Separate companies - redundant topology

This one is real life scenario where two companies, which owns two Ethernet lines (50Mb each), want to get separate but keep load-balance, on regular basis, and redundant path in case of failure, for each one of them.

I will use the following topology:

Network setup:

      -          Each Ethernet line is provided with two separate VLAN’s from the service provider

      -          CPE1 represent the companies physical location, two LAN’s, one for each company 

             (Company 1-192.168.11.0/24 and Company 2-192.168.12.0/24)

      -          CPE2 represent some sort of data center where servers, firewalls and internet exit 

             located (Company 1-192.168.21.0/24 and Company 2-192.168.22.0/24)

      -          BGP is used as routing protocol (real customer demand) while I would prefer OSPF or  

            EIGRP which were much suitable for this topology.

      -     HSRP is used between CPE and routers for keeping the default route

I had to use  VRF lite in order to separate the routing tables for each company

After I finished configuring the interfaces, in their correct VRF, I started to configure the BGP and noticed that no neighbor had come up and I started to get the following messages:

*Mar  1 18:14:48.719: %BGP-4-NORTRID: BGP could not pick a router-id. Please configure manually.

And:

R4#sh ip bgp
% BGP cannot run because the router-id is not configured

After a little investigation I saw that the routers don’t have router-ID:

R4#sh ip bgp vpnv4 vrf CO1
% BGP cannot run because the router-id is not configured

BGP table version is 1, local router ID is 0.0.0.0
Status codes: s suppressed, d damped, h history, * valid, > best, i - internal,
              r RIB-failure, S Stale
Origin codes: i - IGP, e - EGP, ? - incomplete

   Network          Next Hop            Metric LocPrf Weight Path
Route Distinguisher: 1:100 (default for vrf CO1)
*  10.1.34.0/24     0.0.0.0                  0         32768 i
*  192.168.21.0     10.1.34.254              0         32768 i

And this is my conclusion although I didn’t check it for sure so I may be wrong:

The routing process should take the highest IP (first loopbacks the interfaces) as the router-ID, in this situation he couldn’t use any IP because there were in VRF and not in the global routing table.

After configuring  a loopback interface is solved this problem.

the rest of this lab, along with configs and NET map, can be found in the following link:

http://www.4shared.com/rar/2NVMjzxb/CONFIG.html