Wednesday, November 2, 2011

Forbes: Cloud Computing is Fueling the Next Startup Boom - role of funding councils and R&E networks

[Here is a great article from Forbes on how cloud computing is fueling the next startup boom.
It is not only the private sector that is creating new opportunities in this field, but funding councils and R&E networks are also playing a critical role in the transformation of the future computing. The recent announcements by Internet 2, SURFnet, JISC/JANET, etc to broker commercial cloud services for researchers are great examples of this trend.

In this era of fiscal constraints many funding councils are looking for industry “partnerships”, rather than simple “partners” in order to accelerate commercialization of academic research. Moving research computing to commercial clouds will enable an entire new eco-system of industry and research applications and tools. A good example is Galaxy (, which is taking the genomics and bio-informatics research community by storm. Researchers are voting with their feet (and wallets) to use these commercial cloud applications because of their ease of use, simplicity, ability to scale quickly, no need for upfront capital costs for computing and no bureaucratic hassles to access university computing resources. As well, young entrepreneurial graduate students, who are fueling the startup boom are discovering that clouds are a great revenue opportunity through click compute initiatives that earn them an ongoing revenue stream every time someone uses their application on a commercial cloud.

Moving research computing to commercial clouds can easily pay for themselves in energy savings alone. Up to 80% of operating a research computer is in its energy consumption according to a recent study (Belady, C., “In the Data Center, Power and Cooling Costs More than IT Equipment it Supports). If the commercial cloud infrastructure uses clean renewable energy, then cloud computing can significantly reduce the carbon footprint of research computing as well.

Most HPC computing at universities is now built around clusters and it estimated that of 50-80% of the applications that run on these computers are loosely coupled and therefore could easily be transferred to the commercial cloud. For those who want to see rapid commercialization of academic research, the time to move to commercial clouds is now.

Some excerpts from the Forbes article—BSA]

With all the gloomy economic headlines in recent months, one can be forgiven for thinking we’re in a hopeless economic morass. But if you look beneath the surface of today’s technology shifts, you may also see potential for one of the biggest economic booms in a generation. How so?

As with all other economic booms, this boom will arise from the spunk and innovation of an emerging class of entrepreneurs, many being young and just out of (or still in) college, and others being veterans of workforce experiences relatively void of opportunities. In this next boom, another thing will be different as well – today’s entrepreneurs have an incredible resource available at their fingertips at minimal cost – cloud computing.
Unemployment is high right now, and there are many, many, many professionals who see the startup route as a more sustainable alternative to seeking full-time employment. There is now an incredible abundance of resources available on demand, for little upfront cost, to businesses.
I’m not just referring to companies that are offering cloud services – rather, companies of all types can now be created and supported. I’m talking about law firms, travel services, insurance brokerages, scientific ventures, entertainment sites and just about everything you can imagine. I’m also talking about groups or departments within existing large organizations, as well as individuals working out of home offices.
Last week, New York Times columnist Tom Friedman made an interestingobservation about the prospects for our economy going forward: cloud computing is driving new growth and opportunities. He quotes Jeff Weiner, CEO. of LinkedIn, who observed that cloud “makes it easier and cheaper than ever for anyone anywhere to be an entrepreneur and to have access to all the best infrastructure of innovation. And despite all of our challenges, it is happening here in America.” Previously, Friedman referred to this as the “DIY economy.”
There was a time when launching a serious startup required serious capital. Seed money was required for hiring talent, marketing and promotion, office space, and for technology to make it all happen. The technology portion of the equation is suddenly diminishing, dramatically. Thanks to cloud computing and social networking resources, it now costs virtually pennies to secure and get the infrastructure needed up and running to get a new venture off the ground.
One relatively recent survey of 550 startups from BestVendor found a majority use cloud-based resources: QuickBooks (71%) for accounting, Google Analytics (70%) for BI, (59%) for customer relationship management, and Dropbox (39%) for storage and backup. A few months back, there was a report about 280 North Inc., a San Francisco-based startup which produces presentation and Web development software. The founders said they incurred initial monthly expenses of about $4,500 a month by using code available free on the Web and renting storage from Amazon Web Services. As Chris Sacca, a 280 North investor and former Google Inc. executive, put it: “The biggest line item in these companies now is rent and food… A decade ago, I don’t think you could write a line of code for less than $1 million.”
Let me emphasize here that we’re talking about startups. The value proposition of the cloud is compelling for the first two or three years, then begin to catch up to on-premises IT costs. And there are small firms thatprefer to build and manage their own IT from the get-go. But it’s at that initial startup phase that cloud lowers the barrier to entry for many innovators. Acompilation of cloud computing stats by O’Reilly Media shows that companies can save up to 30% in IT costs over a three-year period employing cloud resources versus on-premises equipment. A relatively small operation with two application servers and two database servers could expect to pay about $106,000 over a three-year period, versus $149,000 for internal IT. For the first year, the capital requirements for a small server operation are near zero with cloud, versus $40,000 and up for standard on premises software and servers.
To get an in-the-trenches perspective, I asked Jason Stowe, founder and CEO of Cycle Computing, about his experiences as an entrepreneur who built his business on the cloud and offers the chance for others to do the same. Cycle delivers high-bandwidth supercomputer capabilities to scientific, engineering and technical firms — many of which are startups. “Any size organization can now tap into supercomputing power, from big companies to start-ups to individual researchers,” he says. He even coined a term for what his firm is offering: “utility supercomputing.” Essentially, thanks to cloud, Cycle can make supercomputing power available to the masses.
And lots of startups and small businesses are taking advantage of this relatively new cloud resource. Stowe gives examples: a chip design firm runs simulations of its digital circuits on his firm’s CycleCloud clusters. Researchers at a bioinformatics start-up use Cycle’s cloud to index and query genomics data to help fight disease. A young, up-and-coming scientific instrument company uses Cycle’s clusters to process the high volume of data that comes off their products.
“In these cases, start-ups can focus on their core-competency while still accessing a supercomputer that only Fortune 100s could build and operate before,” says Stowe. Many of the startups he works with would not have been able to get off the ground without cloud offerings such as that Cycle is offering. “Science-heavy start-ups would require much larger capital investments to get off the ground if they didn’t take advantage of cloud and utility supercomputing offerings,” says Stowe. “For example, 30,000-core cluster for top-five pharma would have cost $5 to $10 million and about six months to build.” With Cycle’s cloud offering, the project took eight hours to implement, at a cost of about $10,000.
Oh, and by the way, Cycle Computing itself is built on the cloud. “Cycle itself owns no servers,” Stowe says. He started the company in 2005 with no investors and $8,000 in capital — a real bootstrap story. “Thanks to the cloud we can test our utility supercomputing software at the scales our clients use, with minimal cost.”
If companies such as Cycle Computing are any indication, and with a confluence of underutilized skills and cheap online resources, we may be on the verge of an explosion in entrepreneurial activity in the decade ahead that will rival anything we’ve seen before. The barriers to entry have been significantly lowered.