The second core issue: to IOE save money?

If you do not save money, is it worth doing?  Why IOE can not significantly reduce costs? How to make all fairies understand this? I have been thinking. The most obvious example is not the case: There is no free cake in the world, there is no silver bullet, can not only remember to remember.  For example, to build a house, IOE infrastructure is a solid cement, the application is to build high-rise buildings on the cement, it is safe and stable. Some people say that cement is too expensive and monopolized by oligarchs, which has seriously affected the scalability of the building (too high to make it up) and even its stability. Now you have to remove the flyy cement and build the house on Not reliable yellow sand, yellow sand cheaper and no one can monopolize. Well, architects rack their brains how to build reliable applications on the base of no fly it? The core idea is that you can not build up a scaleup and build a scaleout; that capacity expands indefinitely, and once a building collapses, it does not collapse as a whole and affects only one bungalow; then, in order to do this That is, the building mode (from waterfall to agile), building architecture (IT architecture from single to distribution), building materials (from IOE to IOE), property (operation and maintenance to sre, operation and maintenance development, Comprehensive operation, and gradually realize AIOps ... ...), all aspects of reinvent the wheel.  In essence, it must be realized that if there is a need to build a reliable house on unscrupulous building materials, there must be another place to fly! Above all those aspects must be transformed and strengthened, have to spend money, especially the cost! Technically, everything has become software-defined, with no stable infrastructure and only a stable architecture. That software is not defined as money? Architecture is not relying on money out of it?  This is the only building materials to replace (hardware, mainly IE) is to save money, and the province is not the cost of investment. Or, go directly to IE after the cost of direct maintenance greatly reduced, but the ensuing overhead and software redefinition of indirect costs. IE away from the application, this is not too serious, close to the application of O this point is particularly prominent, so go O is the most difficult thing to IOE.
TCO point of view, due to the statistical size is not uniform, I can not give the detailed figures, but from an experience point of view, personally think that IOE overall significant investment savings, but the cost increase at least doubled, the overall TCO flat or slightly lower. The main savings in IE, and in the O point of view is likely to rise and fall TCO (see how the statistics, if the total sold by the comprehensive TCO should still be reduced, if by imperialism and rogue operators certainly more expensive .In addition to How to count, the purely cost point of view, must have risen substantially, doubled are conservative.