Guest post: This is a post by Barry X Lynn, CEO of 3Tera, in response to a discussion about whether cloud computing is industrial strength. Lynn can be found on the 3Tera blog.
There have been multiple white papers and articles written on the topic - Is Cloud Computing Ready for the Enterprise? The question is asked so many times now - Is Cloud Computing ready for the enterprise? But there is an equally relevant question: Is the enterprise ready for Cloud Computing?
The following analogy may appear to imply that people who operate corporate data centers are crazy or stupid or both. But nothing could be farther from the truth. Enterprise CIOs and IT managers have the most difficult jobs in the corporate universe. They are the brains and the central nervous systems of large enterprises. They are also the most taken for granted of all executives. They represent cost centers and get no credit for their corporations’ profits, while keeping the corporation alive. If they achieve 99.99% availability of their services, an iota of kudos is given for that 99.99%. But a mountain of wrath is doled out for the other 0.01%.
If you woke up in the morning and read in the Wall Street Journal that an eCommerce company like Overstock.com had stopped using the USPS, UPS, FedEx, DHL, etc. to deliver their goods and, instead, leased airport hubs all over the world, bought a fleet of jets and bought thousands of trucks and started delivering the stuff themselves, you’d think they were out of their minds. So, why is it not equally insane for financial services companies, health care institutions, manufacturing companies, bio-tech companies, pharmaceutical giants, etc. to be spending a billion dollars or much more every year on information technology infrastructure?
Of course, CIOs and IT managers are intellectually ready for Cloud Computing. And of course they have the experience and skills to adopt Cloud Computing. And they have the resources. Most significantly, they have always risen to the occasion when disruptive technologies have been thrust upon them.
But, practically speaking, though the future holds a world where all enterprises will get computing on demand and only pay for what they consume, we know that this will not happen over night. It most assuredly though will be very much the norm in the not so distant future, and corporations owning data centers will be the exception to the rule.
So, here’s a kind of Darwinian Theory of the Corporate Data Center’s Evolution.
Cloud Computing is the most disruptive technology that has come along in a very long time. Respected technology analysts say it will be bigger than e-Business and it’s potentially a quarter of a trillion dollar market. So, is Cloud Computing a revolution or an evolution?
The answer is a resounding “Both”.
All evolutionary changes start with a revolutionary change. In Darwin’s “Origin of the Species” evolutionary changes start with a mutation. Those mutations are the revolutions that result in evolution. In most cases the mutation comes about as a mechanism to heighten the chance of survival - you know, to make the species more fit. Subsequent to those revolutions, the evolutionary process gradually occurs as the most fit survive and the mutation becomes the norm - the standard… until the next mutation starts.
Cloud computing is the mutation - the revolution. Enterprise IT and Corporate CIOs/IT Managers will jump on the opportunity to evolve as they always have when revolutionary technology mutations have occurred.
So, here’s an example of a scenario of how this evolution will happen.
In the very near future two things will occur in the enterprise data center.
First, today, the hardest things to plan for with regard to capacity, performance, etc., are on line applications offered on the web. Enterprises have no control over who may log on, how many may log on, when they may log on, what they may do once they log on, etc. So, the natural evolutionary step to mitigate these uncertainties is to run those applications on massively scalable infrastructure that scales up and down dynamically as needed, using resources on demand, always there when needed and only paying for what is consumed. These infrastructures are what we are now calling Clouds.
For the time being, the mission critical data and systems of records that are the enterprises’ life blood residing in their data centers need to be isolated from these on line applications exposed to every internet user. This will be accomplished through the use of secure virtual gateways in the Cloud, connecting, in a loosely coupled manner, rather than a fully integrated manner, to the enterprise data centers, their databases and systems of record.
These gateways will take many forms. They may be SOA gateways using XML and virtual XML firewalls, virtual messaging systems such as MQ, virtual EAI appliances or customized appliances encapsulating organizations’ proprietary techniques for reliably and securely communicating among systems (and anything new that comes along to supplement or replace these things).
Second, infrastructure/architecture agnostic Cloud platforms (what we at 3tera call Cloud Computing Without Compromise) will be installed in enterprise data centers. There will be two factors that will drive this.
(1) As more and more applications are offered on line, those same applications will often be used internally by the enterprise employees. Why incur the cost of having separate experiences for employees and customers who are accessing the same information and functionality? Also, when connecting the on line applications in the Cloud to the data center and SORs, having them on similar platforms will make it seamless and efficient.
(2) A Cloud infrastructure done right, behind the corporate firewall, enables the enterprise to run their data centers as metered utilities. It enables them to more efficiently use their hardware resources by provisioning what is needed for each application on demand and releasing those resources when no longer needed for other applications to use. It enables them to more efficiently use intellectual capital by shifting IT administrators from managing machines to managing applications. And, most importantly, it greatly decreases time to market because the lengthy provisioning, configuring, etc., of hardware and infrastructure resources is, pardon the pun, virtually eliminated. So albeit humongously significant, forget all the talk about cost reduction and avoidance. Cloud Computing in the enterprise has the potential to greatly increase revenue and beat the heck out of competitors implementing like products using traditional data center deployment methods.
OK - so what’s the next step in the evolution of the data center?
While enterprises are growing comfortable with applications in Clouds and realizing the upside of dynamic provisioning and scaling, they will be developing new applications and replacing/changing existing ones. They will start building the new applications in Clouds and as they change existing applications, will consider migrating them to the Cloud in the process. This will afford them the advantages of much faster time to market, the ability to run applications on demand in multiple data centers (globally if appropriate) creating their first truly complete disaster recovery abilities and concentrate on their core businesses which may be financial services, health care, manufacturing, etc., but certainly is not data center operations (they will leave that to the companies whose core business IS data center operations).
As a result, enterprises will find themselves with data centers that only contain data. Finally, a data center will be what its name implies. All of their functionality - all the non-data tiers of their services, will be in Clouds connected to the data centers’ data. At that point, evolution will treat the data center as an appendage. Over time, the corporate data will move to the Cloud just as many smaller businesses without data centers are using storage services in Clouds today. The corporate data center will be a vestige, and eventually evolution will cause it to disappear.
In short, the corporate data center is not a stupid useless entity. There have been no alternatives. They have been profound necessities. But necessity truly is the mother of invention, and the corporate data center, with all of its overhead, has bred Cloud Computing.
And for those who prefer an analogy to Creationism rather than evolution, all you have to do is depict the corporate data center as the eventual dinosaur with a saddle on its back being ridden by a member of the Cloud Computing species.