In order to maintain high availability for services running on Azure virtual machines in the event of a single data center failure, what are two effective strategies for deploying these virtual machines?