Add support of remote* table

Why it matters

This feature enables users to leverage parallel replica queries in remote* table functions when connecting to a cluster, improving query performance and fault tolerance by distributing queries across multiple replicas. It also addresses issue #73295 to enhance the reliability and scalability of remote table functions.

How to use it

Specify a cluster name in the address_expression argument of any remote* table function. ClickHouse will automatically connect to parallel replicas within the cluster, allowing query execution to be distributed and parallelized across those replicas.