| Embracing the Cloud Disruption  By Randy Bias, Co-founder and CTO, Cloudscaling
Dr. Donald Ferguson knows about IT complexity.
The father of IBM’s Websphere and current CTO of CA Technologies was recently quoted in a BBC interview saying, “IT departments spend 75 to 85 percent of their budgets just to keep existing IT environments running; that leaves little room for innovation.”
Ferguson explains that a primary driver of this cost profile is that modern IT systems are too complicated, over-designed and, in the end, expensive obstacles to innovation.
Cloud Computing = Simplified IT
In the past five years, public Infrastructure-as-a-Service vendors like Amazon Web Services (AWS) began offering a simpler way to provision IT services.
These webscale clouds offered a basic set of storage, compute and networking resources at low prices with credit card billing. Automated provisioning slashed deployment times. Relentless price cutting and contract-optional pricing essentially converted IT infrastructure from a fixed cost to a variable cost. A new model was born; one that launched countless Web 2.0 ventures by eliminating traditional barriers to entry posed by the high up-front costs and ongoing operational costs of legacy IT.
Of course, the webscale cloud model had its drawbacks. Limited system transparency, offsite data storage and a limited set of platform options meant that you could have your IT “in any color you want... so long as it’s black.”
Some in corporate IT saw these limitations as fatal flaws. In the past two years, however, webscale cloud vendors have added new services and platform support for a growing list of traditional enterprise applications. All the while, they’ve kept cutting prices and attracting a growing list of large enterprise clients like NASA, Eli Lilly, ING and Hitachi. In fact, there might be developers working for you right now who are using a webscale cloud for testing or skunkworks projects because they’re fast, cheap and scale quickly.
Disruption, Anyone?
What’s happening with these new entrants to IT services closely maps the classic definition of an industry in disruption. As Clayton Christensen explains in his groundbreaking research on the subject, four milestones define a market in a state of disruption:
- Disruptive innovation takes root in simple applications at the bottom of a market, where margins are thin and entrenched players are not interested in competing.
- As uncontested adoption grows, new entrants move up market, adding features that appeal to a broader audience while preserving the utility, simplicity and cost performance that made them appealing to the early adopters.
- A whole new population of consumers gains access to the product or service that was historically accessible only to consumers with more money or skill.
- Finally, new entrants become formidable competitors to entrenched players, attracting capital and customers that enable them to compete on equal or preferred footing to the established players.
Webscale cloud computing upstarts – AWS, Google, Rackspace and others – fit Christensen’s definition of disruptors with stunning accuracy.
Dawn of the Legacy Cloud
Service providers soon realized that they needed to add cloud offerings to their portfolios. Some arrived at this conclusion through a market opportunity analysis, while others simply wanted to reap IT cost savings through outsourcing and managed services strategies. To build their clouds, most called upon the same folks who had helped them build their data centers.
Understandably, these legacy IT vendors built ‘legacy clouds’ that essentially automated and virtualized the same complex IT stacks they had been selling for years. This approach offered several important advantages, most notably the ability to run existing applications with minimal reconfiguration and the ability to architect applications in a familiar way. And while some had self-service portals to automate provisioning, they unfortunately followed very few of the new design patterns being used by the webscale disruptors.
The gap between legacy clouds and webscale clouds is profound. Webscale clouds employ a fundamentally unique architecture that offers a 3-7x TCO advantage. Disruptors are using this to drive relentless price competition and gain share. Legacy clouds, by contrast, simply virtualize and automate the existing, expensive architecture. As a result, many carriers have seen minimal or even negative ROI on their legacy cloud deployments. The high TCO of their legacy clouds is also forcing them to price their services far above the webscale competition, leading to significantly lower growth rates (<30 percent) than the compound annual growth rates of between 90 and 100 percent that the new disruptors like AWS are experiencing. The difference in these approaches is massive from a revenue perspective. In 2010, Amazon is estimated to have generated approximately $500 million in revenue for their cloud service, with some analysts predicting sales of $1 billion in 2011.
A Closer Look: How is Webscale Different?
Carriers and service providers can avoid disruption by embracing the reality that there are really two different kinds of cloud platforms, one built on a legacy architecture and the other on a webscale architecture. It’s akin to the transition from mainframe to client/server that began more than 20 years ago:
- Legacy cloud architecture – Ideal for legacy apps, with automation and virtualization to optimize workloads and automate provisioning. It’s more expensive than the webscale architecture and has lower commercial growth rates when offered as a service.
- Webscale cloud architecture – Ideal for new (greenfield) apps and a growing list of enterprise apps. It’s less expensive than the legacy architecture and has higher growth rates.
Carriers and service providers that have already built legacy clouds can continue using them to run the workloads they are uniquely suited to run. However, the disruption underway in IT suggests that carriers also should add a webscale cloud strategy to their product portfolios, offering this low-cost, highly adaptive and flexible offering to their internal and external customers.
On the other hand, carriers who have not yet launched their cloud initiatives have a unique opportunity. They can skip the legacy cloud step altogether and move directly to the webscale cloud model, with the assumption that with the disruption well underway, those 'simple, inadequate' clouds will rapidly move up the value chain and soon be significant competitors to traditional IT type systems.
KT, one of Korea’s dominant players in both fixed line and wireless, took the second course. Its ucloud offering is built on a webscale architecture, which the company employs across public and private cloud architectures, as well as a hybrid model it calls a ‘mobile cloud.’ The company’s leadership has been very vocal about the value they see in embracing the same cloud model that’s worked for the most successful public cloud players.
KT’s results are tantalizing: Its private cloud yielded a 79 percent CapEx reduction and 50x decrease in provisioning time compared to the legacy model, and its public cloud beat AWS pricing by 18 percent in KT’s core geography.
Summary: Moving to the Webscale Cloud
Legacy clouds and webscale clouds are fundamentally different animals. The advantages that legacy clouds currently enjoy are eroding as webscale clouds continue to disrupt and move upmarket. The economics and agility of webscale clouds have made them a disruptive force in a market seeking simpler solutions to replace the complexities of legacy IT.
If you’ve built a legacy cloud, keep running it to serve your complex IT workloads. If not, consider skipping the legacy model and going straight to webscale cloud. As KT’s experience has proven, the advantages could move you ahead in both cost structure and competitive positioning.
Randy Bias will be presenting on the webscale cloud for carriers along with KT at Management World 2011 in Dublin on Wednesday, May 25 from 11:15 am - 12:45 pm.
| |