Now the fun really begins…

By Jay Kidd, Chief Technology Officer, NetApp


Yesterday we announced clustered Data ONTAP 8.2. This release brings a number of new capabilities to ONTAP, such as storage Quality-of-Service and support for Microsoft’s SMB 3.0 protocol. There is a lot more detail posted elsewhere on all the great stuff that is now available.


I am particularly excited about this release because I expect it to be the trigger for a large number of our existing customers to make the leap to clustered ONTAP.   We have had great adoption and uptake in the 11 months since ONTAP 8.1 became generally available. Sales have doubled each quarter for the past year,  and the responses from the roughly 1,000 customers using the clustering capabilities has been phenomenal.


The core value of clustered ONTAP is the ability to eliminate planned downtime caused by storage tasks. Most vendors offer hardware resiliency to avoid unplanned outages, but the ability to do expansion, migration, and full tech refresh without the application owners having to know is unique to NetApp.  Data migrations are a common practice and in some cases an organization can perform hundreds a year, many of which require the application to be offline. Each application outage may only take an hour or so, but the planning for the event and the weekend work, can easily roll up to 40 man hours of time. All of that work goes away with the non-disruptive operations capability of clustered ONTAP.


I hear excitement about this from customers all the time.  A big semiconductor shop in Europe who moved from ONTAP 7-mode to clustered ONTAP talked about ‘eliminating the pockets of stranded storage capacity’ on their systems since any capacity can be accessed from any node of the cluster, and volumes can be moved to fill in the pockets as needed. Peak Colo, an innovative service provider in the US, is running upwards of 10,000 VMs on their 6 node cluster and non-disruptively moving customer workloads between nodes frequently to balance performance and capacity as new tenants are provisioned.  Luke Norris, CEO at Peak even told a story of a customer who thought storage was the root cause of a performance problem, which led Peak to move their data volumes from SAS disk to SSDs while the application was running. Turns out, storage was definitely not the issue, so they moved it back.  All without an outage to the application.


Several customers have talked about the leasing flexibility that clustered ONTAP provides.   With discreet storage systems, they had to interrupt normal work to migrate the data to the new equipment when the lease expired.  This often took a few months of work at an inconvenient time, and always required some outage to the application.   With clustered ONTAP, nodes of the cluster can have lease expiration dates at different time and can be rolled out and rolled in without disrupting the applications. Immortal storage is an ambitious term, but clustered ONTAP enables that to become true.


Photobucket is a long-time NetApp customer who has recently added clustered ONTAP to their environment. In addition to the non-disruptive load balancing capability, the global cluster namespace simplified the operations team life by reducing the number of discreet storage mount points from over 50 to 1.


Add this to the great storage efficiency that ONTAP is known for, plus support for NAS and SAN, primary and secondary, and a rich set of application support, and it is a big deal for the industry. 


There is also a lot of discussion lately about software defined storage, which promises to allow applications to provision and control virtual storage containers on a range of hardware. This is not a new idea, just a new term.  ONTAP delivers highly efficient block and file storage from virtual storage machines that can be placed on NetApp disk, on disk arrays from other vendors via NetApp V-series, or on commodity servers via ONTAP-Edge.   Applications can provision and manage this storage according to policies in an automated way. This is an interesting component of a software-defined datacenter, and we look forward to extending the capabilities and agility that clustered ONTAP delivers today.


I haven’t had this much fun in years. 


Now lets wait for the SnapManagers

Projected for August according to my SE