
Maximize & Sustain Performance
Eliminate manual performance tuning. AOS intelligently tiers and places data close to workloads that need it most.

Keep Data Safe & Resilient
Built-in encryption, replication, and self-healing ensures data is always safe and available. Never lose data.

Optimize Resource Utilization
Continuous deduplication and erasure coding squeeze maximum efficiency without compromising performance.
Co-locates applications & data to optimize performance
Storing data locally means the majority of reads are handled by local hardware, improving latency and resource utilization. The replica location of a write is dynamically chosen based on current utilization metrics, preventing replica traffic from burdening other nodes. This enables consistent performance by preventing network congestion and by keeping resources balanced. When a workload migrates, data is pulled local as it is accessed to avoid reading remotely.

Accelerate applications with local NVMe
AOS architecture is optimal for leveraging the performance benefits of NVMe drives without the need to redesign applications. Data locality enables applications to benefit from the low latency and high throughput possible with NVMe devices, and optimizations such as RDMA based replication accelerate application writes while ensuring data availability.
Ensure Quality of Service
Give application owners peace of mind with fine-grained Quality of Service (QoS) control for their workloads, without the need to create silos for different classes of applications. Configure limits on throughput or IOPS for misbehaving workloads, and pin critical applications to flash in hybrid clusters.

Self-heal quickly without impacting applications
Distributed self-healing is made possible due to the way AOS protects data: writes are replicated across at least two nodes in a cluster, with one local copy for quick access and the other copy sent to another node. Replicas are distributed evenly across the cluster, which enables fast rebuilds and enables consistent performance at scale, even when a rebuild is taking place.
Maximize storage utilization
Get the most out of your storage investment with advanced data reduction capabilities that deliver maximum savings without needing to compromise.
AOS optimizes capacity utilization without hurting performance or resiliency by dynamically compressing, deduplicating, and erasure coding blocks of data based on potential capacity savings as well as application access patterns.
Optimize price versus performance
AOS combines all persistent storage devices across the cluster into a single pool, instead of wasting capacity by reserving fast drives for caching. Data is dynamically moved between tiers in granular segments to keep the most active data in the fastest drives, thereby optimizing for performance while reducing cost.
Respond quickly with efficient clones and snapshots
Snapshots enable quick local restores for backup and disaster recovery, and clones make it easy to quickly and reliably spin up workloads.
AOS tracks snapshots with a fine-grained redirect-on-write algorithm that allows fine-grained control and maximizes both efficiency and performance. Clone hundreds of new workloads in seconds, and create deep snapshot chains without impacting applications.
Rely on a highly scalable & self-managed data service
Unlike other HCI solutions in the market, AOS tracks data in small segments enabling it to manage data much more efficiently. Its distributed metadata service uses algorithms like paxos and map/reduce to ensure consistency and to enable distributed data processing.
This architecture enables AOS to continuously optimize data management and sustain performance while the cluster grows and as more workloads are added, long after day 1.

Become an Expert at Nutanix University
Learn more about the Nutanix platform architecture and design considerations

The Nutanix Bible
The Nutanix Bible breaks down the inner-workings of AOS and explains how it all comes together.

Five Secrets About NVMe
NVMe is the next generation of performance media, but there are some secrets the traditional storage vendors don’t want you to know.

Elasticsearch on AOS
Elasticsearch is a scale-out workload that benefits from distributed storage. Learn about how you can improve resiliency of your Elasticsearch deployment with AOS.