In today’s economic environment, it’s critically important to make sure that any data storage system an organization purchases has an optimally low total cost of ownership (TCO) over its lifetime.
Many storage buyers today use capital expenditure (CapEx) as the sole factor for determining the value of a storage purchase. In doing so, they overlook multiple significant operating expenditure (OpEx) cost variables such as:
- Length of warranty period;
- Post-warranty hardware and software maintenance costs;
- Software licensing fees;
- Administration costs;
- Power and cooling costs; and
- Incremental cost of storage growth and the cost of lost business value caused by system downtime.
Further complicating the total cost equation is that these variables impact storage systems differently depending on the vendor. In the end, as David Vellante points out in his Wikibon article Storage CapEx vs. OpEx, OpEx is 64% of TCO, compared to 36% for CapEx.
In this article, I will explain several key TCO measures to calculate when evaluating the total lifecycle cost of your storage system. This should remove the complexity surrounding true cost of storage, help you purchase the best system for your unique environment, and improve your return on investment.
The CapEx Equation
CapEx is the easiest cost to determine since it is essentially the storage system’s sticker price. However, it’s important to not make a purchasing decision based on sticker price alone. A purchasing decision needs to take into account the underlying long-term cost variables. To better estimate the cost and manage the impact of CapEx, focus on these four primary strategies:
Purchase Only What You Need Now– Purchasing the full amount of capacity needed for the life of a storage system at the point of sale may be simple, but it’s an inefficient use of resources and will decrease the system’s utilization rate. By deferring hardware purchases, you benefit from declining hardware costs over time. Disk drive prices have historically declined 30% year-over-year according to Gordon Moore in his April 1965 article “Cramming More Components onto Integrated Circuits.” ftp://download.intel.com/research/silicon/moorespaper.pdf As a result, purchasing three or even five years of capacity at the point of sale could cost you up to 150% more over a five-year period than adding the capacity each year, only when needed.
The legacy view of purchasing storage capacity upfront is that it will be cheaper than making incremental add-on purchases and does not introduce downtime for subsequent installation. But with today’s storage technology, this is less of a concern and certainly not enough to justify such a TCO hit.
Reduce TCO with Higher System Utilization– There is a double benefit to purchasing storage capacity as it is needed. Not only does it reduce upfront and overall system purchase cost, but it also lowers OpEx by reducing electricity usage and software licensing fees through higher system utilization. Fewer disk drives means less maintenance, reduced licensing fees, and lower power and cooling requirements. However, you must ensure that you also get the I/O performance you need even as system utilization increases.
Few storage systems can provide this benefit and an incorrect choice may force you to accept lower system utilization not for capacity but for performance. That’s because some systems may only have sufficient I/O capacity when writing to or reading from the outermost cylinders of disk platters, a practice known in the industry as “short stroking.” Make sure you get the capacity and the I/O performance you need at full utilization. You are paying for the system, after all. Take full advantage of it and do not settle for less than optimal utilization.