered for
             nexthop verification.  Adj-RIB-In and Loc-RIB are created
             automatically and used as default.

     rde route-age (ignore|evaluate)
             If set to evaluate, the best path selection will not only be
             based on the path attributes but also on the age of the route,
             giving preference to the older, typically more stable, route.  In
             this case the decision process is no longer deterministic.  The
             default is ignore.

     route-collector (yes|no)
             If set to yes, the route selection process is turned off.  The
             default is no.

     router-id address
             Set the router ID to the given IP address, which must be local to
             the machine.

                   router-id 10.0.0.1

             If not given, the BGP ID is determined as the biggest IP address
             assigned to the local machine.

     rtable number
             Work with the given kernel routing table instead of the default
             table, 0.  Note that table 0 is used for nexthop verification.
             Routes in the specified table will not be considered for nexthop
             verification.  This is the same as using the following syntax:

                   rde rib Loc-RIB rtable number

     transparent-as (yes|no)
             If set to yes, AS paths to EBGP neighbors are not prepended with
             their own AS.  The default is no.

ROUTING DOMAIN CONFIGURATION
     BGP supports the setup and distribution of Virtual Private Networks.
     It is possible to import and export prefixes between routing domains.
     Each routing domain is specified by an rdomain section, which allows
     properties to be set specifically for that rdomain:

           rdomain 1 {
                   descr "a rdomain"
                   rd 65002:1
                   import-target rt 65002:42
                   export-target rt 65002:42
                   network 192.168.1/24
                   depend on mpe0
           }

     There are several routing domain properties:

     depend on interface
             Routes added to the rdomain will use this interface as the
             outgoing interface.  Normally this will be an MPLS Provider Edge,
             mpe(4), interface that is part of the rdomain.  Local networks
             will be announced with the MPLS label specified on the interface.

     descr description
             Add a description.  The description is used when logging but has
             no further meaning to BGP.

     export-target subtype as-number:local
     export-target subtype IP:local
             Specify an extended community which will be attached to announced
             networks.  More than one export-target can be specified.  See
             also the ATTRIBUTE SET section for further information about the
             encoding.  The subtype should be set to rt for best compatibility
             with other implementations.

     fib-update (yes|no)
             If set to no, do not update the Forwarding Information Base,
             a.k.a. the kernel routing table.  The default is yes.

     import-target subtype as-number:local
     import-target subtype IP:local
             Only prefixes matching one of the specified import-targets will
             be imported into the rdomain.  More than one import-target can be
             specified.  See also the ATTRIBUTE SET section for further
             information about the encoding of extended communities.  The
             subtype should be set to rt for best compatibility with other
             implementations.

     network arguments ...
             Define which networks should be exported into this VPN.  See also
             the nexthop section in GLOBAL CONFIGURATION for further
             information about the arguments.

     rd as-number:local
     rd IP:local
             The sole purpose of the Route Distinguisher rd is to ensure that
             possible common prefixes are destinct between VPNs.  The rd is
             neither used to identify the origin of the prefix nor to control
             into which VPNs the prefix is distributed to.  The as-number or
             IP of a rd should be set to a number or IP that was assigned by
             an appropriate authority.  Whereas local can be chosen by the
             local operator.

NEIGHBORS AND GROUPS
     BGP establishes TCP connections to other BGP speakers called
     neighbors.  Each neighbor is specified by a neighbor section, which
     allows properties to be set specifically for that neighbor:

           neighbor 10.0.0.2 {
                   remote-as 65002
                   descr "a neighbor"
           }

     Multiple neighbors can be grouped together by a group section.  Each
     neighbor section within the group section inherits all properties from
     its group:

           group "peering AS65002" {
                   remote-as 65002
                   neighbor 10.0.0.2 {
                           descr "AS65002-p1"
                   }
                   neighbor 10.0.0.3 {
                           descr "AS65002-p2"
                   }
           }

     Instead of the neighbor's IP address, an address/netmask pair may be
     given:

           neighbor 10.0.0.0/8

     In this case, the neighbor specification becomes a template, and if a
     neighbor connects from an IP address within the given network, the
     template is cloned, inheriting everything from the template but the
     remote address, which is replaced by the connecting neighbor's address.
     With a template specification it is valid to omit remote-as; BGP will
     then accept any AS the neighbor presents in the OPEN message.

     There are several neighbor properties:

     announce (all|none|self|default-route)
             If set to none, no UPDATE messages will be sent to the neighbor.
             If set to default-route, only the default route will be announced
             to the neighbor.  If set to all, all generated UPDATE messages
             will be sent to the neighbor.  This is usually used for transit
             AS's and IBGP peers.  The default value for EBGP peers is self,
             which limits the sent UPDATE messages to announcements of the
             local AS.  The default for IBGP peers is all.

     announce (IPv4|IPv6) (none|unicast|vpn)
             For the given address family, control which subsequent address
             families (at the moment, only none, which disables the
             announcement of that address family, unicast, and vpn, which
             allows the distribution of BGP MPLS VPNs, are supported) are
             announced during the capabilities negotiation.  Only routes for
             that address family and subsequent address family will be
             announced and processed.

     announce as-4byte (yes|no)
             If set to no, the 4-byte AS capability is not announced and so
             native 4-byte AS support is disabled.  The default is yes.

     announce capabilities (yes|no)
             If set to no, capability negotiation is disabled during the
             establishment of the session.  This can be helpful to connect to
             old or broken BGP implementations.  The default is yes.

     announce refresh (yes|no)
             If set to no, the route refresh capability is not announced.  The
             default is yes.

     announce restart (yes|no)
             If set to yes, the graceful restart capability is announced.
             Currently only the End-of-RIB marker is supported and announced
             by the restart capability.  The default is no.

     demote group
             Increase the demotion counter on the given interface
             group, when the session is not in state ESTABLISHED.
             The demotion counter will be increased as soon as
             BGP starts and decreased 60 seconds after the session went to
             state ESTABLISHED.  For neighbors added at runtime, the demotion
             counter is only increased after the session has been ESTABLISHED
             at least once before dropping.

     depend on interface
             The neighbor session will be kept in state IDLE as long as
             interface reports no link. 
             The state of the network interfaces on the system can be viewed
             using the 'route print' command.

     descr description
             Add a description.  The description is used when logging neighbor
             events, in status reports, for specifying neighbors, etc., but
             has no further meaning to BGP.

     down    Do not start the session when bgpd comes up but stay in IDLE.

     dump (all|updates) (in|out) file [timeout]
             Do a peer specific MRT dump.  Peer specific dumps are limited to
             all and updates.  See also the dump section in GLOBAL
             CONFIGURATION.

     enforce neighbor-as (yes|no)
             If set to yes, AS paths whose leftmost AS is not equal to the
             remote AS of the neighbor are rejected and a NOTIFICATION is sent
             back.  The default value for IBGP peers is no otherwise the
             default is yes.

     holdtime seconds
             Set the holdtime in seconds.  Inherited from the global
             configuration if not given.

     holdtime min seconds
             Set the minimal acceptable holdtime.  Inherited from the global
             configuration if not given.

     local-address address
             When BGP initiates the TCP connection to the neighbor system,
             it normally does not bind to a specific IP address.  If a
             local-address is given, BGP binds to this address first.

     max-prefix number [restart number]
             Terminate the session after number prefixes have been received
             (no such limit is imposed by default).  If restart is specified,
             the session will be restarted after number minutes.

     multihop hops
             Neighbors not in the same AS as the local BGP normally have
             to be directly connected to the local machine.  If this is not
             the case, the multihop statement defines the maximum hops the
             neighbor may be away.

     passive
             Do not attempt to actively open a TCP connection to the neighbor
             system.

     remote-as as-number
             Set the AS number of the remote system.

     rib name
             Bind the neighbor to the specified RIB.

     route-reflector [address]
             Act as an RFC 4456 route-reflector for this neighbor.  An
             optional cluster ID can be specified; otherwise the BGP ID will
             be used.

     set attribute ...
             Set the AS path attributes to some default per neighbor or group
             block:

                   set localpref 300

             See also the ATTRIBUTE SET section.  Set parameters are applied
             to the received prefixes; the only exceptions are prepend-self,
             nexthop no-modify and nexthop self.  These sets are rewritten
             into filter rules.

     softreconfig (in|out) (yes|no)
             Turn soft reconfiguration on or off for the specified direction.
             If soft reconfiguration is turned on, filter changes will be
             applied on configuration reloads.  If turned off, a BGP session
             needs to be cleared to apply the filter changes.  Enabling
             softreconfig in will raise the memory requirements of BGP
             because the unmodified AS path attributes need to be stored as
             well.  The default is yes.

     tcp md5sig password secret
     tcp md5sig key secret
             Enable TCP MD5 signatures per RFC 2385.  The shared secret can
             either be given as a password or hexadecimal key.

                   tcp md5sig password mekmidasdigoat
                   tcp md5sig key deadbeef

     transparent-as (yes|no)
             If set to yes, AS paths to EBGP neighbors are not prepended with
             their own AS.  The default is inherited from the global
             transparent-as setting.

     ttl-security (yes|no)
             Enable or disable ttl-security.  When enabled, outgoing packets
             are sent using a TTL of 255 and a check is made against an
             incoming packet's TTL.  For directly connected peers, incoming
             packets are required to have a TTL of 255, ensuring they have not
             been routed.  For multihop peers, incoming packets are required
             to have a TTL of 256 minus multihop distance, ensuring they have
             not passed through more than the expected number of hops.  The
             default is no.

FILTER
     BGP has the ability to allow and deny UPDATES based on prefix or AS
     path attributes.  In addition, UPDATES may also be modified by filter
     rules.

     For each UPDATE processed by the filter, the filter rules are evaluated
     in sequential order, from first to last.  The last matching allow or deny
     rule decides what action is taken.

     The following actions can be used in the filter:

     allow     The UPDATE is passed.

     deny      The UPDATE is blocked.

     match     Apply the filter attribute set without influencing the filter
               decision.

PARAMETERS
     The rule parameters specify the UPDATES to which a rule applies.  An
     UPDATE always comes from, or goes to, one neighbor.  Most parameters are
     optional, but each can appear at most once per rule.  If a parameter is
     specified, the rule only applies to packets with matching attributes.

     as-type as-number
             This rule applies only to UPDATES where the AS path matches.  The
             as-number is matched against a part of the AS path specified by
             the as-type.  as-number may be set to neighbor-as, which is
             expanded to the current neighbor remote AS number.  as-type is
             one of the following operators:

             AS           (any part)
             peer-as      (leftmost AS number)
             source-as    (rightmost AS number)
             transit-as   (all but the rightmost AS number)

             Multiple as-number entries for a given type or as-type as-number
             entries may also be specified, separated by commas or whitespace,
             if enclosed in curly brackets:

                   deny from any AS { 1, 2, 3 }
                   deny from any { AS 1, source-as 2, transit-as 3 }
                   deny from any { AS { 1, 2, 3 }, source-as 4, transit-as 5 }

     community as-number:local
     community name
             This rule applies only to UPDATES where the community path
             attribute is present and matches.  Communities are specified as
             as-number:local, where as-number is an AS number and local is a
             locally significant number between zero and 65535.  Both
             as-number and local may be set to `*' to do wildcard matching.
             Alternatively, well-known communities may be given by name
             instead and include NO_EXPORT, NO_ADVERTISE, NO_EXPORT_SUBCONFED,
             and NO_PEER.  Both as-number and local may be set to neighbor-as,
             which is expanded to the current neighbor remote AS number.

     ext-community subtype as-number:local
     ext-community subtype IP:local
     ext-community subtype numvalue
             This rule applies only to UPDATES where the extended community
             path attribute is present and matches.  Extended Communities are
             specified by a subtype and normally two values, a globally unique
             part (e.g. the AS number) and a local part.  See also the
             ATTRIBUTE SET section for further information about the encoding.

     (from|to) peer
             This rule applies only to UPDATES coming from, or going to, this
             particular neighbor.  This parameter must be specified.  peer is
             one of the following:

             any          Any neighbor will be matched.
             address      Neighbors with this address will be matched.
             group descr  Neighbors in this group will be matched.

             Multiple peer entries may also be specified, separated by commas
             or whitespace, if enclosed in curly brackets:

                   deny from { 128.251.16.1, 251.128.16.2, group hojo }

     (inet|inet6)
             This rule applies only to routes matching the stated address
             family.  The address family needs to be set only in rules that
             use prefixlen without specifying a prefix beforehand.

     max-as-len len
             This rule applies only to UPDATES where the AS path has more than
             len elements.

     max-as-seq len
             This rule applies only to UPDATES where a single AS number is
             repeated more than len times.

     prefix address/len
             This rule applies only to UPDATES for the specified prefix.

             Multiple address/len entries may be specified, separated by
             commas or whitespace, if enclosed in curly brackets:

                   deny from any prefix { 192.168.0.0/16, 10.0.0.0/8 }

             Multiple lists can also be specified, which is useful for macro
             expansion:

                   good="{ 192.168.0.0/16, 172.16.0.0/12, 10.0.0.0/8 }"
                   bad="{ 224.0.0.0/4, 240.0.0.0/4 }"
                   ugly="{ 127.0.0.1/8, 169.254.0.0/16 }"

                   deny from any prefix { $good $bad $ugly }

     prefixlen range
             This rule applies only to UPDATES for prefixes where the
             prefixlen matches.  Prefix length ranges are specified by using
             these operators:

                   =       (equal)
                   !=      (unequal)
                   <       (less than)
                   <=      (less than or equal)
                   >       (greater than)
                   >=      (greater than or equal)
                   -       (range including boundaries)
                   ><      (except range)

             >< and - are binary operators (they take two arguments).  For
             instance, to match all prefix lengths >= 8 and <= 12, and hence
             the CIDR netmasks 8, 9, 10, 11 and 12:

                   prefixlen 8-12

             Or, to match all prefix lengths < 8 or > 12, and hence the CIDR
             netmasks 0-7 and 13-32:

                   prefixlen 8><12

             prefixlen can be used together with prefix.

             This will match all prefixes in the 10.0.0.0/8 netblock with
             netmasks longer than 16:

                   prefix 10.0.0.0/8 prefixlen > 16

     quick   If an UPDATE matches a rule which has the quick option set, this
             rule is considered the last matching rule, and evaluation of
             subsequent rules is skipped.

     rib name
             Apply rule only to the specified RIB.  This only applies for
             received updates, so not for rules using the to peer parameter.

     set attribute ...
             All matching rules can set the AS path attributes to some
             default.  The set of every matching rule is applied, not only the
             last matching one.  See also the following section.

ATTRIBUTE SET
     AS path attributes can be modified with set.

     set can be used on network statements, in neighbor or group blocks, and
     on filter rules.  Attribute sets can be expressed as lists.

     The following attributes can be modified:

     community [delete] as-number:local
     community [delete] name
             Set or delete the COMMUNITIES AS path attribute.  Communities are
             specified as as-number:local, where as-number is an AS number and
             local is a locally-significant number between zero and 65535.
             Alternately, well-known communities may be specified by name:
             NO_EXPORT, NO_ADVERTISE, NO_EXPORT_SUBCONFED, or NO_PEER.

     ext-community [delete] subtype as-number:local
     ext-community [delete] subtype IP:local
     ext-community [delete] subtype numvalue
             Set or delete the Extended Community AS path attribute.  Extended
             Communities are specified by a subtype and normally two values, a
             globally unique part (e.g. the AS number) and a local part.  The
             type is selected depending on the encoding of the global part.
             Two-octet AS Specific Extended Communities and Four-octet AS
             Specific Extended Communities are encoded as as-number:local.
             Four-octet encoding is used if the as-number is bigger then 65535
             or if the AS_DOT encoding is used.  IPv4 Address Specific
             Extended Communities are encoded as IP:local.  Opaque Extended
             Communities are encoded with a single numeric value.  Currently
             the following subtypes are supported:

                   rt       Route Target
                   soo      Source of Origin
                   odi      OSPF Domain Identifier
                   ort      OSPF Route Type
                   ori      OSPF Router ID
                   bdc      BGP Data Collection

             Not all type and subtype value pairs are allowed by IANA and the
             parser will ensure that no invalid combination is created.

     localpref number
             Set the LOCAL_PREF AS path attribute.  If number starts with a
             plus or minus sign, LOCAL_PREF will be adjusted by adding or
             subtracting number; otherwise it will be set to number.  The
             default is 100.

     med number
     metric number
             Set the MULTI_EXIT_DISC AS path attribute.  If number starts with
             a plus or minus sign, MULTI_EXIT_DISC will be adjusted by adding
             or subtracting number; otherwise it will be set to number.

     origin (igp|egp|incomplete)
             Set the ORIGIN AS path attribute to mark the source of this route
             as being injected from an igp protocol, an egp protocol or being
             an aggregated route.

     nexthop (address|blackhole|reject|self|no-modify)
             Set the NEXTHOP AS path attribute to a different nexthop address
             or use blackhole or reject routes.  If set to no-modify, the
             nexthop attribute is not modified.  Unless set to self, the
             nexthop is left unmodified for IBGP sessions.  self forces the
             nexthop to be set to the local interface address.

                   set nexthop 192.168.0.1
                   set nexthop blackhole
                   set nexthop reject
                   set nexthop no-modify
                   set nexthop self

     prepend-neighbor number
             Prepend the neighbor's AS number times to the AS path.

     prepend-self number
             Prepend the local AS number times to the AS path.

     rtlabel label
             Add the prefix to the kernel routing table with the specified
             label.

     weight number
             The weight is used to tip prefixes with equally long AS paths in
             one or the other direction.  A prefix is weighed at a very late
             stage in the decision process.  If number starts with a plus or
             minus sign, the weight will be adjusted by adding or subtracting
             number; otherwise it will be set to number.  Weight is a local
             non-transitive attribute and a bgpd-specific extension.  For
             prefixes with equally long paths, the prefix with the larger
             weight is selected.

Contact a Digi expert and get started today! CONTACT US

Desktop Site