Bandwidth scaling and High Availability are built into the Transit Gateway inherently as well via the multiple Availability ZOne (AZ) attachments to our VPCs. However; for connectivity back to our Datacenter, we have some things to consider. In the real world, we would create another customer gateway on a totally separate physical device. Ideally this is physically as fault-isolated from the first Customer Gateway as we can make it (think, in across the room or even in another communications room. With separate connectivity and power if we have it). But in order to distribute load across both Customer gateways from the datacenter, you typically would use another tier of routers (shown below via a core router) to balance the traffic. For our demo purposes, lets build it on the same CSR we have in our simulated Datacenter so we can see ECMP in action.
In the AWS Management Console change to the region you are working in. This is in the upper right hand drop down menu.
In the AWS Management Console choose Services then select VPC.
From the menu on the left, Scroll down and select Transit Gateway Attachments.
You will see the VPC Attachments listed, but we want to add one to connect our Datacenter. Click the Create Transit Gateway Attachment button above the list.
Fill out the Create Transit Gateway Attachment form.
For Inside IP CIDR for Tunnel 1 use 169.254.12.0/30 for CIDR. Note: we are different addresses from the previous VPN
For Pre-Shared Key for Tunnel 1 use awsamazon
For Inside IP CIDR for Tunnel 2 use 169.254.13.0/30 for CIDR. Note: we are different addresses from the previous VPN
For Pre-Shared Key for Tunnel 2 use awsamazon
Once the page is filled out, click Create attachment at the bottom right.
While we are on the Transit Gateway Attachments page, lets go back to the top and give the VPN connection a name. Scan down the Resource type column for the VPN Attachment. *note: you may have to hit the refresh icon in the upper right above the table to get the new VPN to show. If you click the pencil that appears when you mouse over the Name column, you can enter a name that’s different than the first VPN. Be sure to click the check mark to save the name.
From the Menu on the Left Select Site-to-Site VPN Connections. From the main panel, you likely will see the new VPN is in State pending. That fine. Lets take a look toward the bottom, and click the Tunnel Details tab. Record the two Outside IP Addresses. We want to record them in the order of the one pairing up with the Inside IP CIDR range 169.254.12.0/30 first. note: You can use cloud9 as a sratch pad, by clicking the + in the main panel and selecting New file. be sure to paste them in the right order!
From the Menu on the Left Select Transit Gateway Route Tables. From the table in the main panel select Green Route Table. Lets take a look toward the bottom, and click the Associations tab. Associations mean that traffic coming from the outside toward the Transit gateway will use this route table to know where the packet will go after routing through the TGW. note: An attachment can only be Associated with one route table. But a route table can have multiple associations. Here in the Green Route Table, We already have two associations, The Datacenter Services VPC and the 1st Site-to-Site Tunnel. Click Create associations in the Associations tab. From the drop-down list, select the new vpn. note:it should be the only one in the list without a Association route table . Click Create association.
While at the Transit Gateway Route Tables, take a look at the Propagations tab. These are the Resources that Dynamically inform the route table. An attachment can propagate to multiple route tables. For the Datacenter, we want to propagate to all of the route tables so the VPC associated with each route table can route back to the datacenter. Lets start with the Green Route Table. We can see all of the VPCs are propagating their CIDR to the route table. Since the Datacenter Services VPC is also associated with this route table, we need to propagate this new second set of VPN routes to the Green Route Table.
Click in Create Propagationon the field “chose attachment to propagate”, select the attachment of the VPN (previously named by you) and click in Create propagation.
Repeat the above step on the propagations tab for the Red Route Table and the Blue Route Table so VPCs associated with these route tables also get four paths to the datacenter.
Take a look at each of the route tables and notice the tab Routes. You can see the routes that are propagated, as well as a static route table that was created for you by the CloudFormation template. That’s the default route (0.0.0.0/0) that will direct traffic destined for the internet to the Datacenter Services VPC and ultimately through the NAT Gateway in that VPC. note: there is also a route table with no name. This is the default route table. In this lab we do not intend to use the default route table.
Back on the Cloud9 browser tab, using the two VPN tunnel endpoint address generated from the step above, cd to tgwwalk on the Cloud9 bash console and run the bash script, ./create2ndcsr.sh. note: THIS is a different Script than above! Also, be sure to put the address that lines up with Inside IP CIDR address 169.254.12.0/30 for ip1. Example from Site-to-Site VPN
cd tgwwalk ##./create2ndcsr.sh ip1 ip2 outputfile ./create2ndcsr.sh 220.127.116.11 18.104.22.168 my2ndcsrconfig.txt
note: AWS generates starter templates to assist with the configuration for the on-prem router. For your real world deployments, you can get a starter template from the console for various devices (Cisco, Juniper, Palo Alto, F5, Checkpoint, etc). Word of Caution is to look closely at the routing policy in the BGP section. you may not want to send a default route out. You likely also want to consider using a route filter to prevent certain routes from being propagated to you.
On the left hand panel, the output file should be listed. You may have to open the tgwwalk folder to see the txt file. Select all text (ctrl-a on pc/command-a on mac). Then copy the text to buffer (Select all text (ctrl-c on pc/command-c on mac))
enter configuration mode, which will take you to a config prompt
ip-10-4-0-17#conf t Enter configuration commands, one per line. End with CNTL/Z. ip-10-4-0-17(config)#
Once in Configuration mode note: you should see (config)# prompt, paste Select all text (ctrl-v on pc/command-v on mac) in the text from the outputfile created in step 4. This will slowly paste in the configuration file.
if you are still at the (config)# or (config-router) prompt, type end and press enter.
Now lets look at the new interfaces: sh ip int br. You should see new interfaces: Tunnel3 and Tunnel4 and they both should show up. *note: if they do not change from down to up after a 2 minutes, likely cause is the ip addresses were flipped in the createcsr script.
Lets make sure we are seeing the routes on the Cisco CSR. first we can look at what BGP is seeing: show ip bgp summary. The most important thing to see is the State/PfxRcd (Prefixes received). If this is in Active or Idle (likely if neighbor statement is wrong: IP address, AS number) there is a configuration issue. What we want to see is a number. In fact if everything is setup correctly we should see 4 for each neighbor.
ip-10-4-2-30#sh ip bgp summ BGP router identifier 169.254.10.2, local AS number 65001 BGP table version is 68, main routing table version 68 5 network entries using 1240 bytes of memory 17 path entries using 2312 bytes of memory 4 multipath network entries and 16 multipath paths 2/2 BGP path/bestpath attribute entries using 560 bytes of memory 1 BGP AS-PATH entries using 24 bytes of memory 0 BGP route-map cache entries using 0 bytes of memory 0 BGP filter-list cache entries using 0 bytes of memory BGP using 4136 total bytes of memory BGP activity 14/9 prefixes, 40/23 paths, scan interval 60 secs Neighbor V AS MsgRcvd MsgSent TblVer InQ OutQ Up/Down State/PfxRcd 169.254.10.1 4 65000 368 391 68 0 0 01:00:41 4 169.254.11.1 4 65000 367 387 68 0 0 01:00:40 4 169.254.12.1 4 65000 374 394 68 0 0 01:01:44 4 169.254.13.1 4 65000 373 392 68 0 0 01:01:39 4
Lets verify Equal Cost Multipathing (ECMP). Making sure we setup ECMP, back in config mode we will add maximum-paths to 8:
ip-10-4-0-17# config t router bgp 65001 address-family ipv4 maximum-paths 8 end
Now, run sh ip ro command . See, both the tunnels are showing up!
...<output omitted> S* 0.0.0.0/0 [1/0] via 10.4.0.1, GigabitEthernet1 10.0.0.0/8 is variably subnetted, 7 subnets, 3 masks B 10.0.0.0/16 [20/100] via 169.254.13.1, 01:01:05 [20/100] via 169.254.12.1, 01:01:05 [20/100] via 169.254.11.1, 01:01:05 [20/100] via 169.254.10.1, 01:01:05 S 10.4.0.0/16 is directly connected, GigabitEthernet1 C 10.4.0.0/22 is directly connected, GigabitEthernet1 L 10.4.2.30/32 is directly connected, GigabitEthernet1 B 10.8.0.0/16 [20/100] via 169.254.13.1, 01:01:05 [20/100] via 169.254.12.1, 01:01:05 [20/100] via 169.254.11.1, 01:01:05 [20/100] via 169.254.10.1, 01:01:05 B 10.16.0.0/16 [20/100] via 169.254.13.1, 01:01:05 [20/100] via 169.254.12.1, 01:01:05 [20/100] via 169.254.11.1, 01:01:05 [20/100] via 169.254.10.1, 01:01:05 B 10.17.0.0/16 [20/100] via 169.254.13.1, 01:01:05 [20/100] via 169.254.12.1, 01:01:05 [20/100] via 169.254.11.1, 01:01:05 [20/100] via 169.254.10.1, 01:01:05 169.254.0.0/16 is variably subnetted, 8 subnets, 2 masks C 169.254.10.0/30 is directly connected, Tunnel1 L 169.254.10.2/32 is directly connected, Tunnel1 C 169.254.11.0/30 is directly connected, Tunnel2 L 169.254.11.2/32 is directly connected, Tunnel2 C 169.254.12.0/30 is directly connected, Tunnel3 L 169.254.12.2/32 is directly connected, Tunnel3 C 169.254.13.0/30 is directly connected, Tunnel4 L 169.254.13.2/32 is directly connected, Tunnel4
Just to verify where those routes are coming from, we can take a look at the Green Route Table. note: remember, it’s under the VPC service and Transit Gateway Route Tables at the bottom of the left menu.