It might not be a good thing to do that for the following.
1. the 100 ms latency might become a problem for some of the messaging system and requests that go to the ESXi hosts, or between Cell/vCenter. I think it needs to be 10 ms or under ... something really narrow.
2. To have 2 vCD Cells clustered, you need to have an NFS share mounted to both of them. This might be an issue if you have to do an NFS mount across geographies.
there are other condiserations too ... might want to read the following : http://www.yellow-bricks.com/2013/01/15/stretched-vcloud-director-infrastructure/