Any storage consolidation effort must involve a serious consideration of disk subsystems and the specific disk types that those systems will contain. As an example, there may be hundreds (even thousands) of file servers distributed across the organization. A proliferation of individual servers can be difficult for storage administrators to utilize and maintain, often leading to wasted space. A storage consolidation initiative might move to replace those disparate servers with a single large network attached storage (NAS), which is easier to maintain and manage as a centralized storage resource.
Storage types and management
The goal of storage consolidation is typically to store more information using fewer disks or systems, so it's inevitable that systems slated for consolidation will support many more disks. Each disk is selected for its mix of storage capacity, performance and reliability. For example, Fibre Channel (FC) drives normally offer the best performance, but they cost the most and have the lowest capacities. SATA drives represent the other extreme, offering the largest capacities and lowest costs, but with the lowest performance. SAS drives fall somewhere in between but have not yet achieved broad acceptance. It's important to select disks and storage systems that can maintain or improve on current storage service level agreements (SLA) for both users and applications.
Tiered storage has become an important element of the storage enterprise, allowing data to be classified and stored according to its relative importance to the enterprise. Traditionally, this involved multiple storage systems. For example, Tier 1 storage may have resided on a high-performance Fibre Channel storage array, while Tier 2 nearline storage and Tier 3 archival storage may rely on an array of SATA disks. Today, storage systems can often support more than one disk type, allowing multiple storage tiers within the same system, further reducing the number of disk systems needed in an enterprise. See our article on tiered storage for more detailed information.
Consider the amount of management required for the new storage platform. Ideally, the storage consolidation process should simplify management by reducing the number of systems that need to be managed, but the new consolidated system will have its own management demands and GUI. If multiple heterogeneous storage systems are being introduced, be sure that the management software can support all of those systems through a single interface.
Any time more data is stored on fewer disks, it's important to consider the role of data integrity or protection in your storage consolidation strategy. RAID is typically the only form of data protection at the hardware level within a storage array, so match the RAID level to your protection requirements. RAID 0 offers simple striping for performance improvement, while RAID 1 provides mirroring. For today's storage arrays, RAID 5 handles single parity disk protection and RAID 6 (sometimes dubbed RAID-DP by proprietary vendors) supports dual-parity RAID that can protect against two simultaneous disk faults. RAID 6 is particularly useful for SATA storage systems where the probability of multiple disk faults is substantially higher than Fibre Channel systems. When a disk fault occurs, it will take time to rebuild the failed disk from parity information and the other disks in the group. Understand the implications of rebuild time and its impact on array performance. Aggressive disk diagnostics and pre-emptive rebuild features can help to keep rebuild times shorter.
Remember that almost all RAID implementations will reduce the overall storage capacity. For example, RAID 1 will use double the disk space because every disk is mirrored to another. RAID 5 requires one additional disk for the RAID group's parity data and RAID 6 requires two additional disks. This will impact the effective storage consolidation in your array.
Data migration and backup
Storage consolidation almost always implies data migration in order to move data from existing disparate systems to the centralized storage platform. However, migration takes time and it's usually disruptive to the production environment. Even though most storage systems provide migration tools to ease the process, it's important to plan the migration between heterogeneous storage systems in order to minimize service disruptions. Systems that cannot be migrated with automated tools require even more direct attention from IT staff.
Data backup and disaster recovery processes will also be influenced by consolidation, so plan any changes needed to ensure that the new data locations or systems will be protected properly. Data replication is also a means of protecting data, so new disk systems intended for storage consolidation should be able to duplicate their data to another local storage system or across a WAN to a storage system in another location. Most disk arrays include some amount of native replication capability.
Decommissioning older systems
Storage consolidation normally replaces one or more existing storage systems. Some organizations choose to liquidate older storage systems, while others may opt to redeploy the older hardware at a backup/disaster recovery site, reinstall the older hardware as a lower tier or relegate it to a lab/testing service. The key here is that older hardware may still have value to the organization, and any consolidation initiative should include plans to decommission or redeploy that leftover equipment.
Also consider the impact of power and cooling on a consolidation plan. When the new storage system is installed, it will demand additional power and cooling until data is migrated and the new system is "cut over" into regular service. After older equipment is removed, those power and cooling demands will drop -- possibly below previous demands. Remember to provide adequate power and cooling during the transition and into the future if necessary.