Yesterday, NetApp made the 8.3 release of Data ONTAP Generally Available. This release marks, in many ways, the completion of our platform transition from 7-mode to clustered Data ONTAP. Each of these individual sub-bullets will eventually make up their own deep-divey post, but there are a couple features I wanted to call out that will truly change the way you manage your storage infrastructure.
- Integrated System Manager – with 8.3 System Manager moves from a client installed application to a simple web app that runs natively on the ONTAP cluster. No more Java incompatibilities/dependencies!
- Automated Non-Disruptive Upgrades – clustered Data ONTAP has always had the ability to non-disruptively perform ONTAP upgrades. However now the process is fully automated across the entire cluster, and is triggered using a single command.
- Version Independent SnapMirror – 8.3 introduces a new type of SnapMirror relationship that is no longer tied to ONTAP version. This means that going forward organizations with complex replication topologies can perform NDU of each cluster as the need arises, without having to do the upgrades concurrently across source and destination, and without having to resync the relationship.
- IP Spaces – logical routing for each SVM. This provides the ability to have overlapping CIDR domains within a cluster, a capability critical to any serious multi-tenant architecture.
- Broadcast Domains – a logical entity that enables the administrator to define layer 2 networks at a cluster level. This also removes the need to create customized failover groups since the interfaces are automatically assigned based on the subnet used.
- Subnets – Are you tired of managing IP address assignment for your LIFs? Using clustered Data ONTAP subnets the cluster manages IP assignment for you, automatically assigning and reclaiming addresses as LIFs are created and destroyed.
- Inline Zero Elimination – Starting with 8.3, ONTAP will perform inline detection and elimination of zeroes. This results in both increased efficiency, as well as dramatic performance improvements in certain workloads.
- Advanced Disk Partitioning – Also starting with 8.3 any AFF or FAS25xx will leverage ADP for the root and data aggregates. This intelligently partitions the disks to serve both types of aggregate using the same physical asset, dramatically improving efficiency by 50+% for many platforms.
- Network Compression for SnapMirror – Further reduce replication intervals and WAN traffic by compressing all SnapMirror communications.
- MetroCluster – Build active-active, synchronously mirrored, storage systems that can be physically stretched for up to 200km. The most available, efficient, and manageable storage platform possible today.
- LUN Move – Starting with 8.3 a LUN can be dynamically relocated to another Qtree or volume within a SVM non-disruptively. This completes the work to fully realize Data Motion within ONTAP, as every element of the system can now be moved without impact operations.
- AFF Personality – Starting now, any FAS80xx can be ordered from the factory to be fully optimized for an All-Flash world. Such systems will not support HHD as they have been tuned even further for the advantages of an All Flash deployment.
- Storage Pools – Building on the Advanced Disk Partitioning technology, SSDs can be shared between the nodes of an HA pair to construct a Storage Pool. This pool can then be provisioned to create a hybrid aggregate utilizing a portion of the available capacity across each of the SSDs. This technology effectively lowers the economics of Flash to the point where any system can now benefit.
- Massive Increases in Cache Maximums – Across the board the amount of FlashCache and FlashPool capacity that can be leveraged went up dramatically. Enabling Virtual Storage Tiering to service even larger datasets without the need to perform a hardware upgrade.
Also check out fellow DatacenterDude.com contributor Justin Parisi’s post on Data ONTAP 8.3!