基于Cisco技术的MPLS ××× 实现[一]
MPLS的出现是因为现有的路由选择以及转发技术无法应对越来越大的internet路由表. ATM技术相对于IP技术的优势在于ATM信元转发的快速性, IP技术相对于ATM技术的优势在于IP技术的实现容易性和灵活性. 虽然IP是可以承载在ATM之上的, 但是由于大容量的ATM交互矩阵的高成本以及每个53字节的ATM信元都会有8字节的ATM头带来的转发低效率, 标签转发技术适时的出现了.
下图是MPLS(多协议标签转发)技术的一个应用图表.
 
 
 
在这个连载中, 我会依次讲解这些基于标签转发技术的应用细节和CISCO实现.
MPLS是一个2.5层技术, 可以参考下面的基于帧模式的一个协议字段图. MPLS的lable是加在帧头和IP头之间的. 也可以插入多标签针对不同的MPLS应用(比如MPLS-×××应用, 流量工程, 基于MPLS-×××的流量工程等).
 
 
多标签情况如下图.
 
由于标签的引入带来的MTU扩大问题. 标签的长度是4字节, 根据CISCO建议, 系统MTU设置为1524如果是以太网封装的话, 多的24字节可以承载6个标签的插入, 一般情况下6个标签适应于几乎所有的应用了.
下面6台路由器组成的一个拓扑环境就是这一部分的实验图.
 
R1(IPS1)和R2(Border1)之间是EBGP邻居, R2(Border1)和R5(Border2)之间是IBGP邻居, R5(Border2)和R6(IPS2)之间是EBGP邻居. R2(Border1), R3(Core1), R4(Core2), R5(Border2)之间跑IGP协议, 图中是RIP, 后面的详细配置是OSPF. R2(Border1), R3(Core1), R4(Core2), R5(Border2)的相邻接口都启用MPLS并建立MPLS邻居.
R1(IPS1)宣告(network)一条R1的32环回路由给R2(Border1). R6(IPS2)宣告(network)一条R6的32环回路由给R5(Border2). Core1和Core2并没有跑BGP协议. 根据BGP原理, R2会从R1学习到一条EBGP路由,并且传递给IBGP邻居R5. R5会将这条IBGP路由传递给EBGP邻居R6. 反之亦然.
根据BGP原理, Core1和Core2并没有跑BGP,也没有将BGP重分发到IGP中, 因此如果从R1 ping 学习到的R6的环回口地址(BGP 路由), 是不会通的, 因为BGP路由黑洞. Core1会丢弃去往R6的包因为路由不可达. 
但是, 我们却可以在R1上(通过以宣告的环回口做源地址)traceroute或者ping学习到的EBGP路由(及是R6的环回口地址), 发现是可以通的.  这个就是MPLS转发起的作用了.
首先我们看一下LDP/TDP协议也就是标签分发协议. TDP是CISCO的私有协议,但是原理基本一致.
LDP/TDP的工作主要分下面几个部分: (1) 邻居发现阶段, 这一阶段通过基于UDP的Hello包发往组播地址224.0.0.2来发现链路上的邻居, 如果丢失三个hello包宣告邻居失效. 也可以通过配置peer来发送单播hello来建立邻居, 由于ldp session是基于TCP的, 所以ldp可以跨越网络建立session, 并不需要直连.
(2) TCP连接的建立 (3) LDP Session的建立 (4) 标签的分发
同一台路由器上不同接口可以分别启用ldp或者tdp协议, 同一接口上也可以启用ldp和tdp双协议.
 
 
下面, 用一个实际的实现例子来解析MPLS的原理. 拓扑使用上面IPS1,IPS2,Border1,Border2,Core1,Core2组成的实验图.
路由器连接如下:
Router1 E1/0 <----> Router2 E1/0
Router2 E1/1 <----> Router3 E1/1
Router3 E1/2 <----> Router4 E1/2
Router4 E1/3 <----> Router5 E1/3
Router5 E1/0 <----> Router6 E1/0
路由器配置如下:
r1#sh run
Building configuration...
Current configuration : 1061 bytes
!
version 12.3
service timestamps debug datetime msec
service timestamps log datetime msec
no service password-encryption
!
hostname r1
!
boot-start-marker
boot-end-marker
!
!
no aaa new-model
ip subnet-zero
!
!
no ip domain lookup
!
ip cef
!
!
interface Loopback0
 ip address 1.1.1.1 255.255.255.255
!
interface FastEthernet0/0
 no ip address
 shutdown
 duplex half
!        
interface Ethernet1/0
 ip address 172.16.1.1 255.255.0.0
 duplex half
!
interface Ethernet1/1
 no ip address
 shutdown
 duplex half
!
interface Ethernet1/2
 no ip address
 shutdown
 duplex half
!
interface Ethernet1/3
 no ip address
 duplex half
 tag-switching ip
!
router bgp 200
 no synchronization
 bgp router-id 10.10.10.10
 bgp log-neighbor-changes
 network 1.1.1.1 mask 255.255.255.255
 neighbor 172.16.2.2 remote-as 100
 no auto-summary
!
ip classless
no ip http server
no ip http secure-server
!
gatekeeper
 shutdown
!
!
line con 0
 exec-timeout 0 0
 logging synchronous
 stopbits 1
line aux 0
 stopbits 1
line vty 0 4
 login
!
!
end
r1#
r1#
 
r2#sh run
Building configuration...
Current configuration : 1374 bytes
!
version 12.3
service timestamps debug datetime msec
service timestamps log datetime msec
no service password-encryption
!
hostname r2
!
boot-start-marker
boot-end-marker
!
!
no aaa new-model
ip subnet-zero
!
!
no ip domain lookup
!
ip cef
mpls label range 200 299
mpls label protocol ldp
tag-switching tdp router-id Loopback0
!
!
interface Loopback0
 ip address 2.2.2.2 255.255.255.255
!
interface FastEthernet0/0
 no ip address
 shutdown
 duplex half
!
interface Ethernet1/0
 ip address 172.16.2.2 255.255.0.0
 duplex half
!
interface Ethernet1/1
 ip address 10.2.2.2 255.255.255.0
 duplex half
 mpls label protocol ldp
 tag-switching ip
!
interface Ethernet1/2
 no ip address
 shutdown
 duplex half
!
interface Ethernet1/3
 no ip address
 shutdown
 duplex half
!        
router ospf 2
 log-adjacency-changes
 network 2.2.2.2 0.0.0.0 area 0
 network 10.2.2.0 0.0.0.255 area 0
 network 172.16.0.0 0.0.255.255 area 0
!
router bgp 100
 no synchronization
 bgp router-id 20.20.20.20
 bgp log-neighbor-changes
 neighbor 5.5.5.5 remote-as 100
 neighbor 5.5.5.5 update-source Loopback0
 neighbor 172.16.1.1 remote-as 200
 no auto-summary
!
ip classless
no ip http server
no ip http secure-server
!
!
gatekeeper
 shutdown
!
!
line con 0
 exec-timeout 0 0
 logging synchronous
 stopbits 1
line aux 0
 stopbits 1
line vty 0 4
 login
!
!
end
r2#
 
r3#sh run 
Building configuration...
Current configuration : 1147 bytes
!
version 12.3
service timestamps debug datetime msec
service timestamps log datetime msec
no service password-encryption
!
hostname r3
!
boot-start-marker
boot-end-marker
!
!
no aaa new-model
ip subnet-zero
!
!
no ip domain lookup
!
ip cef
mpls label range 300 399
mpls label protocol ldp
tag-switching tdp router-id Loopback0
!        
!
interface Loopback0
 ip address 3.3.3.3 255.255.255.255
!
interface FastEthernet0/0
 no ip address
 shutdown
 duplex half
!        
interface Ethernet1/0
 no ip address
 shutdown
 duplex half
!
interface Ethernet1/1
 ip address 10.2.2.3 255.255.255.0
 duplex half
 tag-switching ip
!
interface Ethernet1/2
 ip address 10.3.3.3 255.255.255.0
 duplex half
 tag-switching ip
!
interface Ethernet1/3
 no ip address
 shutdown
 duplex half
!
router ospf 3
 log-adjacency-changes
 network 3.3.3.3 0.0.0.0 area 0
 network 10.2.2.0 0.0.0.255 area 0
 network 10.3.3.0 0.0.0.255 area 0
!
ip classless
no ip http server
no ip http secure-server
!
!
gatekeeper
 shutdown
!
!
line con 0
 exec-timeout 0 0
 logging synchronous
 stopbits 1
line aux 0
 stopbits 1
line vty 0 4
 login
!
!
end
r3#
 
r4#sh run
Building configuration...
Current configuration : 1173 bytes
!
version 12.3
service timestamps debug datetime msec
service timestamps log datetime msec
no service password-encryption
!
hostname r4
!
boot-start-marker
boot-end-marker
!
!
no aaa new-model
ip subnet-zero
!
!
no ip domain lookup
!
ip cef
mpls label range 400 499
tag-switching tdp router-id Loopback0
!
!
interface Loopback0
 ip address 4.4.4.4 255.255.255.255
!
interface FastEthernet0/0
 no ip address
 shutdown
 duplex half
!
interface Ethernet1/0
 no ip address
 shutdown
 duplex half
!
interface Ethernet1/1
 no ip address
 shutdown
 duplex half
!
interface Ethernet1/2
 ip address 10.3.3.4 255.255.255.0
 duplex half
  mpls label protocol ldp
 tag-switching ip
!
interface Ethernet1/3
 ip address 10.4.4.4 255.255.255.0
 duplex half
  mpls label protocol tdp
 tag-switching ip
!
router ospf 4
 log-adjacency-changes
 network 4.4.4.4 0.0.0.0 area 0
 network 10.3.3.0 0.0.0.255 area 0
 network 10.4.4.0 0.0.0.255 area 0
!
ip classless
no ip http server
no ip http secure-server
!
!
gatekeeper
 shutdown
!
!        
line con 0
 exec-timeout 0 0
 logging synchronous
 stopbits 1
line aux 0
 stopbits 1
line vty 0 4
 login
!
!
end
r4#
 
r5#sh run
Building configuration...
Current configuration : 1353 bytes
!
version 12.3
service timestamps debug datetime msec
service timestamps log datetime msec
no service password-encryption
!
hostname r5
!
boot-start-marker
boot-end-marker
!
!
no aaa new-model
ip subnet-zero
!
!
no ip domain lookup
!
ip cef
mpls label range 500 599
tag-switching tdp router-id Loopback0
!
!
interface Loopback0
 ip address 5.5.5.5 255.255.255.255
!
interface FastEthernet0/0
 no ip address
 shutdown
 duplex half
!
interface Ethernet1/0
 ip address 192.168.5.5 255.255.0.0
 duplex half
!
interface Ethernet1/1
 no ip address
 shutdown
 duplex half
!
interface Ethernet1/2
 no ip address
 shutdown
 duplex half
!
interface Ethernet1/3
 ip address 10.4.4.5 255.255.255.0
 duplex half
 mpls label protocol tdp
 tag-switching ip
!
router ospf 5
 log-adjacency-changes
 network 5.5.5.5 0.0.0.0 area 0
 network 10.4.4.0 0.0.0.255 area 0
 network 192.168.0.0 0.0.255.255 area 0
!
router bgp 100
 no synchronization
 bgp router-id 50.50.50.50
 bgp log-neighbor-changes
 neighbor 2.2.2.2 remote-as 100
 neighbor 2.2.2.2 update-source Loopback0
 neighbor 192.168.6.6 remote-as 300
 no auto-summary
!
ip classless
no ip http server
no ip http secure-server
!
!
gatekeeper
 shutdown
!
!
line con 0
 exec-timeout 0 0
 logging synchronous
 stopbits 1
line aux 0
 stopbits 1
line vty 0 4
 login
!
!
end
r5#
 
r6#sh run
Building configuration...
Current configuration : 1055 bytes
!
version 12.3
service timestamps debug datetime msec
service timestamps log datetime msec
no service password-encryption
!
hostname r6
!
boot-start-marker
boot-end-marker
!
!
no aaa new-model
ip subnet-zero
!
!
no ip domain lookup
!
ip cef
!
!
interface Loopback0
 ip address 6.6.6.6 255.255.255.255
!
interface FastEthernet0/0
 no ip address
 shutdown
 duplex half
!        
interface Ethernet1/0
 ip address 192.168.6.6 255.255.0.0
 duplex half
!
interface Ethernet1/1
 no ip address
 shutdown
 duplex half
!
interface Ethernet1/2
 no ip address
 shutdown
 duplex half
!
interface Ethernet1/3
 no ip address
 shutdown
 duplex half
!
router bgp 300
 no synchronization
 bgp router-id 60.60.60.60
 bgp log-neighbor-changes
 network 6.6.6.6 mask 255.255.255.255
 neighbor 192.168.5.5 remote-as 100
 no auto-summary
!
ip classless
no ip http server
no ip http secure-server
!
!
gatekeeper
 shutdown
!
!
line con 0
 exec-timeout 0 0
 logging synchronous
 stopbits 1
line aux 0
 stopbits 1
line vty 0 4
 login
!
!
end
r6#
 
MPLS技术领域中需要熟记于心的控制层面,转发层面结构图如下所示.
 
上图描述的还不是很全面. 控制层面中的routing protocol指的是IP层面的路由控制交互协议,比如RIP,OSPF等等. LDP指的是标签分发协议LDP或者TDP. IP Forwarding Table及是Cisco IOS中的cef table(FIB) 有ip routing table 和LIB 共同作用生成. 所有存在Label Forwarding Table(LFIB) 向IP Forwarding Table做的一个标签回灌动作. 同时, LFIB需要LIB和FIB信息产生, 所以通常可以认为LFIB = LIB + FIB
进入MPLS路由器的报文是查询FIB表还是LFIB表是根据报文是否带有标签决定的. 带有标签的报文,以太帧头中的类型字段是不一样的.
同理, 出MPLS路由器的报文既有可能是IP包(pop或者untag动作后)也有可能是标签包.
[常用命令]
(1) show mpls ldp bindings (查看LIB标签表)
(2) show mpls forwarding-table detail (查看LFIB的标签转发表)
(3) show  ip cef detail (查看FIB IP转发表)
[注!!!]
(1) CISCO的MPLS实现, ip cef 必须启用
(2) mpls ldp route-id 必须ip层面可达
(3) 次末跳弹出原则, ldp协议给本路由器直连网段分配空标签, 下一跳为空标签的情况下, 执行标签pop动作
(4) ldp协议不为BGP路由分配标签(在MPLS ×××的应用中为顶层标签), 该标签取自BGP路由下一跳地址所分配的标签
(5) 以上的实现中, R2与IPS1的网段和R5与IPS2的网段宣告进了OSPF. 如果这两个网段不宣告进OSPF, R2和R5互为IBGP邻居的配置中, 配置neighbour xx.xx.xx.xx next-hop-self, 从R1以lo0为源访问R6的lo0仍然是通的. !!!但是如果R2和R5的BGP router id 不取环回口lo0, 而取与Core1和Core2的直连接口做router id, 这样会有路由黑洞产生. (因为次末跳弹出原则, 从R1到R6的包, 在Core2处就会pop标签而查询FIB表,这个时候Core2并没有BGP路由,导致黑洞.) 所以在设计好的MPLS网络时候, 要仔细考虑协议.