The cloud computing battleground takes shape. Will it be winner-take-all?
This year has been one of relatively grand alliances between emerging cloud computing vendors as they fill holes in their capabilities and try to create appealing one-stop enterprise cloud services.We’ve seen major announcements so far from IBM and Juniper, Cisco/EMC/VMware, and most recently BMC and Salesforce. There are many other smaller initiatives that have formed as well and all of these efforts underscore several key points for those businesses trying to understand the real strategic benefits of the cloud including cost, agility, and scalability.In the end we have some momentus choices; here's how to take the high road when it comes to enterprise cloud computing.
You can bet that the industry will be playing for keeps yet businesses can increasingly reap real bottom line benefits. This year has been one of relatively grand alliances between emerging cloud computing vendors as they fill holes in their capabilities and try to create appealing one-stop enterprise cloud services.
We've seen major announcements so far from IBM and Juniper, Cisco/EMC/VMware, and most recently BMC and Salesforce. There are many other smaller initiatives that have formed as well and all of these efforts underscore several key points for those businesses trying to understand the real strategic benefits of the cloud including cost, agility, and scalability:
First, there is no single vendor that can today provide an end-to-end cloud computing solution for businesses, hence the reason for all the alliances. The cloud computing stack (facilities, bandwidth, compute power, storage, operations, management, etc.) is deep and comprises not only most of the elements that you would find in a corporate data center but a great deal more besides. This includes R&D, product development, support capabilities, developer networks, and capabilities such as compliance monitoring and additional layers of security and governance.
Second, it's unclear how the cloud computing vendor landscape is going to shape up. Everyone is in early days yet with only Amazon with anything approaching operational maturity, with Google and Force.com vying for the lower end of the enterprise. Making long-term decisions isn't a good idea in this environment, though using cloud computing tactically does make good sense at this point, especially if you're experimenting with private cloud technology that will likely translate well to public clouds, such as Eucalyptus.
Third, and perhaps most importantly, standards for cloud computing are just emerging and only cover today an incomplete portion of the cloud computing stack. This means scenarios where you can seamlessly move your cloud computing workloads from your private cloud and public clouds of choice are fairly far off still, unless you are willing to commit to one of the alliances that will enable it with proprietary approaches. This is the core scenario that businesses are interested in as dabble with cloud technology internally today and then want to move outside to get cost and quality advantages as they get more confidence in the cloud. But it's one that is currently rife with lock-in and those that remember the platform wars of the 90s are wise to recall.
While the benefits of cloud computing are enormous in terms of reducing costs, increasing utilization, and providing scalability, there's a significant risk of lock-in. Given the early state of cloud technology, there simply aren't adequate standards to offset this.
My fellow ZDNet colleague Phil Wainewright recently pointed out what I like to call the Faustian bargain of cloud computing approaches not based on open standards. There are, however, clear benefits to be had -- particularly with the real state of IT affairs in most organizations -- as long as critical issues like transparency and public scrutiny exist hand-in-hand (hint: they don't today):
It is no surprise that the heritage of buggy, unproven and unwarrantied software that businesses and individuals have been saddled with by the established vendors over many years has led us to instinctively mistrust any computing that forces us to rely on a third party.
Yet despite our understandable caution, it is far better to trust the cloud, where security and performance are continuously open to public scrutiny, where costs can be predictably mapped to actual value delivered and where the technology is constantly kept up-to-date for no extra cost or disruption to the customer. Provided the buyer makes proper due diligence and precautions, there is in my view no alternative form of computing that is more trustworthy.
The incumbent, on-premise establishment on the other hand can overprice, under-deliver, cause massive overruns, suck out 80% of our IT budgets for routine work – but we need to keep trusting them.
So we have immature cloud computing technology, at least when it comes to enterprise-wide capabilities, incomplete standards, and the potential that market leaders will convert their existing online strengths (see Amazon and Google) into a winner-takes-all strategy, and you have a recipe for hesitancy and foot dragging on cloud computing of even the most enlightening IT thinker. Yet, most will agree that the current state of affairs in most IT organizations could be greatly improved by pay-only-for-what you use, switch-at-any-time cloud computing services that support swift and easy two-way public/private cloud migration.
The Cloud Computing High Road
The high road to get here, and by this I mean what will smart companies do that want to stay off the increasingly slippery slope of shadow IT, cloudsourcing dictated by the board, worst of all, potential lock-in to technologies and providers. They will insist on a some hard-won ground rules that we learned from the platform wars of yore. This won't be easy, just like it wasn't back then. But unlike the Web, for now cloud computing is being driven more by vendors and less by customers or even standards organizations, so far. That has to change. Here's what I believe will get things heading in a direction that will be good for the industry in the long-term and customers in particular. And yes, vendors eventually.
Widely accepted open standards. The ability to transition quickly and easily between different public clouds and your own private cloud will largely be driven by standards. Whether these will be de facto or format remains to be seen. Although it's often observed that standards promote the least-common denominator, for the foreseeable future, that will be an attractive place for most organizations.
Transparency and public scrutiny on cost, SLA, and security. Cloud providers today don't offer much scrutiny into their actual operations and technology stacks since this can reveal competitive advantage they'd otherwise prefer not to share. Insisting on identification products being used internally and access to feeds of the same management and uptime information that the provider is seeing will be the minimum required. The latter is making much more progress than the former with most public clouds providing some sort of online dashboard, even if it's still too high level today.
Complete cloud computing stacks. We are still learning what a complete cloud computing stack is and the continued announcement of major new features from many providers shows that the industry is learning what it will take to fully meet enterprise needs.
Technology agnostic clouds. While Google App Engine adding Java support to its cloud this year it a major step in the right direction, clouds-based on proprietary approaches such as Force.com's APEX language will have limited acceptability to most enterprises. True, enterprise-class cloud computing stacks will be as technology agnostic as it reasonable while still offering an on-ramp to next-generation approaches.
Seamless switching between public and private clouds. This appears to be the focus of the bigger enterprise cloud computing alliances, such as IBM/Juniper. And it's a smart bet that this is also the focal point of enterprise adoption; namely by providing a clear private cloud approach that can evolve gracefully into full, on-demand public cloud capability at the proverbial push of a button. Real choice has to built into the foundation of cloud computing for it to be a mutual win for customers and vendors.
Security, compliance support, and governance. Good security and governance are checklist items for any cloud but best-of-breed will help in a winner-takes-all scenario, especially in hybrid cloud scenarios. Compliance is the odd-man out, however, since no cloud computing solution offers this today this as a major features (though they often don't preclude it). By this I mean explicitly supporting requirements such as the EU's geographic restrictions of the storage of citizen's personal information. The addition of cloud compliance support will help provider guarantee that the use of their product won't land them in a local court.
There are undoubtedly other enablers as well, especially on the technical side such as migrating vast amounts of data quickly into cloud environments. Gilder's Law will hopefully address this though other approaches, such as bulk physical shipments of data are becoming popular today for efficiently moving tens or hundreds of terabytes of business data to public clouds.
The bottom line: The cloud computing wars are just beginning and the stakes are enormous as an estimated $350 billion to a $1 trillion dollar new business and technology market heats up. You can bet that the industry will be playing for keeps and yet businesses can increasingly reap real bottom line benefits. Provided, that is, if they keep clear heads and insist on ground rules that work for everyone.