Wikibon is advising clients that migrating existing databases to new stand-alone storage systems is often of marginal value at best. Specifically, Wikibon CTO David Floyer writes, Oracle Corp. customers using NetApp, Inc.’s, ONTAP 7-Mode storage should carefully assess migration options before moving their data to NetApp Clustered ONTAP.

The break-even point for most of those migrations is over 50 months into the future and the internal rate of return (IRR) is an anemic 12 percent at best, Floyer estimates. Such migrations are inherently risky, and customers would be better advised either to stay on their existing storage for now or move either to a server SAN or true private or public cloud. Overall, enterprise IT should consider pure play storage (and other component) vendors to be tactical, not strategic partners unless the provider can demonstrate the ability to improve business results dramatically.

Floyer offers a detailed analysis of ONTAP customers’ choices in “Assessing the Cost of NetApp ONTAP Migrations in Oracle Environments,” ending with a consideration of NetApp’s potential future in a rapidly changing environment that is moving toward converged and hyper-converged systems and private and public cloud platforms.

Workload Breakdown

Floyer breaks down workloads into five categories and provides recommendations for NetApp ONTAP customers for each, with a particular focus on Oracle workloads:

1. Latency-sensitive workloads: An Oracle engineered system is the best target for migrating these workloads. In most cases this provides the lowest cost and least risky option, and these systems offer many of the features of true private cloud. For very mixed latency sensitive environments, the best strategic option is an interim all-flash scale-out solution such as Kamlinario, Inc.’s K2, Dell, Inc./EMC Corp.’s XtremIO, NetApp’s SolidFire or Oracle’s FS when that becomes available. As true private cloud systems increase in flexibility, they will become the best strategic option.
2. Standard SAN workloads: These are the design point for Clustered ONTAP and migration from 7-MODE will allow a decrease in headcount. However, the cost of migrating is significant and return on investment low, making the migration of doubtful benefit. Wikibon recommends that clients focus on migrating these workloads to true private clouds as they become available.
3. High-bandwidth filers: NetApp’s write anywhere file layout (WAFL) architecture is not well suited for workloads that are constrained by bandwidth and data write speeds. If bandwidth requirements are acute, an alternative solution is required. Wikibon’s research shows that the Oracle ZFS NAS appliance is “in a league of its own for high-bandwidth, write-heavy Oracle backup, ETL [extract/transform/load] and batch environments.”
4. Standard filers: Cluster ONTAP provides several advantages for these workloads. However, migration is a significant disruption, and alternatives including Dell’s scale-out solutions, EMC Isolon and Oracle’s ZFS will often offer lower cost solutions as interim platforms. This area will also be covered by public cloud and true private cloud offerings, which will provide more compelling business cases.
5. Capacity storage and WORN: Write Once, Read Never or almost never (WORN) applications such as long-term data archiving require very low-cost storage. NetApp Clustered ONTAP is overkill. No traditional array can compete in this area against very low cost basic archiving solutions from public cloud services. If a private solution is mandated, the best solution is flape (flash/tape), in which the extensive meta-data needed to access specific data in the archive is kept on an all-flash front-end while the data itself is stored on low-cost tape.

Whither NetApp?

Floyer concludes his analysis with a look at NetApp’s future. The company, he says, has “great products and a loyal customer base,” and its markets move slowly. However, its traditional business is being completely disrupted by server SAN, true private cloud and public cloud economics. Over time NetApp, like all IT component companies, will see a steady decline in its business. Floyer identifies five choices for NetApp’s longer term survival:

1. Manage declining revenues as a traditional cash cow.
2. Focus on gaining market share from EMC and the other storage vendors.
3. Become a systems vendor by acquiring or merging with a systems platform vendor and integrating the NetApp Cloud software.
4. Become an OEM supplier of NetApp Cloud software while becoming a high-functioning cloud software provider.
5. Follow EMC’s strategy and sell the company to a larger vendor that needs a strong storage division.

The number of organizations that have successfully disrupted themselves is short, with Netflix and Charles Schwab as stand-outs. Both achieved this trick by creating an impervious wall between the traditional and disruptive business operations. NetApp, Floyer suggests, could combine options one and four with such a wall between them as a strategy for self-disruption.

CIOs should move away from strategic partnerships with vendors providing individual technology components and move toward partnerships with true private cloud and public cloud providers. They should push responsibility for design, testing and full maintenance to these cloud service providers, Floyer concludes.

Graphic courtesy NetApp, Inc.