What you will learn: How to avoid common NAS upgrade mistakes.
NAS upgrades are usually deployed to add capacity or performance to file-based storage. While it's easy to see any additional capacity, users often fail to consider the added performance requirements. For example, it might make sense to replace two or three older file servers or NAS boxes with a single large NAS platform. Consolidation eases hardware maintenance needs, lowers power demands and can add substantially more storage capacity to the data center.
The problem is that some organizations add disks without considering the underlying performance needs. Having fewer storage systems is usually a plus, but now more users are vying to access that one consolidated NAS box. This leads to contention, which can impair file-based storage performance. NAS upgrade plans should extend beyond the notion of capacity. New NAS systems must include the additional connectivity and internal disk controller architecture needed to match an organization's needs for greater storage performance. And regardless of the vendor's reputation, you can't just take them at their word. In-house testing must be a core element of the purchasing plan.
Clustered NAS systems do not necessarily provide more performance or scalability than nonclustered NAS platforms. Businesses that accept that premise at face value may be disappointed by the results.
Data migration is another critical issue that gets overlooked during a NAS upgrade or consolidation cycle. Once the new NAS is deployed, data must be migrated from existing NAS systems. The migration process can be time-consuming, possibly resulting in substantial disruption to the business. You want to develop a data migration plan that will impose minimal disruption. If you intend to let the new NAS vendor handle your data migration, you should fully understand the business impact, and see that the vendor commits to all your migration objectives.
"Look beyond the basic speeds and feeds," says Greg Schulz, founder and senior analyst at the Storage IO Group. "Look at any applicable or relevant performance comparisons for different applicable workloads to get an idea of how those extra disks, ports, memory, controllers and nodes in a cluster actually work to deliver what you are expecting."
Stephen J. Bigelow is the Features Writer for SearchStorage.com
