One of the advantages of ASM in that scenario is that ASM mirroring supports "preferred mirror reads". ASM knows about which part of a mirror pair is at the local and which one is at the remote site and will read from the local LUN instead from the remote one to avoid the latency hit that would come with querying the other lun.
It solves or improves one downside of an extended distance rac cluster but there are still other things to consider, like network latency, disk writes and more
a.) Make sure your latency between the nodes is low. Otherwise the performance might suffer. It is also for that reason that normally you are restricted about the distance to around 5km for Ethernet and 50km for DWDM.
b.) Extended Cluster is NOT DR: E.g. a corrupt block on one side will likely be mirrored to the other side as well. If you are looking for a disaster recovery solution, take Data Guard.
Also see this whitepaper: