This conference talk from FAST '25 explores scalable integrity checking mechanisms for secure cloud storage. Learn how researchers from the University of Wisconsin–Madison analyze the performance costs of Merkle hash trees—the standard method for protecting data integrity and freshness in storage systems. Discover their innovative Dynamic Merkle Trees (DMTs) solution, which intelligently exploits workload patterns to deliver up to 2.2X improvements in both throughput and latency compared to existing approaches. The presentation details how the team quantified storage-level hash tree overheads in realistic settings and developed optimizations based on identified performance bottlenecks. Gain insights into this promising new direction for achieving efficient, scalable integrity guarantees in cloud storage systems.