For skeptics of the “putting all your eggs in one basket” strategy, history shows that building fewer, sturdier baskets can create strategic advantage. Take a look at the construction of airplanes. Bigger planes create economies of scale that make flying safer, faster, and more efficient. Oil shipments also move more safely and cost-effectively in double-hulled, automation-packed tankers than on large fleets of small craft.
Data storage is simply another successful application of this consolidation strategy. Market statistics from Gartner and IDC show storage capacity shipments increasing every year, even as the number of storage arrays shipped annually continues to decline. The math doesn’t lie. Enterprises are consolidating their storage systems because it works. It increases data storage areal density, improves performance, usable availability, and lowers costs by improving data center operational and financial efficiencies.
CIOs and IT directors are moving ahead with consolidation strategies because they’ve developed confidence in the availability, resiliency, performance and cost-effectiveness of high-end storage arrays. Their experience has led them to conclude that the benefits far outweigh the perceived downsides of putting their eggs in one basket.
The main reason enterprises are consolidating their storage systems is that concerns over the practice are being debunked.
For years the biggest argument against consolidation was that creating larger fault domains or “blast radii” was risky—especially in large, multi-petabyte scale environments. Putting storage in fewer but larger storage systems increases the number of users impacted by any single problem. Also concerning was the fact that recovering petabytes of data is intrinsically more difficult than recovering terabytes. More users are impacted and more time is needed to recover from a loss of data accessibility because recovery times are ultimately proportional to the amount of data being recovered. Consolidation also generally meant users had to invest in a higher-grade storage architecture, which generally increased their costs per terabyte.
Modern arrays that are increasingly reliable and resilient are putting many of these concerns to rest. As a result, enterprises are looking more closely at the benefits of consolidated storage strategies. With microprocessor performance increasing 5-6 times faster than data transfer rates on and off storage media, well designed storage systems can increase the ratio of terabytes per controller—without sacrificing performance. Decreasing the number of controllers and ports in the SAN infrastructure reduces the frequency of repairs and lowers power and cooling requirements.
Storage consolidation using modern high-end arrays simplifies infrastructure topologies. It does this by reducing the number of arrays being managed and the opportunities for misconfigurations—two major causes of downtime. Modern arrays are also efficient. They create larger pools of free space that improve operational efficiency relative to collections of smaller dual controller storage systems that may contain lots of stranded capacity: chunks of capacity that are too small to use and too difficult to aggregate into usable quantities.
Benefits of Consolidated Storage
Consolidating storage on fewer arrays benefits organizations in several ways.
It provides more usable availability by increasing the storage system’s—and, hence, the enterprise’s—ability to meet service-level objectives during software updates or controller hardware fails. Multi-controller architectures also reduce the inherent or implied performance degradation when one controller is offline. Multi-controller arrays don’t go into panic mode if one controller gets taken offline. By consolidating, you reduce the impacts of microcode updates and controller failures, and you’ve made your microcoding update process inherently more reliable.
Consolidation also simplifies the deployment of AI-augmented data analytics. At petabyte scale, the more data you have, the better you’ll be at training your systems and identifying patterns. Therefore, you don’t have to worry about different performance profiles. That simplifies the process of deploying new applications such as artificial intelligence-augmented analytics and statistical packages. Providing more access to the data makes the infrastructure more agile and usable.
Consolidated, higher-end storage deployments also are more cost-effective than a dispersed collection of mid-range arrays. While the base acquisition costs per terabyte are still typically higher for higher-end storage, the cost gap is narrowing, and the economics improve as the capacity increases. Further, the sturdier basket approach pays off by optimizing the use of free space and reducing costs related to downtime and managing the system.
Storage Consolidation Best Practices
Storage consolidation projects frequently fail because they challenge the status quo, and change creates winners, losers and risks. Following are some of the most important best practices that maximize the probability of consolidation projects being successfully completed.
- Don’t over-promise—Promise just enough benefits to senior management when trying to obtain their support to proceed with a consolidation project: better to over-deliver than over-promise.
- Build a cross-functional team—Build a team that can prioritize the infrastructure refresh design objectives. This should include all the important stakeholders: storage architects, operations, developers, finance and legal.
- Involve vendors—Give your vendors a vested interest in the success of your consolidation projects by having them share data migration costs and risks with you.
- Empower users—and gain their support by giving them the things that matter: higher availability, more performance and better agility with continuously improving price/performance.
Tilting the Scale to Your Advantage
Successful storage consolidation projects create opportunities to tailor the infrastructure to current and future workloads. When enterprises monetize the value of improved staff productivity, reduced downtime costs, and simplified D/R, the scale almost always tilts toward the deployment of high-end multi-controller storage arrays that are intelligent and self-managing versus staying with a collection of midrange arrays. Stated differently, the benefits of storage consolidation win.