We're very early in the private cloud business -- which is precisely why such large and influential vendors as Oracle, Intel, HP, VMware, Citrix and Red Hat are jumping into the market with initiatives and pledges for standards and support. We're seeing some whoppers here at Oracle OpenWorld, from Oracle, Intel and HP in particular.
Why? The early birds that can establish de facto standards on data portability and resources governance -- minding the boundaries between the private and public clouds and their digital condensates -- will be in a position to define the next abstraction of meta operating system (for lack of a better term).
In just the last two weeks, VMware, Citrix and now Oracle have pledged to come to market with the infrastructure that enterprises and service providers alike will want. The cloud wanters need cloud makers, the picks and shovels, to build out on the vision of next-generation data center fabrics -- of dynamic resource pools of infrastructure, platform, data applications and management services.
How these services are supported, and how they are managed to inter-relate with each other and the services-abstracted older IT assets, forms the new uber platform -- the new target through which to attract developers, architects, partners and users -- lots and lots of users all feeding off of huge clouds of dynamic, low-cost services.
Yes, a market critical-mass cloud platform standard implementation could create yet a new way to lock in huge multi-billion-dollar markets to ... need. To need, and to want, and to buy, and to have a heck of a hard time stopping that needing. The picks and shovels. The lock-in, the black hole-pull of the infrastructure, hard to resist, and then ... impossible.
Such a prize! And just like in the past, the crass business interests side of the vendors will want to own, dominate and lock-in to their proprietary platform implementations. Opposing forces, also inside the same vendors, will opine on the need (correctly) for openness and standards to provide the real value the users and ecology players demand. The new lock-in, they will say (correctly) is not technical but in terms of convenience, simplicity, power, and cost. Seduce them, don't force them, might be the mantra.
So seduce or lock-in, early-days private cloud platform definitions require the best management of two sets of boundaries -- one that properly falls between the pubic-facing clouds, and the nascent "private" or on-premises or enterprise clouds. The pay-off comes not just from operating efficiencies but on how well the services generated from either types of cloud can interoperate and play well in supporting extended enterprise and B2C processes.
This need to cross boundaries well will also prompt the handful of public cloud providers (Amazon, Google, Yahoo, Microsoft, Apple, etc.) to embrace sufficient levels of standards-based interoperability. Think of it as mass markets balancing interests ... like globalization ... where economics more than proprietary technologies wins the day.
The second boundary to to be defined properly is between the legacy systems, SOAs, business applications and middleware -- and the private cloud fabrics that will increasingly be where new applications/services are "natively" deployed, and where the integrations to the old stuff occurs. We can really have two kinds of clouds -- one for IT and one for consumers. There needs to be one cloud that suits all of the digital universe, within certain (as yet undefined) parameters. They really need to bet this boundary right so that B2E and B2B is also B2C.
Clouds will, of course, be highly virtualized, and so they will be able to support many of the older proprietary and standard-based IT systems and development environments. But why virtualize the new stuff, too? Why have B2E/B2B old and separately B2B/B2C new? We should want one cloud approach that newer apps and services can target directly, and then virtualize al the older stuff.
The question then is what constitutes the new "native" platform that is of, for, and by the standard cloud. If there is a fairly well-defined, standards-based approach to cloud computing that manages all these boundaries -- between public and private, between the old and the new of IT -- and which can serve as the target for all the new apps, services, data abstractions, modeling tools, workflow/policy/governance/ESBs and development needs -- well that's a business worth shooting for.
Who cares how the lock-in occurs, this is the next $100 billion company business. In other words, getting this right is a very big deal. The time is nigh for defining IT for at least a decade, maybe longer.
But like the Unix wars of old (and the app server wars of not-so-old) there will be jockeying for cloud implementation supremacy, brinkmanship over whose this or that is better, and the high-stakes race for who gets the definitions of the boundaries correct best for the users, developers, channel, and partners. Who can woo the best?
What is different this time, in cloud time, is that there are few players that can play this game, less of a channel to be concerned about, and fewer developer communities to woo. Far more than in the past, developers can use the tools and frameworks of their choice, and the clouds will support them. Users also have new choices -- not between a Mac and a PC, between Unix and x86, between Java and .Net, between Linux and Windows -- but between cloud ecologies of vast services providers. The better the bundle of services (and therefore interop and cooperation), the better the customer attraction and loyalty. The seduction, the lock in, comes from packaging and execution on the services delivery.
More important than in past vendor sporting events, the business model rules. The cloud model that wins is the "preferred cloud model" that gives IT shops in enterprises high performance at manageable complexity and dramatically lower total costs. That same "preferred" cloud attracts the platform as a service developer crowd, allows mashups galore, allows for pay-as-you-use subscription fees. Viral adoption on a global scale. Oh, and the winning cloud also best plays out the subsidy from online advertising in all its forms and permutations.
Yes, we can expect several fruitful years of jockeying among the major vendors, the rain makers for the cloud providers -- and see gathering clouds of alliances among some, and against others. We're only seeing the very beginning of the next chapter of IT in the last few weeks of IT vendor news.
The cloud wars, however, won't be won on technical merits alone, it will be a real beauty pageant too. It will be more of a seduction and an election, less of a slight of hand and leveraging of incumbency ... and that will be a big switch.