Is the word “cloud” too big to fail?
Suddenly, overnight, cloud computing broke the barrier of if and entered into the boundless expanses of when. Suddenly everyone is discussing their cloud strategy and any semblance of a hosted offering is now a cloud offering. The very act of being connected to the internet has become an activity in the clouds. News of cloud outages (like all news really) has become sensationalized; Oh noes! The clouds is fallin’! Will we have to rethink our cloud strategy? Withdrawing in a panic caused by network doldrums? Nah, everyone knows that we can’t move off the cloud–it’s just too cheap and easy. The result of this flurry has been to galvanize the word cloud into IT discourse.
President Obama has deployed his own Cloud Vanguards, including the first ever Federal Chief Information Officer Vivek Kundra, and his appointed “Cloud Czar” Federal Cloud CTO Patrick Stingley. They’re armed with billions to spend on building cloud infrastructure and procuring cloud services, spreading the wealth of money and innovation around. Contracts for government use of cloud computing have already been awarded and large batch processing jobs are being doled to public clouds and platforms. Government websites like usa.gov have been moved up into the clouds. Moreover the NIST had done the unthinkable just a few weeks earlier during the heat of Manifestogate; they had the audacity to apply a definition of the cloud. Although they predictably try not to make anyone all pissy about it, noting that this is a “new paradigm” with “many definitions” in a “large ecosystem” of providers. This draft, like the Open Cloud Manifesto, are signs of things to come. The Obama administration’s interest has thrust the cloud into the realm of politics fostered much spirited debate over cloud legislation and international business law with efforts to make US-based cloud services more attractive to the rest of the world.
Globalization is also a catalyzing factor here. IT resources and innovation have been spread thinner around the world over the past decade. The cloud paradigm has offered a unique opportunity to level the international playing field by creating a common marketing and infrastructure ideology. The United States is currently 7th in search frequency for cloud computing on Google. Number 6 is Taiwan. Using WolframAlpha (one of my new favorite tools) we can easily use the cloud calculate that Taiwan has only 6.619% of the total number of internet users of the US (14.76 to 223 million). While it’s obvious that there is a demographic difference for internet use, it certainly illustrates the point. Public and private clouds are popping up in all geographic regions. Companies are using clouds to bring IT to corners of the world that never seen this type of infrastructure investment before. Small companies in Cambodia, Lithuania and Venezuela are leveraging public clouds to deliver services to new markets, IBM has built cloud datacenters in Brazil, South Korea, Vietnam and India, governments all over the world are rapidly building clouds to compete both economically and militarily. Emerging and growth economies have proven to be very agile in their positioning against the lumbering giants and this has largely been due to the global adoption of the word cloud to describe the metered consumption of resources over an external network.
Of course, cash is king. The Motley Fool is blasting newsletters to their millions of followers about investment opportunities in Amazon and Google, saying that their clouds are going to help take down Microsoft’s empire and cloud companies are the next big stock picks. This is pushing the term cloud computing into the common dialog and attaching to it a brand awareness. Control of the terminology is already breaking away of the sphere of IT influence. Even Oracle CEO Larry Ellison, who’s become a somewhat notorious cloud villain, always seems to hedge his bets. After his infamous cloud rant, where he flippantly asks “what the hell is cloud computing?”, Oracle releases database images for EC2 weeks later (although he doesn’t change the licensing model), and then buys Sun for billions after their big cloud announcement and IBM’s failed offer. Despite Larry’s tirade about there being “no money” in cloud computing, he sure has been throwing a lot at it. Makes me wonder what is really insane; swooping in to buy Sun certainly seems like a more aggressive and perhaps overcompensating play then just altering the wording on Oracle marketing materials.
“They’ve gone to plaid!”
The cloud as an idea, as a term, and as a method of doing business in IT, is too big to fail. The height of marketing hype and cash investment directly into so-called cloud solutions has provided what is now evolved into a stable platform and a common thread for discussion and innovation. Whether companies are ready to forge ahead is going to play a big part in the reformation of the IT world and the word cloud will ultimately define the tech giants of the early 21st century.
P.S.: Is the phrase “too big to fail” too big to fail?
Cloud Management is Cost Mitigation
In an analogy I attribute to Michael Crandell; Instances are like raw diamonds. It’s great to have quality base materials, but the real value shines when you cut and polish them.
The investment in the server resources is important, but that’s the easy part. It’s what you do with them that is going to be the most critical part of any application of infrastructure (yes, that’s a pun). Cloud Management Platform vendors are aware that we’re on our way to a multi-cloud world. Every major player has jumped into the game with some type of platform or infrastructure offering–I’m assuming my audience at this point won’t require me to list them all. The point is, it should matter to companies interested in deploying their applications or services on the cloud that the vendor they select is pushing hard for support of multiple infrastructure providers and creators. Tools for cloud management will be the biggest contested space in the race for cloud supremacy.
It’s obvious really, why would anyone think that the War for the Clouds would take place on the ground?
Cloud management software should be a big decision in the strategy and investment of any organization moving their infrastructure into the cloud. Don’t just stare at numbers on a spreadsheet and compare the cost of the cheap commodity hardware purchased and racked at deep volume discounts with state of the art server management technology. There are many sunk costs associated with developing without a multi-cloud management platform. Consider these, all of which could be classified as $$ per hour per cloud:
Learning Curve(s)
Commensurate costs should be estimated inclusive of RnD costs, trial and error, and scattered nature of educational resources, documentation and support resources for DIY integrations. They should also account for ongoing education around new cloud infrastructure developments. Just trying to keep up with the firehose of information about cloud computing is a full time job in and of itself (although most of the work is sifting through the hype).
Integration Development/Training
Engineering cycles will be spent integrating an application or service into the cloud regardless of which way it’s approached. Cloud computing is not plug and play, so you should take care to work with a company that is working fervently on creating switchboards where it can be. Isn’t that what you wanted in the first place? Then there is a matter on training staff on operating procedures around development, testing, deployment and administration. How many times can you afford to do that? Invest those cycles into something with longevity in interoperability. A cloud agnostic API is also a big plus. Eventual standardization of an API would be great and many vendors will likely adopt components of it; but ultimately an API that you pay for should provide much more then the basics.
Revisioning and Lifecycle Management
Internal engineering resources would be required to revision and support all cloud server components. In addition to your own applications. Cloud management tools should focus on more then just launching servers. They should focus on making it easier to maintain a variety of useful components and utilities and should address the fact that there is both power and confusion in numbers. Disposable servers mean lots to keep track of over a long period of time. Software should track these resources beyond their instantiation at a given moment. This is about business continuity after all.
Ongoing Cloud Development/Maintenance/Support
Resources would be needed to evaluate, develop, test, QA, and implement, new features as they are released to the public from various infrastructure providers. Do you have time to keep track of that all? Your revenue is directly generated by your application or service; not by spending time on more RnD in a space that’s rapidly changing the IT landscape by the day. Cloud platforms are competing for your dollars in this place. So you can bet that they are paying close attention to what’s going on so you don’t have to. This is the cloud mentality; companies should start acclimating.
Monitoring, Emergency Support
Commensurate costs would include development/integration of cloud monitoring systems, on-call engineering resources and incident response time. These costs would be multiplied by the number of servers you’re running. Some infrastructure providers do offer these services as an additional cost, so you’re going to pay for it either way. This is an obvious and critical component of a cloud management platform and there should be both I and AI based resources available to leverage in the case of an issue.
Remediation
Warn and Automate! Not all problems would need to be addressed manually with an engineering resource; and problems are not 9 to 5 employees. Constant systems administration is required for any infrastructure, yet much of it can and should be automated, especially in the cloud. Incidents that require manual intervention could potentially lead to additional overtime costs at less-then-social hours. Using cloud management software to set up automated remediation decreases the likelihood of a critical escalation.
Partner Technology
The cloud is the ultimate technology deliver platform for ISVs as well. Independent Software Vendors should also be leveraging a cloud platform when considering their deployment strategy. How will you deliver your technology to multiple infrastructures and reach the largest number of users with the least amount of effort? End users should also welcome the emergence of cloud appliance federation tools. Cloud platform vendors again will be competing for a piece of the cloud app-store. Apple and Facebook’s model examples will be followed.
The bottom line is that investment in cloud management software will mitigate many of these costs in the long term through software innovation. We’re still at the tip of the iceberg. Make sure you’re comitting your valuable engineering and dollar resources to something that will ultimately provide you with better agility in the clouds–especially when it comes to base infrastructure providers. When you first heard buzz about hosting your servers in the cloud you probably envisioned cloud management software…
Invest in your instinct.
-
Archives
- February 2011 (1)
- October 2009 (1)
- May 2009 (1)
- April 2009 (2)
-
Categories
-
RSS
Entries RSS
Comments RSS
