Beruflich Dokumente
Kultur Dokumente
(VSNL)
Shailesh Gupta, Vice President Raju Raghavan, Manager
Contents
Tata Communications ISP Network Journey since Aug15th, 1995 Advantages of IPV6 over IPV4 IPV6 Drivers & design in TCL Network IPV6 Customers of TCL, India Expectations from GOI/DOT/TEC for promoting IPV6 in India Thanks
KDDI-2XDS3
China
Hong Kong
Santa Clara
Chicago
TCL -7xGig
5
CAT Thailand DS3
TCL -2xSTM-4
Sydney
MSN-Gig
* Dual stack edge and core . ** Independent core topologies for IPv4 and IPv6
International Connectivity
FLAG SEA-ME-WE3
Kolkat a
MUMBAI
TATA Indicom Chennai, India India Singapore Cable 5.12 Tbps System
Atlantic Route
Submarine Cable & Carrier Level Redundancy TATA Indicom is the Only Carrier having capacity on 5 out of 5 cables into India
SINGAPORE
SAFE
117 location across the the India. 3-tier Hierarchical topology for better management. 9 Big Tier 1 cities including 4 metros 7 Major Tier 2 cities. 101 Tier 3 cities
* Serves the length and breadth of India. ** IPv6 Dual stack edge with MPLS tunneled core.
0.064 10
n'9 6
24
n'9 7
g'9 5
n'0 6 Ja
Au
De
Ja
Ja
Ja
Ja
Ja
Ja
Ja
Ja
Ja
c'0
The IPv6 routes in the global routing table has grown from 500 prefixes in 2004 to around 1200 prefixes in 2008. The IPv6 ASes in the global routing table has gone from 400 ASes in 2004 to around 900 ASes in 2008. A two fold increase. (IPv4 ASes - 28000 )
One of the prime drivers for IPv6 adoption globally has been Service Differentiation and the pursuit of competitive advantage (First Mover).
13 Fields
Remove unwanted header fields Simple header.
8 Fields
Less fields means lesser processing good for devices which have lesser processing power?
Multi-Topology ISIS protocol builds separate topologies for IPv4 and IPv6 A B
C E D F
o TCL promoted the experimentation of IPv6 and the 6bone/6TAP initiative as a member of the Canarie Policy Board. o TCL is a founding member of the IPv6 Forum . o TCL provided IPv6 connection to the last SANOG meeting.
Tata - NIXI Mumbai connectivity Getting few prefixes from RIC & Netmagic 3 prefixes ,one /48 from NIXI (2001:DE8:1::/48), one each from Netmagic & RIC Honesty (ISP) at Mumbai IIT Kanpur 1 Gb 40+ international IPV6 customers
IPV4 - 3.4 L+ routes in global routing table IPV6 15000 routes in global routing table
The more seemingly impossible the goal, the more glorious is the triumph of achievement!
Backup Slides
Customer Expectations: Reachability of customer IPv4 and IPv6 prefixes in Internet Domain. Customer required Full IPv6 route from Tata Communications (15 K) Both IPv4 and IPv6 sessions on same link.
Scenario Customer is having IPv6 prefixes with his own AS number. wants to run IPv6 BGP session with Tata communications on his existing IPv4 link. Customer is already running IPv4 BGP session on same link for his IPv6 Prefixes.
Challenges: Propagating Customer IPv6 prefixes in Internet Domain. ensuring propagation within Tata communication domain ensuring propagation within Tele globe domain. ensuring proper Hand-off between these Two domains. Establishment of IPv6 BGP session and IPv4 BGP session on same link. Configuring proper route-map and IPv6 Prefix lists. Proper network and mask configuration w.r.t. /128 mask to avoid route leakage and improper announcements to Customer as well as to upstream provider. Testing
PE Configs :
interface Vlan2455 description 022100167783/Honesty net solutions (India) Private Limited/Bhandup (West) Mumbai/STDILL/20MB/VBL ip address 115.113.128.85 255.255.255.252 ipv6 address 2403:0:100::1/126 ipv6 enable address-family ipv6 no synchronization network 2403:0:100::/126 neighbor 2403:0:100::2 activate neighbor 2403:0:100::2 prefix-list HONESTY_IPV6 in neighbor 2403:0:100::2 maximum-prefix 1000 100 restart ipv6 prefix-list HONESTY_IPV6: 1 entries seq 10 permit 2401:4800::/32 le 64
Testing:
PE#sh bgp ipv6 sum BGP router identifier 192.168.196.5, local AS number 4755 Neighbor V AS MsgRcvd MsgSent TblVer InQ OutQ Up/Down State/PfxRcd 0 0 3w4d 1
Testing:
Trace from Looking Glass
Router: gin-mtt-mcore3 Site: CA, Montreal - MTT, TATA COMM. INT. CENTER Command: traceroute ipv6 2401:4800:1::2 Tracing the route to 2401:4800:1::2 1 if-13-0.mcore4.nqt-newyork.ipv6.as6453.net (2001:5A0:300:100::2) 8 msec 8 msec 8 msec 2 2001:5A0:400:100::16 204 msec 200 msec 200 msec 3 2001:5A0:2300:100::6 204 msec 204 msec 200 msec 4 2401:4800:1::2 208 msec 204 msec 208 msec 5 2401:4800:1::2 204 msec 208 msec 204 msec
2. IPV6 Deployment NIXI, Mumbai mu-mhp-vas-mi04-rt01 192.168.132.232 7606 Gi2/13 Tata-Mumbai-Vashi Fibre-to-Elect Convertor(gig) NIXI router 7606
Peering IP:2001:DE8:1::1 ASN : 24029 Note 1:Advertising specific 4755 IPv6 network, 2403::/32 on this session while accepting specific NIXI customer (Reliance & Netmagic) ipv6 n/w with no-export Note2:The ipv6 networks from 6453 (around 1600 prefixes) and 4755 ipv6 n/w available in Mumbai ICG were passed to Vashi Micropop during this implementation to cater NIXI customers as well as our customers
mu-mhp-vas-mi04-rt01#sh run | inc 2001:DE8:1::1 neighbor 2001:DE8:1::1 remote-as 24029 neighbor 2001:DE8:1::1 description ***eBGP IPV6 with NIXI *** no neighbor 2001:DE8:1::1 activate neighbor 2001:DE8:1::1 activate neighbor 2001:DE8:1::1 route-map NIXIV6-IN in neighbor 2001:DE8:1::1 route-map NIXIV6-OUT out neighbor 2001:DE8:1::1 maximum-prefix 1000 100 restart 30 mu-mhp-vas-mi04-rt01#sh route-map NIXIV6-IN route-map NIXIV6-IN, permit, sequence 10 Match clauses: as-path (as-path filter): 100 Set clauses: community 4755:98 4755:2255 no-export Policy routing matches: 0 packets, 0 bytes mu-mhp-vas-mi04-rt01#sh ip as-path-access-list 100 AS path access list 100 permit ^24029$ permit 18101 permit 17439 mu-mhp-vas-mi04-rt01#sh route-map NIXIV6-OUT route-map NIXIV6-OUT, permit, sequence 10 Match clauses: ipv6 address prefix-list TATAV6 Set clauses: Policy routing matches: 0 packets, 0 bytes mu-mhp-vas-mi04-rt01#sh ipv6 prefix-list TATAV6 ipv6 prefix-list TATAV6: 1 entries seq 5 permit 2403::/32
mu-mhp-vas-mi04-rt01#sh run int gi2/13 Building configuration... Current configuration : 381 bytes ! interface GigabitEthernet2/13 description NIXI DEMO LINK FOR ipv6 TEST | Direct Physical connection to NIXI witch through Tata owned Opt-to-Elect Convertor | 16Dec08 no ip address ipv6 address 2001:DE8:1::3/64 ipv6 enable wrr-queue cos-map 3 1 7 priority-queue cos-map 1 5 6 rcv-queue cos-map 1 4 1 3 rcv-queue cos-map 1 6 0 2 4 7 rcv-queue cos-map 1 8 5 6 end mu-mhp-vas-mi04-rt01#ping ipv6 2001:DE8:1::1 Type escape sequence to abort. Sending 5, 100-byte ICMP Echos to 2001:DE8:1::1, timeout is 2 seconds: !!!!! Success rate is 100 percent (5/5), round-trip min/avg/max = 0/0/4 ms
Currently, only a /36 (2403::/36) from the Global space has been used. 8 numbers of /40 used for infrastructure and other 7 cities namely Mumbai, Chennai, Ernakulam, Delhi, Bangalore, Hyderabad, Kolkata (excel sheets attached). As per IPv6 LLD we have calculated /40 subnets into smaller subnets to arrive different WAN/loopback and Customer IP E.g. Mumbai 2403:0:100:0:0:0:0:0 /40 subnetted to /48s First 2403:0:100:0:0:0:0:0/48 subnetted to /64s. For customers we will allocate /64 LAN IP. First 2403:0:100:0:0:0:0:0/64 is further subnetted till we get 2403:0:100:0:0:0:0:0 /112 and WAN IP given to first customer. Please go through the IPv6 LLD.
Diffserv / IP Precedence
Diffserv / IP Precedence
autoconfig flag.
Src= Link Local Address (FE80::/10) Dst = All nodes multicast address (FF02::1)
IPv6 has link local addresses, so on a local lan even without any protocol exchanges the IPv6 can start communicating.
BGP Update source IPv4 address is converted to IPv6 mapped address ::FFFF:192.168.1.1
The destination PE installs the route in the routing table with ::FFFF.192.168.1. as next hop.
IPv6 Packet
The IGP label corresponding to ::FFFF.192.168.1.1 (192.168.1.1) is configured as the outer label .
IPv6 Packet