Published: 12 Nov 2005
All in all, midrange array users are a satisfied bunch, but the winner of the Quality Award in this category might surprise some people.
Midrange storage arrays aren't just for smaller companies and remote offices anymore--they've earned their place in the data center right next to enterprise arrays. The trend toward tiered storage environments has taken root to the point that Diogenes Labs estimates 70% of Global 2000 companies have a defined multitier storage strategy. Although the cost of midrange storage arrays is 30% to 50% less than enterprise storage on a per-terabyte basis, many organizations have enough midrange storage that they spend in absolute dollars the same amount for midrange and enterprise storage. Moreover, based on conversations with users and vendors, we expect the share of midrange storage to continue growing to the point that the split between the two types will be 60% midrange to 40% enterprise by 2008 to 2009. In more than 50% of organizations, the midrange and enterprise array vendors are different, indicating little technical advantage to having a single supplier for both.
The following midrange arrays were included in the Quality Awards survey:
The top midrange array is ...
Long recognized for its dominant market position among tape products, StorageTek doesn't necessarily come to mind when considering excellence in disk products. But that perception should change. In our survey, StorageTek not only won, it dominated the competition in achieving a No. 1 ranking in every survey category. (StorageTek's midrange disk products are manufactured by Engenio Information Technologies, as are several other vendors' products in our survey.) "StorageTek has always been kind of a trailblazer of technologies," says Dan Evans, director of technical services at Cleveland's St. Vincent Charity Hospital. "I think StorageTek has a lot of benefits to bring to Sun--it'll be interesting to see how those product lines evolve as a result."
NetApp finished a strong second, placing in that position in every category except sales-force competence (where it came in third). In fact, StorageTek and NetApp were the only vendors with products that placed in the top half in every category (see "Overall ranking of midrange arrays," next page).
Users apparently have a much more favorable impression of midrange arrays than enterprise arrays. In our enterprise array survey, the majority of responses fell between a rating of 4.7 and 5.7 (4.5 is the neutral marker on an 8.0 scale). Average responses for the midrange survey were a point higher, falling between 5.7 and 6.7. For the first time in any of our surveys, a product achieved a category rating above 7.0: StorageTek did it twice, earning a 7.11 for initial product quality and a 7.03 for product features. As impressive as that is, all of the products received high scores, reflecting a high level of satisfaction in a competitive field. Even EMC's Clariion, which was the only product to place in the bottom half in every category, had an overall rating of 5.75, which is substantially above the 4.5 neutral marker. Nevertheless, readers who have followed these Quality Awards will note that EMC had the lowest ranked product in both the enterprise and midrange array surveys.
The sales-force competence category was designed to assess a vendor's ability to sell the proper system. Inappropriate products will result in dissatisfied users. Doug McDaniel, IT director at St. Louis-based Maritz Inc., experienced some disappointment in dealing with EMC's sales force. "They're very aggressive in their sales approach," he says, "and some of the things they promised don't necessarily pan out after the product is on the floor."
But Evans, who also has EMC Clariions in his shop, likes working with EMC and feels the firm anticipates his needs well. "I'm getting notifications of things that I'm thinking about," says Evans, "and I'm getting that information without actually calling them."
In formulating the survey, we surmised that a significant number of systems were acquired through resellers and we wanted to know if this would influence user satisfaction. Only 35% of the systems in our survey were bought through resellers, with the rest purchased directly from the manufacturers. We didn't find any trend with respect to particular resellers, and the data didn't provide any conclusive evidence that one channel or the other resulted in greater satisfaction.
StorageTek, the winner of the sales-force competence category, had a 50/50 split between sales through resellers and direct sales. No vendor had more than 50% of its sales through channels. Sun had the highest percentage of direct sales (81%).
One significant result was that the lowest top and overall scores in our survey were for sales-force competence and technical support. This is an important finding as the scores reflect the "people" aspect of the survey, whereas the other categories (product features, initial quality and reliability) are primarily "technology" issues. Vendors that wish to improve their scores should therefore focus on sales-force competence and technical support.
The product features section determined how well each product meets a user's needs based on their perception, rather than on a list of functions. Our survey inquiries focused on software functionality. StorageTek's 7.03 was not only the highest score, but was more than a half-point higher than second place NetApp (see "Features ranking of midrange arrays," previous page). StorageTek was almost a full point higher than IBM and 1.11 points higher than Sun, even though the respective hardware is basically the same. For the statement "Overall, this product meets my needs," StorageTek's 7.24 score was more than a full point higher than for IBM (6.10) and Sun (6.17). In fact, StorageTek scored a 7.0 or higher on five of the seven evaluated items in the category; no other vendor reached the 7.0 mark on any item in this section.
Scalability appears to be less of an issue for midrange users than for users of enterprise arrays. For the statement "This product's capacity scales to meet my needs," all of the products scored a 6.0 or higher, with StorageTek topping the group at 7.15, easily beating out HDS and IBM (who notched solid scores of 6.39 and 6.38, respectively).
Initial product quality
The initial product quality section assesses how easily a product is installed, configured and put into production. In this section, one key statement, "This product delivers good value for the money," generated a 7.35 for StorageTek. NetApp and HDS tied for second with a 6.13. IBM, however, had a relatively low score of 5.51, which was the lowest of the group (see the related graph "Quality ranking of midrange arrays," above).
|About the survey|
The Diogenes Labs–Storage magazine Quality Award for midrange arrays complements our Quality Award for Enterprise Arrays (see "Quality Awards: Enterprise arrays," Storage, August 2005). The Quality Awards identify and recognize products that have proven their quality and reliability in actual use. It's important to note that the awards aren't based on the subjective opinion of a few individuals, but are derived from a survey of qualified Storage readers that assesses products in five main categories: sales-force competence, product features, initial quality, reliability and technical support. Our methodology incorporates statistically valid polling that eliminates market share as a factor. Indeed, our objective is to identify the most reliable product on the market regardless of vendor name, reputation or size. Products were rated on a 1.0-to-8.0 scale, where 8.0 is the most favorable possible score.
The midrange array survey was completed by 339 respondents. Respondents were allowed to evaluate more than one array if they indicated a qualification to do so. Thirty-six percent of midrange respondents evaluated only one system vs. 26% in our earlier enterprise array survey. The study's margin of error is +6% with a 95% confidence factor.
Although product reliability generated the fewest number of average scores of 6.0 or higher, every vendor among the finalists scored a 6.0 or higher on the statement "This product meets my service-level requirements." Interestingly, every vendor, except for IBM, had its lowest score in this section in response to the statement "Patches can be added non-disruptively." HP, which had respectable scores on most items, scored a surprisingly low 4.95. However, most vendors scored well for the statement "This product does not require unplanned patches." From these results, it's clear that users aren't unhappy with product stability; however, every vendor would be well served to improve its patch management process.
With both NetApp filers and HP EVAs in his data center, Enron's Ambrocik is pleased with both offerings. "From a reliability point of view, we run a lot of litigation data through the [NetApp] filers and we also run Exchange on the filers," says Ambrocik. "From an uptime perspective, we're very happy."
Support is typically the area where users cite their highest levels of dissatisfaction; but overall, our survey indicates that midrange users are generally pleased with the system support they receive (see "Technical support ranking of midrange arrays," below). In this category, the StorageTek juggernaut continued, with an overall score of 6.45, highlighted by a score of 7.0 for the statement "Vendor supplies support as contractually specified" and a 6.9 for "Vendor takes ownership of the problem."
NetApp's support gets the thumbs up from Tom Morrier, a systems analyst with Inco Ltd., a mining company in Sudbury, Ontario. "They usually let me know about problems before I know I've got them," says Morrier. He also likes the way NetApp monitors his systems and reacts quickly to problems. "I actually had a [drive] failure in the middle of the night," says Morrier. "When I came into the office to start my regular day, there was a drive waiting for me."
One EMC Clariion user, a storage administrator for a graphics hardware and software firm in the Northwest, says EMC's after-hours support is a big problem for his company. "I've had a number of cases where I have started with offshore support, worked through two cycles of offshore support and come back to U.S. support," he notes. He typically has to explain the problem again for EMC's U.S. support team. "I've spent all night working through Ireland and Singapore to get back to the States," he notes.
The lowest overall support score was EMC's 5.71, but four vendors all scored a 6.0 or higher--HP, NetApp, StorageTek and Sun. This is in stark contrast to the results of the enterprise array survey, where HDS' 5.76 was the top score.
But even among highly satisfied midrange users, training is an issue. "The vendor provides adequate training" was the only statement for which no vendor scored a rating of 6.0 or better--the best score was StorageTek's 5.95, while the lowest was a 5.43 for HDS.
Summing up the scores
Although EMC and Dell have similar systems in nearly every respect, including third-party field support provided by Unisys, the two didn't track together in our survey. Dell finished third overall, but varied between second (sales-force competence) and seventh (initial quality). Because our methodology includes user perception as a factor in the ratings, we hypothesize that perception is the key difference. In the case of Dell, Clariion is its top-of-the-line product, whereas Clariion takes a back seat to Symmetrix in the EMC product line. However, the products tracked closely on both initial quality and technical support.
The initial quality section of our survey showed the largest spread among vendors, from StorageTek's 7.11 to EMC's 5.7. Similar to the EMC/Dell situation, the scores from StorageTek didn't track with IBM and Sun, even though the hardware is the same (the bundled software varies by vendor). Our hypothesis is the same in this situation, given that the D-series and FLA are the primary array products for StorageTek (mainframe products weren't included in our survey). However, the IBM FAStT/DS series and Sun 6920 products also have enterprise-array siblings.
As part of our analysis, we compared the scores of products in cases where more than one product was evaluated by the same respondent. Having two vendors vie for business is a key part of some storage shop's strategy. A two-vendor strategy "is critical for us, especially when it comes to negotiating," says Maritz's McDaniel. "That's helped us immensely [in getting a lower price]."
In this head-to head comparison, StorageTek was preferred in 63.6% of such cases, while NetApp was preferred 51.4% of the time. All other vendors fell below 50%, with Dell the lowest--just 17.6% of respondents prefer its products. We also asked respondents, "All things considered, would you buy this product again?" More than three-quarters (77%) said "Yes." This indicates a high degree of product satisfaction regardless of a vendor's order of finish. StorageTek lead among our vendors with 86.4% positive responses, but Sun was a close second (85.7%). Even though EMC trailed the group in the overall standings, it had a strong 73.6% buy-again rating (see "Would you buy this company's array again?," previous page).