Part 1 - The Rise of the Utility
Mark Hurd, President of Oracle, made some extraordinary claims regarding the performance of new computing equipment from Oracle. He told members of the press and attendees of a plenary session at Oracle Open World about the substantial improvements with in-memory computing and throughput capabilities of the revamped ExaData, ExaLogic and ExaLytics machines.
He said that reports could be produced as much as 100 times (not 100%) faster in some cases. He argued that data can be compressed by as much as tenfold. He claimed that certain workloads could be processed as much as 2000 times faster than previously accomplished. He described machines capable of holding approximately 26 TB of in memory DRAM. When several machines are clustered together, in memory databases of 222 TB (via flash memory) are possible.
For the Oracle readers, I may not have captured every number exactly right at the event (nor may we ever see these numbers verified by an independent source) but that is not the point. It's the magnitude of these processing and memory achievements that sends a signal change in the market: computers may be capable of a scale and have gotten so powerful that we have entered the world of utility computing.
Oracle is not alone in achieving massive improvements in processor speed, in-memory computing and other frontiers. Firms like IBM and Hewlett Packard are moving in similar paths and are also introducing ever more powerful computers.
But, if you are a CIO of a small to midsized firm, are you willing to put forth the capital expenditure required to buy one of these 7 foot tall mega machines? That decision is looking ever more doubtful as these CIOs will turn more and more to large utility providers of computing capabilities. Likewise, the market for these mondo-machines may be moving to cloud provisioners, systems integrators, hosting firms, large governments, very large businesses and some universities. It’s the old Cray super-computer space all over again.
I recently conducted a webinar with the CIO of a very large credit union. His enthusiasm for moving his applications to a third-party cloud service center was equal parts infectious and enthusiastic. He has already witnessed firsthand the benefits of relying upon the scalar capabilities found in the solutions from large cloud provisioners. Granted, he is but one of some 1400 other customers using this facility, but, he is able to take advantage of the cloud provisioner’s deep, broad and talented bench of individuals, their disaster recovery capabilities, their enhanced security capabilities, etc. And, better still, it costs less, too.
The old world, where every business had its own data center, its own programmers, its own security protocols, etc. may be coming to an end. We should all liken this to what occurred a few centuries ago when local artisans and craftsmen in most every town and village were displaced by more powerful, larger scale, industrial age mechanized factories. It was scale and access to powerful water or coal-fired power systems that facilitated the Industrial Revolution. There’s a reason why few businesses today have their own power generation systems. They get their power from firms that make power generation their core competency. In today's technology world, the introduction of new in-memory or flash memory-based computing systems brings scale and power to a massively new level.
If I were a CIO of a small to midsized firm I would get out of having a data center. For the same reason that very few businesses today have their own water wheels or coal fired power generation plants attached to their business. Modern firms will likely begin to transition away from their own internal data centers.
The implications of this change will be profound. Hardware vendors must re-focus their sales efforts to cloud solution providers and less to the corporate world. Application software vendors will likely find cloud provisioners (i.e., firms offering cloud computing services) a surprising force in software negotiations. Why? Smart cloud provisioners will want to wield their market aggregation power and negotiate deep software discounts for their customers. This gives their service a competitive edge. Middleware and database vendors will likewise find themselves dealing more with cloud provisioners, too.
Software consumers will use more of the spot computing capabilities of these utility providers as they embrace more analytic applications, especially those that rely on big data. But, they won’t use their own tech to do these analyses. It just won’t be cost effective to do so.
Bottom line: on-premise is fading and utility computing has arrived. Cloud powered applications will be in ascendancy.
(Continue to Part 2 - The Permanent Leaderboard Change in ERP)