:了解MPLS
提示:
1、单域mpls 架构回顾
2、Option A的设计思想
3、 Option A的命令行
4 、Option B的设计思想
5、Option B的的命令行
5.基于思科华为的配置
提示:这里可以添加计划学习的时间
例如:
1、 周一至周五晚上 7 点—晚上9点
2、 周六上午 9 点-上午 11 点
3、 周日下午 3 点-下午 6 点
内容:
zz
mpls 处理行为:isp内网运行igp 内网启一个mpls ldp。 在pe上创建 instance ,针对每一个独立的客户拿不同的 instance 拿跟客户关联。关联最简单的方式就是。直接拿创建出来的 instance跟客户去做一个绑定。通过这个接口学到的路由都会加入 instance对应的子路由表。并且由于pe之间有 v4的邻居建立,在 instance中 设置路由的rt 和rd ,结合路由前缀掩码+rd和rt 将ipv4路由 转变成 v4 路由 给其它pe传递走。其它pe rt和rd做导入导出操作 将 v4路由还原成ipv4路由,给所连接的另一个pe将路由传递走。 内层标签是pe路由器在通告 v4路由时 自动通告的。收到标签是不能直接使用的,因为没有运行 insatnce的路由器是看不懂的。需要启mpls 针对 v4邻居更新源地址对应的igp路由来做一个标签分发
通过pe的公网接口转发时,会给数据打上标签,一般是两层,,内层标签:通常是 v4的标签 。外层 :去往 v4邻居针对路由下一跳所对应igp路由的标签,中间路由器只会查看外部标签,只会认为是是发给pe的,到达pe后通过php次末跳标签弹出 就只剩内层标签 pe在通过查找 instance路由表来转发数据 。
ps :在日常的网络环境中 单域是不常见的,因为做单域的条件是非常苛刻的,因为需要总部和公司需要连接到同一家运营商 并且需要在同一个as。 所以当我们想实现跨域 跨运营商的架构时,就有4种解决方案:option a b c1 c2.
option a设计思想 :asbr对于一个isp不要将另一个isp当isp来看,当客户来看
r1,r2:asbr,isp之间互联的链路两端的设备。 as域之间asbr不需要运行mpls ,只建立ipv4邻居关系 向对方发送简单的路由
2跟6之间启 v4邻居 1跟5 v4邻居。2和1能接收到 v4路由 但不能加入路由表 加路由表的前提是创建vrf r7 r8做rr.他们不在小数据层面的骨干流量传递路径.
9.10是路由反射器因为在运营商的网络环境中,需要大量路由器运行bgp,需要建立ibgp邻居,没有反射器存在的话,会导致邻居建立臃肿。rr会无视 v4 r5过滤器的机制。 rt过滤器:我通过我的 v4邻居收到路由,全部接收。只能给ibgp邻居。但ibgp可以通过放行的rt传给其他EBGP邻居,和ibgp邻居没联系。
当pe路由器传递 v 4路由的前提是,创建了vrf,vrf里会收集到ce的路由,给邻居发送走,邻居收到会查看路由携带的rt,有没有被 他导入,没有就丢弃。
option a:设计思想,把r2 r1把对方都当ce来看,创建vrf,asbr之间建立ebgp或igp。 彼此之间互传的都是ipv4路由,但都会把对方传过来的路由加入vrf路由表。
缺点:
在r1,r2之间直接建立 v4邻居,不用子接口。直连链路要使用mpls ,创建跟r6相似的vrf 一样的导入导出rt.需要关闭 v4的rt过滤器,只有关闭过滤器会收到 v4才能加入到本地的 v4 bgp 表里,才能把路由从asbr传递给另一端的asbr
缺:虽然标签是没有中断的,在左边isp传输标签时,标签是由r2通告的,在asbr之间传递的时候,标签被交换成对端的asbr所通告的标签,在右边isp中传递标签才是他真正产生这条路由的pe路由器所以分发的标签。不同运营商需要互连,当有第三方运营商中断了,作为穿越路由中断不同运营商的互连,并且他并没有直接连接客户 不可行
r9把路由传递给r10的时候 下一跳会变成r9,r5收到下一跳是r9.会出现问题,流量路由会很奇怪,就算能通rr做路由反射器对cpu 内存已经是极大的消耗了,如果还在流量路径中,需要做流量传递,那么在 控制层面 和数据层面 的双重打击下,rr可能 会hold 不住。 通过next-hop-unchange 来解决 能保证路由下一跳是不变的
基于bgp的解决方案
r6,r5k跟rr建立 v4邻居, asbr之间建立ipv4邻居,rr需要把 v4路由传递给ebgp邻居。 r9 r10需要关闭 v4过滤器。
打三第三重标签:关于asbr的igp路由标签
bgp端到端传递
r2把所属as路由发送给r1, asbr跟rr不用建立内部的igp邻居关系,
通过重分发把igp 路由到bgp 。 带有标签的路由在重分发时,需要标签继承、
ps:华为默认不会把带标签的报文发送给ebgp邻居,收到带标签的ebgp路由发现给ibgp 邻居时标签默认不能 ,pe收到pe的路由 加表的前提是需要有对端32位的主机路由
r7 : 192.168.7.7 192.168.57.7
r8 172.16.8.8 172.16.68.8
其它的:xy.1.1.x xy.1.1.y
r1
router isis
net 49.0001.0000.0000.0001.00
log-adjacency-changes all
is-type level-2-only
metric-style wide
interface range loopback 0 . ethernet 0/1
ip router isis
r3:
router isis
net 49.0001.0000.0000.0003.00
log-adjacency-changes all
is-type level-2-only
metric-style wide
intface range ethernet 0/0-2 ,loopback 0
ip router isi
r5:
router isis
net 49.0001.0000.0000.0005.00
log-adjacency-changes all
is-type level-2-only
metric-style wide
int range e0/0-1 ,loopback
ip router isis
r10:
r5:
router isis
net 49.0001.0000.0000.000A.00
log-adjacency-changes all
is-type level-2-only
metric-style wide
int range e0/0 ,loopback 0
ip router isis
r2
router ospf 110
router-id 2.2.2.2
intface range ethernet 0/0 ,loopback 0
ip ospf 110 area 0
r4:
router ospf 110
router-id 4.4.4.4
int range e0/0-2 ,loopback 0
ip ospf 110 area 0
r6:
router ospf 110
router-id 6.6.6.6
int range e0/1 ,loopback 0
ip ospf 110 area 0
r9:
router ospf 110
router-id 9.9.9.9
int range e0/0,loopback 0
ip ospf 110 area 0
r1,r2,r3,r4,r5,r6 ,r9 ,r10:
ip cef
mpls ip
mpls label protocol ldp
mpls ldp router-id loopback 0 force
内网接口启用 mpls ip
show mpls ldp neighbor
r6
vrf definition a-bj
address-family ipv4 unicast
exit
rd 10:10
address-family ipv4 unicast
route-target both 10:10
exit
int eth0/0
vrf forwarding a-bj
ip add 172.16.68.6 255.255.255.0
ping vrf 172.16.68.8
r8:
router bgp 8
no auto-summary
no synchronization
bgp router-id 8.8.8.8
redistribute connected
neighbor 172.16.68.6 remote-as 8
r6:router bgp 10
no auto-summary
no synchronazation
bgp router-id 6.6.6.6
addres-family ipv4 vrf a-bj
neighbor 172.16.68.8 remote-as 8
neighbor 172.16.68.8 active
r5:
vrf definition a-sh
rd 10:10
address-family ipv4 unicast
route-target 10:10 both
int e0/2
vrf forwarding a-sh
ip add 192.168.57.5 255.255.255.0
router bgp 5
no synchronization
no auto-summary
bgp router-id 5.5.5.5
address-family ipv4 unicast vrf a-sh
neighbor 192.168.57.7
r7
router bgp 7
no synchronization
no auto-summary
bgp router-id 7.7.7.7
neighbor 192.168.57.5 remote-as 5
reditribute connected
show bgp vrf 名 v4 unicast summary 查看邻居
r1 r2创建vrf
r2:
vrf definition a-bj
rd 10:10
address-family ipv4 unicast
router-target both 10:10
interface e0/1.10
encapsulation dot1q 10
vrf forwarding a-bj
ip adddress 10.1.1.2 255.255.255.0
no shut
r6:
no bgp default ipv4-unicast
neighbor 9.9.9.9 remote-as 10
address-family v4 unicast
neighbor 9.9.9.9 active
r9
router bgp 10
no auto-summary
no synchronization
bgp router-id 9.9.9.9
no bgp default ipv4-unicast
bgp router-id 9.9.9.9
neighbor 6.6.6.6 remote-as 10
neighbor 2.2.2.2 remote-as 10
neighbor 6.6.6.6 update-sorce loopback 0
address-family v4 unicast
neighbor 6.6.6.6 active
neighbor 2.2.2.2 active
**neighbor 6.6.6.6 route-reflector-client
neighbor 2.2.2.2 route-reflector-client**
r2:
router bgp 10
no auto-summary
no synchronization
bgp router-id 2.2.2.2
no bgp default ipv4 unicast
neighbor 9.9.9.9 remote-as 10
neighbor 9.9.9.9 update-source loopback 0
address-family v4 unicast
neighbor 9.9.9.9 active
router bgp 10
address-family ipv4 vrf a-bj
neighbor10.1.1.1 remote-as 20
neighbor 10.1.1.1 activate
r10 :
router bgp 20
no auto-summary
no synchronizaiton
bgp router -id 10.10.10.10
no bgp default ipv4-unicast
neighbor 5.5.5.5 remote-as 20
neighbor 1.1.1.1 remote-as 20
neighbor 5.5.5.5 update-source loopback 0
neighbor 1.1.1.1 update-source loopback 0
address-family v4 unicast
neighbor 5.5.5.5 activate
neighbor 5.5.5.5 activate
neighbor 5.5.5.5 route-reflector-client
neighbor 1.1.1.1 route-reflector-client
r5:
router bgp 20
bgp router-id 5.5.5.5
no bgp default ipv4 unicast
neighbor 10.10.10.10 remote-as 20
neighbor 10.10.10.10 update-source loopback 0
address-family v4 unicast
neighbor 10.10.10.10 activate
r1
router bgp 20
no auto-summary
no synchronization
bgp router-id 1.1.1.1
no bgp default ipv4 unicast
neighbor 10.10.10.10 remote-as 20
neighbor 10.10.10.10 update-source loopback 0
address-family v4 unicast
neighbor 10.10.10.10 activate
vrf definition a-sh
rd 10:10
address-family ipv4 unicast
route-target both 10:10
int e0/1.10
encapsulation dot1q 10
vrf forwarding a-sh
ip add 10.1.1.1 255.255.255.0
no shut
r1
router bgp 20
address-family ipv4 vrf a-sh
neighbor 10.1.1.2 remote-as 10
neighbor 10.1.1.2 activate
ar1:
routter bgp 20
neighbor 12.1.1.2 remote-as 10
address-family v4 unicast
neighbor 12.1.1.2 activate
r2
router bgp 10
neighbor 12.1.1.1 remote-as 20
address-family v4 unicast
neighobr 12.1.1.1 activate
但现在是看不到传过来的v4路由的 要关闭rt过滤器
no bgp default route-target filter
清路由 clear bgp v4 unicast * soft
r1
router bgp 20
addres-family v4 unicast
neighbor 10.10.10.10 next-hop-self
clear bgp v4 unicast * soft
r2
router bgp 10
addres-family v4 unicast
neighbor 9.9.9.9 next-hop-self
clear bgp v4 unicast * soft
按常理来说:是通不了的 还需要启用mpls
思科模拟器 当针对物理链路启用v4 邻居时,物理接口会自动启用mpls vrf forwarding
1跟2之间不用建立v4 bgp 建立ipv4 传公网路由
r1
router bgp 20
no neighbor 12.1.1.2 remote-as 10
neighbor 12.1.1.2 remote-as 10
address-family ipv4 unicast
neighbor 12.1.1.2 activate
neighbor 12.1.1.2 send-label
r2
router bgp 10
no neighbor 12.1.1.1 remote-as 20
neighbor 12.1.1.1 remote-as 20
address-family ipv4 unicast
neighbor 12.1.1.1 activate
neighbor 12.1.1.1 send-label
做ipv4 bgp
r6: router bgp 10
nextwork 6.6.6.6 mask 255.255.255.255
address-family ipv4 unicast
neighbor 9.9.9.9 active
neighbor 9.9.9.9 send-label 通告ipv4路由携带标签
r9:
router bgp 10
address-family ipv4 unicast
network 9.9.9.9 mask 255.255.255.255.255
neighbor 6.6.6.6 activate
neighbor 6.6.6.6 route-reflector-client
neighbor 6.6.6.6 send-label
neighbor 2.2.2.2 activate
neighbor 2.2.2.2 route-reflector -client
neighbor 2.2.2.2 send-label
r2:
network 2.2.2 mask 255.255.255.255
address-family ipv4 unicast
neighbor 9.9.9.9 activate
neighbor 9.9.9.9 send-label
去掉v4
r2:
router bgp 10
address-family v4 unicast
no neighbor 9.9.9.9 activate
r9:
router bgp 10
address-family v4 unicast
no neighbor 2.2.2.2 activate
r1
router bgp 20
address-family v4 unicast
no neighbor10.10.10.10 activate
r10
内网ibgp ipv4
r10
router bgp 20
addressfaily ipv4 unicast
network 10.10.10.10 mask 255.255.255.255
neighbor 5.5.5.5 activate
neighbor 5.5.5.5 route-reflector-client
neighbor 5.5.5.5 send-label
neighbor 1.1.1.1activate
neighbor 1.1.1.1route-reflector-client
neighbor 1.1.1.1send-label
r5:
router bgp 20
address-family ipv4 unicast
neighbor 10.10.10.10 activate
neighbor 10.10.10.10 send-label
r1
router bgp 20
address-family ipv4 unicast
neighbor 10.10.10.10 activate
neighbor 10.10.10.10 send-label
1对10 2对9 指next-self-self 因为在v4子地址簇中删除x.x.x.x activate 相关的命令都会被删除
9和10之间建立v4 这里
r9
router bgp 10
no bgp default ipv4 unicast
no bgp default route-target filter
neighobr 10.10.10.10 update-source loopback 0
neighbor 10.10.10.10 ebgp-multihop
addresss-family v4 unicast
neighbor 10.10.10.10 next-hop-unchange
r10:
router bgp 20
neighbor 9.9.9.9 remote-as 10
neighbor 9.9.9.9 ebgp-multihop
no bgp default route-target filter
addresss-family v4 unicast
neighbor 9.9.9.9 activate
neighbor 9.9.9.9 next-hop-unchange
show 标签
show mpls forwarding-table
show bgp ipv4 unicast labels
9和6 5和10 的ipv4 bgp 去掉
r2只需要跟r1 建立 ipv4 bgp路由
r9
router bgp 10
no neighbor 2.2.2.2
address-family ipv4 unicast
no network 9.9.9.9 mask 255.255.255.255
no neighbor 6.6.6.6 activate
no neighbor 2.2.2.2 activate
r6:
router bgp 10
address-family ipv4 unicast
no network 6.6.6.6 mask 255.255.255.255
no neighbor 9.9.9.9 activate
r2:
route bgp 10
address-family ipv4 unicast
no neighbor 9.9.9.9 activate
no network 2.2.2.2 mask 255.255.255.255
r10
router bgp 20
no neighbor 1.1.1.1
adress-family ipv4 unicast
no network 10.10.10.10 mask 255.255.255.255
no neighbor 5.5.5.5 activate
r1:
router bgp 20
no neighbor 10.10.10.10 // 直接去掉所有关于10.10.10.10 的配置
address-family ipv4 unicast
no network 1.1.1.1 mask 255.255.255.255
no neighbor 10.10.10.10 activate
r5
router bgp 20
address-family ipv4 unicast
no netwrok 5.5.5.5 mask 255.255.255.255
no neighbor 10.10.10.10 activate
r2做route-map
r2
ip prefix-list ccie seq 10 permit 6.6.6.6/32
ip prefix-list ccie sq1 20 permit 9.9.9.9/32
route-map o-2-b permit 10
match ip address prefix-list ccie
router bgp 10
address-family ipv4 unicast
redistribute ospf 110 route-map o-2-b
r1
ip prefix-list ccie seq 10 permit 5.5.5.5/32
ip prefix-list ccie sq1 20 permit 10.10.10.10/32
route-map i-2-b permit 10
match ip address prefix-list ccie
router bgp 20
address-family ipv4 unicast
redistribute isis level-2 route-map i-2-b