When you think of the heavyweights of the storage industry, names like EMC and NetApp are two of the first to come to mind. But what has allowed them to grow so large and establish such a solid footing in the highly competitive storage industry? We all understand now that the same factories in Asia make hard drives for every application, and the world learned how fragile the supply chain really is when flooding localized around Bangkok reduced the entire industry’s output by 25%; causing an increase in disk prices across the board that we are still feeling the effects of today. The reason these two firms have such a dominant position in the marketplace is not because they build ‘magic’ hard drives that are somehow faster and more efficient than their less expensive counterparts, the real ‘magic’ is in the software that controls those drives. It is assumed the readers of this article know what storage arrays are and how they can help businesses manage and protect information assets, our goal is to shine a light on some of the capabilities that differentiate NetApp’s Clustered Data ONTAP Operating System from the rest of the storage industry.
Unified Architecture – One of the most underappreciated features of NetApp storage is that every NetApp system runs the same operating system called Data ONTAP. Why does this matter? Consider the fact that Southwest Airlines standardizes on a single aircraft platform, the Boeing 737. Southwest saw the benefits of having a unified platform for pilots to fly and mechanics to maintain, on top of the logistical benefits derived from stocking parts for a single platform instead of a diverse fleet of aircraft. Since every crew is trained on the 737 it’s easy to adjust staffing and make substitutions that suit the business rather than being required to retain specialized teams. The same benefits are available to business leaders when your storage array, regardless of capacity or performance profile, presents a unified approach to data management even as the hardware or the people managing it inevitably change.
Data Coherency – All NetApp systems write and protect data with the same RAID level, RAID-DP. Disks are combined to form flexible pools called aggregates, and these aggregates can survive two simultaneous failures of any disk in the aggregate. Because all NetApp systems use RAID-DP, the guesswork about the appropriate RAID level for a workload type is removed. With a data protection level identical to RAID-6 and a performance profile similar to RAID-10, management complexity and overhead is dramatically reduced. Now anyone can set up storage aggregates without being a ‘storage guy’ and having to fully understand the lifetime trajectory of the workload being placed on a group of disks. RAID-DP aggregates can be expanded on the fly to enable seamless growth and continuation of services, persisting even through major upgrades.
Data Integrity – NetApp snapshots leverage the redirect-on-write methodology for tracking changes to snapshots. By tracking metadata about the individual blocks of data being stored, the file system knows where everything is and by making a copy of those pointers, and not the data, an instantaneous image of the entire file system can be captured. By design NetApp systems are optimized for write performance by sequentializing IO in system NVRAM before writing on the data to new blocks instead of inefficiently re-writing old blocks. Instead of traditional snapshot methodology that writes one copy of the original data to a snapshot reserved space, NetApp snapshots write only the changed data to new blocks while leaving snapshotted data in place. As snapshots expire according to customer-specified retention policies, space is freed from the array.
Deterministic IO – New in Clustered Data ONTAP is the ability to establish deterministic IO (Storage QOS) for any workload. This gives administrators the ability to establish IO ceilings and floors for workloads to ensure that during times of contention that critical work still gets done to meet business needs. Storage QoS workload management allows an administrator to control the resources that can be consumed by storage objects such as volumes, LUNs, VMDKs, or Storage Virtual Machines (detailed later) to manage performance expectations for individual workloads or even business units. You can set throughput boundaries expressed in terms of MB/sec (for sequential workloads) or I/O operations per second (for transactional workloads) to achieve fine-grained control, and features tight integration with VMware’s Storage IO Control (SIOC) capabilities.
Data Mobility – The crown jewel of Clustered Data ONTAP is it’s entirely unique methodology for moving volumes around the cluster. NetApp systems running cDOT may be connected with a dedicated 10 Gb Ethernet network between nodes, along with a software layer above some or all of the nodes operating in the cluster. This software layer is referred to as a global namespace, and NetApp refers to this software layer as ‘Storage Virtual Machines’. Instead of your VM environment or end-users connecting to physical NetApp hardware to read and write data from, they instead always connect to Storage Virtual Machines running on the NetApp controller. This Storage Virtual Machine handles data protocols as well as the connectivity to the physical disk subsystem and entirely abstracts physical controller/disk resources from the end user. Because endpoint connectivity stays the same, an administrator may online relocate volumes from one NetApp controller pair to the other, as well as between different types of disks. This is possible because Storage Virtual Machines may span across physical resources and a low latency contention free 10 Gb network allows for movement of data. This data movement requires no interruption to services or reconfiguration of servers and is unique within the storage industry.
Tego Data is a NetApp Platinum Partner and collaborates with businesses of all sizes take a holistic view of emerging and established datacenter technology to find the right solution for your project’s goals. More information is available at www.tegodata.com