There are two different meanings to SAN storage consolidation. The first reduces the number of multiple, separate SAN fabrics (networks) into fewer and larger fabrics. The second reduces the number of disparate SAN storage systems into fewer larger and more manageable SAN storage systems. This tip focuses on important considerations of the latter.
Here are the critical questions to answer and plan for when considering a consolidated SAN storage project:
- How will the consolidated SAN storage scale to meet the future needs of the organization?
- What's the realistic usable capacity per system that can be expected based on each required RAID group (e.g., the amount of capacity that can actually be used by the applications)?
- What type of capacity management is built in or required with third-party software?
- What tools are provided for predictive performance and capacity management?
- Is the system management significantly simpler than the storage systems being replaced? In other words, will it be intuitive to the current storage administrators?
- Does the system management make it easy to troubleshoot?
- How will it provide actual logical unit number (LUN) expansion -- dynamically or with thin provisioning?
- How does the consolidated SAN storage support the required throughput and IOPS (now and as future requirements)?
- How will it simplify management (policy-based thresholds, alerts, automated actions based on admin policies, etc.)?
- How does the SAN storage consolidation manage application performance prioritization? Will it enable secure separation? (An example of this is LUN masking, where one application can't see or access another's data or storage.)
- Where are the potential single points of failure, if any? If single points of failure are acceptable, what will be put in place to mitigate failures (dual systems, dual fabrics, multipathing, load balancing, etc.)? If one application goes down, how will the others be impacted (if at all)?
- How will the consolidated SAN storage system minimize application and user disruption when migrating from multiple systems?
- What does the consolidated SAN storage system do to meet the organization's data protection requirements?
- Is the consolidated SAN storage certified with all of the applications that will be running on it?
- How will the consolidated storage minimize silent data corruption? What type of end-to-end error detection and consolidation does it provide beyond RAID?
- What type of self-healing does it provide for the hard disk drives beyond RAID? How will it reduce application disruptions from RAID rebuilds and potential data loss?
- Will the total cost of ownership of the consolidated SAN storage (that is, operating expenses [OpEx] and capital expenditures [CapEx]) be less than the multiple SAN storage systems? If not, how will the cost be justified, if at all?
If the answers to the above questions are all positive or mostly positive, you can move on to the following preparation and implementation items.
- Pick the vendor(s), partner(s) and system(s).
- Plan the implementation in great detail.
- Provide implementation checklists for every department affected, including servers, SANs, networks, cable, plant, applications, etc.
- Communicate in detail with all of the departments, administrators and application owners about the transition. Over-communicate if necessary.
- Document everything. You'll learn from any of your initial mistakes and that knowledge will help you during any further consolidations. Make sure to always capture, quantify and correlate all projected consolidation savings to determine if your promised value does indeed exist. This needs to be done if there's to be any credibility for future consolidation efforts.
In the end, a successful SAN storage consolidation project comes down to a clear understanding of current/future requirements, solid preparation, effective planning, efficient implementation and disciplined follow-up.
This was first published in March 2009