What underlying mechanisms govern an AWS EBS Snapshot when creating point-in-time, incremental block-level backups across scalable cloud infrastructures? Furthermore, why is offloading volume state directly to Amazon S3 the definitive engineering approach to ensure rapid disaster recovery, data integrity, and cost-effective storage durability?