Is the Mainframe Cool Again?
Only a few years ago, analysts and pundits were declaring the mainframe a dead technology platform. Distributed computing had won the war for the hearts and minds of business leaders.
Trends seemed to reinforce their view of the end of the mainframe era.
-
Demographic analysis suggested that the average age of the mainframe worker was 53. With their retirements looming, a mainframe skills shortage was widely anticipated. Simply put, the lack of mainframe "cool" was seeing collegians sidestepping university degree programs that focused on mainframe skills. Despite efforts of companies like IBM and CA to promote mainframe studies in colleges and universities, the appeal of the mainframe disciplines couldn't hold a candle to that of Java, the Web, or Google apps.
-
Survey after survey found the majority of business IT planners plotting a speedy exit from Big Iron, migrating workload once processed there onto shiny new X86 based servers.
-
A fundamental shift in the mix of data being generated by business favoring files over transactional database output was seen as the nail in the proverbial coffin. Mainframes were, after all, optimized to be the corporate calculator and not the corporate filing cabinet.
-
And let's not forget cost. Buying a mainframe represented a significantly greater CAPEX expenditure than buying a server. Software licenses for mainframes were exhorbidantly expensive compared to server computing. In truth, whatever economies favored mainframes (and there are many) most business people did not perceive them: mainframe budgets were presented in a coherent and unified way, while defining the real costs of distributed computing costs required forensic accounting skills.
While the case against mainframes has not disappeared, it has diminished of late. But the question remains, can the mainframe reclaim its technology cool again?
What Has Changed?
Demographic arguments against the technology have shifted subtly to reflect a new and nagging reality: even distributed computing confronts a skills shortage going forward. Since the end of the dotcom era, enrollments in all computer technology degree programs have fallen off a cliff. Part of this trend is related to the gender mix in colleges and universities, where women now outnumber men. Information technology is widely perceived as a "bloke culture" and hence a less attractive career path for women. Another factor is the younger generation's view that tech is not the path to instant wealth that it was during the 1990s.
New survey data shows a reversal of plans to migrate workload off of the mainframe. Partly this reflects a slowing economy, since re-platforming ERP and mission critical databases to servers costs real money. Some surveys suggest that companies have defunded those projects for now.
There is also some evidence that a reversal of thinking about mainframes is afoot that actually favors the migration of some workload previously handled on distributed computers back onto the mainframe platform. The reasons are simple and related to the current cost-consciousness brought on by the Great Recession. In the words of one medical CIO: "We have six people running our mainframe and it has been operating without downtime for almost five years. We have 500 staffers in our distributed systems environment and their platforms are down at least once a week."
With companies seeking to contain costs by making fewer staff more productive, dependency on automation is at an all time high. The resiliency of the mainframe and the discipline of mainframe operations in terms of both the professionalism of its operations staff and their preoccupation with procedural formality is making the mainframe shop more attractive in lean economic times.
There is mounting evidence that the distributed systems vendors have shot themselves in the foot with their preoccupation with innovations that lack a strong business value case.
- Fibre Channel "SANs" were a painful reinterpretation of Enterprise Network Storage Archictecture (ENSA) concepts first developed at Digital Equipment Corporation then introduced by Compaq following the DEC acquisition. Storage Area Networks weren't really networks at all and delivered none of the ENSA promise. Instead, they were simply the most expensive way ever invented to host data, benefitting their vendors far more than consumers.
- Server virtualization is tracking along much the same curve. While it might make sense to economize on distributed computing resources by stacking applications in a hypervisor-driven server, the impact of the strategy on application performance, resiliency and cost is the elephant in the room. For all the noise used to promote the technology, the business value case is increasingly in doubt.
Mainframes Get Their Mojo Back
For their part, mainframe vendors have proven to be more agile competitors than the distributed computing vendors anticipated.
- IBM now seeks to convince business planners that the greater scalability and resource control in mainframe environments trumps the strong performance numbers touted by distributed server vendors. Mainframes can, in fact, do more with less.
- IBM argues that Logical Partitions (LPARs) in a mainframe deliver more value than hypervisor-based server virtualization. It should: LPAR technology has a 30 year head start.
- IBM has been building a case for file system-centric storage on mainframes, first by embracing Linux, then by introducing storage peripherals and network file system supports to enable file data to migrate to the mainframes as readily as they can be hosted on distributed storage.
- And leading mainframe software vendors like CA Technologies have been working diligently to simplify and refine the mainframe experience.
Here are a few examples of what the vendor is doing.
CA Technologies Mainframe Software Manager (MSM) brings software installation and deployment into the 21st Century by borrowing the concept of InstallShield from the distributed computing environment. This brief interview with Marie Godfrey, Senior Principal Product Manager at CA Technologies, provides an overview.
CA Technologies Mainframe Encryption Key Manager continues to evolve into a universal and standards-based toolset for unifying the management of encryption keys used to secure data per federal and state mandates -- an increasingly important issue for business today. This quick interview with Tim Bruce, Principal Software Architect, gives the highlights.
CA Technologies Mainframe Virtual Tape (VTAPE) is providing a cost-effective solution for facilitating data protection and enabling both disk-to-disk replication and disk-to-tape backup without investing in a costly VTL stovepipe. This brief chat with Principal Software architect, Dave Helsley, provides the details.
Last but not least, CA Technologies Mainframe Chorus is possibly the most exciting thing to happen to mainframe computing in years: a comprehensive refresh of the mainframe experience that addresses, and potentially solves in one fell swoop, the challenges of the aging mainframe workforce, the appeal of mainframe computing for IT newbies, and the changing roles of workers within the mainframe computing environment today. Watch this brief interview with Troy Coleman, Principal Product Manager for CA Technologies Mainframe Chorus, then take a read of a white paper we developed about Chorus.
Can CA Technologies Mainframe Chorus and the other innovations occurring in the mainframe technology space reinvigorate interest in the mainframe computing platform? They can certainly help. In fact, in an ironic turn of events, current cloud computing woo is actually helping to bolster the case for mainframes. If you listen closely, vendor descriptions of the "perfect cloud" devolve invariably into a litany of descriptions of the features and functions of -- and the business value case for -- a compute platform that we already have today: the mainframe.
Main Menu
Login Form
