A lire sur: http://futureofthedatacenter.com/look-to-the-dinosaurs-to-see-why-the-data-center-wont-disappear-fast/
Technology pundits
often set up modern data strategy options in binary terms, as Cloud
versus On-Premise, for example. This has the advantage of making life
sound simpler than it really is, painting the picture of a religious
schism, a black-and-white choice. But the vast majority of organisations
have little interest in taking sides on one side or the other in IT
squabbles. They’d much rather do what’s best for the business.
Twenty-five years ago it was fashionable to say The Mainframe Is Dead and journalists wrote reams of copy on the subject, saying PCs and commodity servers based on Intel’s x86 processors would spell the end of ‘Big Iron’ host computing. Mainframes were mocked as dinosaurs and people lined up to say that they were brain-dead relics, legacy systems and the like. But what happened was that, although sales of PCs and servers grew very quickly, sales of mainframes and related software and services continued to grow quote nicely. At worst, sales were roughly flat and at other times – when e-commerce boomed, for example – mainframes did very nicely. That’s because their strengths of security, scalability and familiarity, backed by large, mature businesses, appealed to the businesses that ran on them.
Today, that Mainframe versus Client/Server-style war of words is being waged again but this time it’s Cloud versus Client/Server. We are told that on-premise client/server systems are dinosaurs and the people that use them are mocked as ‘server huggers’ and the like. But there are valid reasons for continuing to run data center operations on corporate premises: familiarity (again) is a big factor because companies tend not to like surprises and are willing to pay a modest premium to avoid them. But also there are issues of data governance and privacy that need to be worked through and there is a growing sense that, while they may initially offer low up-front costs, cloud services when all the add-ons are counted up, can be just as expensive as dealing with in-company server rooms.
Look back on any review of history in IT, the baby is never thrown out with the bath water. When the PC era came along, few companies switched off their mainframes. When the Web changed commerce, retailers did not shut up all their shops or call centres. When laptop PCs became available, desktop models continued to sell. When mobile phones appeared, most companies retained desk phones. The trend is always to peaceful coexistence where the incumbent technology, if it really does die, does so over a period of many years and even decades.
Some people say that Enterprise Software Is Dead, ERP Is Dead and so on. But they’re not, and neither are company data centers. In fact, many companies, probably a large majority, will use a blend of approaches, tapping resources inside the firewall as well as cloud services and third-party data centers, and matching each tactic to current needs.
And even for those who suggest that the on-premise model is a dinosaur, remember that the dinosaur ruled the world for quite a long time: the best part of 200 million years, give or take…
Jul 8 2013
Twenty-five years ago it was fashionable to say The Mainframe Is Dead and journalists wrote reams of copy on the subject, saying PCs and commodity servers based on Intel’s x86 processors would spell the end of ‘Big Iron’ host computing. Mainframes were mocked as dinosaurs and people lined up to say that they were brain-dead relics, legacy systems and the like. But what happened was that, although sales of PCs and servers grew very quickly, sales of mainframes and related software and services continued to grow quote nicely. At worst, sales were roughly flat and at other times – when e-commerce boomed, for example – mainframes did very nicely. That’s because their strengths of security, scalability and familiarity, backed by large, mature businesses, appealed to the businesses that ran on them.
Today, that Mainframe versus Client/Server-style war of words is being waged again but this time it’s Cloud versus Client/Server. We are told that on-premise client/server systems are dinosaurs and the people that use them are mocked as ‘server huggers’ and the like. But there are valid reasons for continuing to run data center operations on corporate premises: familiarity (again) is a big factor because companies tend not to like surprises and are willing to pay a modest premium to avoid them. But also there are issues of data governance and privacy that need to be worked through and there is a growing sense that, while they may initially offer low up-front costs, cloud services when all the add-ons are counted up, can be just as expensive as dealing with in-company server rooms.
Look back on any review of history in IT, the baby is never thrown out with the bath water. When the PC era came along, few companies switched off their mainframes. When the Web changed commerce, retailers did not shut up all their shops or call centres. When laptop PCs became available, desktop models continued to sell. When mobile phones appeared, most companies retained desk phones. The trend is always to peaceful coexistence where the incumbent technology, if it really does die, does so over a period of many years and even decades.
Some people say that Enterprise Software Is Dead, ERP Is Dead and so on. But they’re not, and neither are company data centers. In fact, many companies, probably a large majority, will use a blend of approaches, tapping resources inside the firewall as well as cloud services and third-party data centers, and matching each tactic to current needs.
And even for those who suggest that the on-premise model is a dinosaur, remember that the dinosaur ruled the world for quite a long time: the best part of 200 million years, give or take…
Aucun commentaire:
Enregistrer un commentaire