Survey Says: Business as Usual for Big Iron
Shops with large mainframe investments tend to double down on Big Iron; small shops, on the other hand, are trying to wean off of them. Call it a case of expansion versus austerity.
Big Iron has seen its share of ups and downs over the last decade. In fact, the upward trajectory of Big Iron's resurgence -- its growth in revenues, MIPS, and other metrics -- has been constrained by downward pressure, too.
It goes like this: shops with large mainframe investments, measured on the basis of MIPS, are doubling down on Big Iron, expanding their capacities and exploiting Big Blue's mainframe technology dividend. Small mainframe shops, on the other hand, are going ever-smaller, weaning themselves off of Big Iron altogether. It's expansion versus austerity.
The seventh edition of mainframe ISV BMC Software Corp.'s Annual Mainframe Survey confirms this trend -- just as it has in virtually every year since its inception.
Large shops, BMC finds, are expanding their mainframe investments; small shops, on the other hand, are cutting back. More to the point, shops with big MIPS footprints typically have big MIPS growth rates: 15 percent of the survey's 1,243 respondents anticipate that their MIPS growth will exceed 10 percent over the next 12 months; 44 percent estimate their growth requirements at between 1-10 percent.
Of the combined 59 percent who are planning for growth, nearly one-third (31 percent) cite the capacity requirements of new and legacy apps alike. Legacy application growth is still -- by far -- the single biggest driver, accounting for almost 20 percent of anticipated growth; new workloads alone account for just 9 percent, according to BMC.
In surveys past, at least one-third of respondents typically reported flat MIPS growth; this number was down slightly in 2012, at 30 percent. The tally of shops reporting declining MIPS growth -- a combined 11 percent -- was up by more than a third from 2011 (when it stood at 8 percent) but in line with historical trends: 2010's tally was a combined 10 percent.
Elsewhere, a solid half of respondents -- 50 percent -- said that they view Big Iron as a long-term investment for both old-growth (z/OS, zVM, VSE, IMS) and new-growth (Linux, Java, database processing) workloads. Another 40 percent said they view the mainframe as a suitable long-term investment for existing (Big Iron-bound) applications.
A fraction of users -- six percent -- said that they are considering mainframe exit strategies. Another four percent of users fell into the requisite "Other" category.
Mainframe optimism was strongest in the Asia-Pacific region, where 57 percent of respondents say they anticipate deploying new workloads. Overall, 93 percent of respondents in the Asia-Pacific region envision either staying put on or expanding their Big Iron presences.
Among shops that say they're staying put on Big Iron, the mainframe platform's availability was cited as a key factor. Almost three-quarters (74 percent) of respondents cited availability -- up one percent from last year. Security, as always, was another key differentiator, although it was down slightly, year-over-year: this year, 70 percent of users cited the mainframe's superior security model as a key differentiator, down from 75 percent last year (71 percent cited security in BMC's 2010 survey, meaning that 2012's result marks a three-year low).
Other top vote-getters were: the mainframe's role as a "Superior Centralized Data Server" (cited by 68 percent of respondents, down from 70 percent last year); Big Iron as a superior transaction processing platform (65 percent, up from 62 percent); an ability to expose mainframe apps in new business applications (37 percent, down from 38 percent); and the mainframe as a player in Web services and SOA integration projects (39 percent, up one percentage point).
Intriguingly, nearly one-third (29 percent) of respondents said that they were staying put on Big Iron because they believed the ROI that could justify a move just isn't there. That was up by more than 11 percent since last year.
At first glance, BMC's findings with respect to the use or prevalence of IBM's zSeries Integrated Information Processor (zIIP), which debuted more than six years ago, seem especially intriguing.
BMC's numbers this year show zIIP use declining over the last three years: in 2010, for example, BMC counted 1,222 customers using zIIPs; in 2011, that number was down to 1,001. In 2012, just 824 customers are using zIIPs.
On the other hand, participation in the BMC survey has fallen off accordingly: in 2010, BMC collected feedback from 1,707 respondents; in 2011, more than 1,347. If zIIP adoption seems to have fallen off, it could be because the size of BMC's survey pool has likewise fallen off.
That said, the sweet spot of zIIP adoption seems to have shifted over the last three years. In 2010, nearly half (47 percent) of Big Iron shops hadn't yet deployed a zIIP (zIIP had been generally available for more than four years at this point, while the first zIIP-enabled products were shipping by the end of 2006) and 41 percent of respondents in 2010 had deployed 1-4 zIIP engines.
Today, fewer than one-third (32 percent) of respondents haven't yet deployed zIIP, while a full half have between 1-4 zIIP instances in-house.