Data model design

I have a requirement that force me to split out the data across buckets and split out buckets across separate clusters:

Cluster A
bucket A
Cluster B
bucket B
Cluster N
bucket N

All of the data in the buckets A, B, C comprises a single record, but for the purpose of security it is split out.

Are there any potential issues i need to be aware of from the standpoint of query/view design, cluster design, deployment, connecting and compositing such data?

Thanks !

For one, a single N1QL query or index cannot span multiple clusters.