Cloud computing is barely old enough to separate out a few pioneers who have made outstanding contributions to it. InformationWeek made its first attempt to do so Nov. 26, 2012, pulling 10 names off a long list of potential cloud leaders and explaining why they made the select group. Now we're calling out another group of people whose contributions have frequently matched those of the first 10, but who have tended to labor a step back from the limelight or sometimes even in the shadows.
They include an open-source programmer, an industry consultant, entrepreneurs, authors, public speakers and software strategists. For the most part they are individuals, but we wanted to start off by looking at a team of pioneers: the Rackspace OpenStack team.
No single individual at Rackspace is responsible for the company deciding to launch the OpenStack open source project -- several individuals played important roles.
President of Rackspace Cloud Lew Moorman points out that even when Rackspace was offering hosted servers for lease, they were Linux servers tied into other open-source code in the Rackspace environment. As Rackspace Cloud was launched, it, too, was heavily tied to open source, offering virtual servers based on Xen. Noting the connection, a team of MySQL developers, members of the Drizzle Project, came to talk to Rackspace sometime after Sun acquired MySQL AB in late 2008. Drizzle was a MySQL fork and the last thing Sun was interested in was forks. Drizzle team members were wondering if Rackspace might offer steady sponsorship.
A Rackspace engineering team met with the Drizzle developers in 2009. Rackspace representatives included Jason Seats, lead engineer after coming to the company with the SliceHost acquisition in late 2008. The two teams met for a day. Adding a MySQL fork wasn't a fit with what Rackspace was trying to do, but the visitors had impressed upon Rackspace how successful open-source code could be as a method of establishing a new software standard. Jason Seats, lead engineer for Rackspace at the time, turned to Moorman and Jim Curry, director of business development at the time (now senior VP and general manager of Rackspace Private Cloud), after the Drizzle team had left and said, "We should make everything we do open source."
Rackspace wanted to differentiate itself on service for cloud users, not the software it was creating to run the cloud. Moorman and Curry agreed it was the right way to go. "Jason was the first one to state explicitly that what we were doing should be made open source," recalls Moorman, but several co-workers immediately endorsed and enlarged the idea. Curry took on the task of organizing an open-source project with an active community; Seats and Moorman went to the board of directors to say they were giving away the company's latest software "and the board bought it."
"We can make money by deploying OpenStack and helping others deploy it," Moorman said.
Although Moorman doesn't put it this way, OpenStack was also needed by Rackspace. Amazon Web Services was off to a big lead in attracting cloud users and a company like Rackspace needed to organize a larger movement to compete. The idea that OpenStack was a good idea as a standard for cloud computing and good for Rackspace to establish filtered down into the ranks of both the business and engineering sides of the company. Plans were laid to launch the open-source project. Rackspace's first contribution was its Swift cloud storage system. A few weeks before the announcement at the first OpenStack Design Summit, Rackspace learned NASA was also planning to make its Nova cloud compute software open source. The two moves were complementary and both teamed up to get the announcement of OpenStack off the ground in July 2010.
Within a few months, Seats moved on to TechStars in Boulder, Colo. But other San Antonio employees were brought into the effort. Jonathan Bryce, organizer of the original Mosso cloud at Rackspace, joined and eventually became executive director of the Rackspace Foundation. CTO John Engates became an effective OpenStack spokesman at various events.
OpenStack had a slow start as various potential backers evaluated Rackspace's central role and the possibility it might one day be a competitor. The Rackspace team realized open source thrives best when it is not the stepchild of a single company. Rackspace needed the status of having played a leadership role in organizing OpenStack, but it mainly needed OpenStack to succeed. It worked to enlarge the community and share responsibility as NASA, under server budget constraints, faded from the project. With a minimum of friction, it helped transfer governing authority from its own offices to the OpenStack Foundation with a board of contributing and elected directors.
To Curry, the latest sign of what their work accomplished was IBM's adoption of OpenStack as its standard for implementing cloud software, adding its name to an already long list of backers. "It's exactly what we wanted to happen," he said in an interview. The move will have far-reaching consequences, he predicted. OpenStack will become an industry standard, Rackspace will be the first and a leading supplier of OpenStack infrastructure as a service (IaaS), and many other providers and enterprises will follow in these footsteps.
Now read on to meet more cloud pioneers.
Reuven Cohen, senior VP at Virtustream, is the co-founder of Enomaly, an open-source consulting company founded in 2004 in Toronto. Enomaly came up with one of the first self-provisioning systems, Elastic Computing Platform, for infrastructure as a service in 2005. In 2011, Enomaly developed one of the first commodity-style cloud-computing markets, SpotCloud, where owners could sell their unused cloud capacity. It was a concept that would eventually find its way into Amazon's EC2 Reserved Instances Marketplace.
Cohen was an early, tireless advocate for the cloud, appearing on panels and as a speaker at industry events. He also performed a peer review of the drafts of the NIST definition of cloud computing. He was the founder of CloudCamp, a cloud computing conference that's taken place in 300 cities and is based on an informal set of procedures that holds elections to set the agenda and elect discussion leaders from among the participants -- after the conference has started. CloudCamp became a great way for people interested in the cloud to get to know one another in the space of a few minutes. Reuven writes the Digital Provocateur column for Forbes Magazine, has co-hosted DigitalNibbles, an Intel-sponsored podcast, and has served as a strategic advisor to Canada's Department of Foreign Affairs & International Trade. He was a North American representative on the Sino-European/America Cooperation Roundtable forum held in Beijing in May 2011.
He describes himself as "an instigator, part-time provocateur, bootstrapper, amateur cloud lexicographer, and purveyor of random thoughts, 140 characters at a time."
In a 2009 issue of the MIT Technology Review, James Urquhart was named one of the 10 most influential thinkers in cloud computing. That was back when the term was still being disparaged by Oracle CEO Larry Ellison and (at the time) HP CEO Mark Hurd. He was also named one of the top three cloud bloggers by The Next Web in 2011. From 2006 to 2008, he was a senior sales engineer positioning the Cassatt data center platform to deliver automated application deployment and operations, prior to its acquisition by CA Technologies. From 2008 to 2011 he was market strategist and technology evangelist for cloud at Cisco Systems. During this period, Urquhart spoke on the role of Linux and open-source code in cloud at the 2010 Linux Collaboration Summit, one of around 60 speaking engagements that year.
He is the former writer of The Wisdom of Clouds blog on CNet, which ran from 2008 to 2011. In his postings and appearances, he made concise, logical arguments for why cloud computing, mistaken as a new technology, was really a new operations model. He dissed the frequently used analogy between cloud and electric utilities, saying using data intelligently and distributing electricity are two different things. He did a five-part series on how cloud and virtualization were forcing a rethinking of the enterprise IT software stack.
He is currently VP of product strategy at enStratus, a cloud workload management firm where he works with another cloud advocate, Bernard Golden. His latest job has cut down his frequent commentaries, but he is still a contributor to the technology news site GigaOm/Cloud, including a Jan. 27 post on DevOps and fragility versus stability in IT operations.
Urquhart is described as having "a profound interest in the broad landscape of the software industry" along with "a great combination of interpersonal and technical skills" by Christopher Kriese, an independent, San Francisco Bay-area senior software engineer with whom he has worked.
Lydia Leong came into the cloud era with her own hands-on background in technology operation and evolution. She was director of server engineering at [email protected] and director of product engineering and operations at Digex. She joined Gartner in 2000 and gravitated into becoming its first and still primary analyst focused on IaaS. In 2010, in the midst of a blossoming of enterprise interest in the cloud, her prominence as a spokeswoman on the cloud prompted Gartner to name her analyst of the year.
The authority of her voice and writing, along with the pivotal position she occupied in a top-drawer, enterprise analyst group, made her instrumental in establishing the viability of cloud computing for the enterprise. She is still engaged in calling it like it is in cloud software and services. In a Dec. 5, 2012, Gartner blog post, "Cloud SLAs Can Be Meaningless," Leong said: "Unfortunately, infrastructure as a service SLAs can readily be structured to make it unlikely that you'll ever see a penny of money back -- greatly reducing the provider's financial risks in the event of an outage." Amazon Web Services, she noted, has voluntarily paid customers for lost computing time, because its SLA didn't cover the nature of the outage. For example, Amazon's Easter outage in 2011 froze up its Elastic Block Storage service, but EBS was not covered by the SLA. The customers' virtual servers continued running, thus meeting the terms of the SLA, but they couldn't do anything without access to their frequently used data. In the case of both AWS and HP clouds, the downtime clock doesn't start ticking until the service has been out five or six minutes, respectively, she pointed out.
Leong is currently a VP of research at Gartner. Jason Hoffman, CTO of Joyent, says: "Every time I have the chance to talk with Lydia, I either learn something new or I gain a deeper understanding of my own thoughts." Hoffman is not known for suffering analysts kindly; with most, he does most of the talking, he says.
Will VMware Challenge Amazon Head On?
David Linthicum had such a strong technology reputation before the cloud era that he isn't associated exclusively with it, as are Werner Vogels, Google's data center architect Urs Hozel or Enomaly's Reuven Cohen. Linthicum was an early explainer of the client/server era, the possibilities of integrating Windows and Linux environments (a common practice today), and the mind-numbing complexity of enterprise application integration. Six of his 13 books mark major phases of computing architecture, including those on client/server, e-business, Web services and cloud computing.
In October 2009, his publisher brought out Cloud Computing and SOA Convergence In Your Enterprise, A Step by Step Guide. It played to mixed reviews because people who wanted to learn about the cloud weren't necessarily approaching it from the perspective of managing data and services in a services-oriented architecture environment. But many IT managers were, and it became a guidebook for enterprise architects who had concluded the cloud was real but didn't know how to approach it.
His role has been more than technical. He has been a constant interpreter of how the role of IT will change in the cloud era and how disruptive it will be to traditional ways of doing things.
With his stone-cold realism, his implementer's practicality and his intellectual unwillingness to rationalize away drawbacks or complexities, Linthicum played a key role in converting doubters to the ranks of cloud computing. His is not the most impassioned voice on cloud; he is, nevertheless, one of the most persuasive to entrenched IT managers, skeptics and non-believers.
He is currently founder and CTO of Blue Mountain Labs, a cloud-computing consultancy, and writes a regular column on cloud computing for InfoWorld and GigaOm Pro.
Jeremy Geelan, senior VP of Sys-Con Media and Events and conference chair of the Cloud Computing Conference and Expo, says Linthicum's 2009 book "does that rarest of things: he manages to combine showing why SOA [service-oriented architecture] and cloud computing complement one another, with a lucid game plan of how a business can take advantage of the synergies."
Will VMware Challenge Amazon Head On?
Sebastian Stadil, founder and CEO of Scalr, maker of cloud management software, is also the founder of the technology forum perhaps most conversant in cloud computing, the 8,000-member Silicon Valley Cloud Computing Group. As such, Stadil has been a leading organizer and presence in the group, creating a gathering place for cloud computing enthusiasts of many stripes and organizing a series of presentations on where cloud computing is headed. Participants say it's a place where you can air your opinions and sometimes gain greater assurance in them in the feedback from your peers.
An example of what might be a future Cloud Computing Group discussion is contained in Stadil's post on his team's benchmarking of the Google Compute Engine, as it came out in June 2012, compared to Scalr's regular platform, Amazon EC2. The March 15 post illustrated that it took Google Compute Engine about 30 seconds to get a virtual server booted; AWS EC2 took 150 to 300 seconds. "We don't know what sort of sorcery Google does here, but they clearly demonstrate engineering prowess," wrote Stadil. The post drew 26 well-informed comments in the first few hours of its existence.
Stadil started as a cloud developer in 2004, developing Web services for e-commerce and then for cloud server resources. He founded Scalr to produce management products for IaaS, whether private or public.
Will VMware Challenge Amazon Head On?
Michael Crandell is the co-founder and CEO of Rightscale. He is a thinker and entrepreneur, equally comfortable in a programmer's t-shirt, which he frequently wears while engaging customers at RightScale's user group meetings, or wearing the Silicon Valley's executive uniform of a dark blue suit and no tie. He has defined a new market -- online cloud management -- and is determined to do it better than any single cloud vendor.
Crandell created in 2006 what is now a powerhouse service that sits at the crossroads of cloud computing, with users paying a fee to use the RightScale hub to gain the flexibility and ability to move between dissimilar clouds. The RightScale system knows enough about the machine images needed by each provider that it can move a workload that is first run in Amazon Web Services out to a VMware partner cloud, then to Rackspace and back to Amazon, if the customer chooses. Rightscale simplified migration into the cloud with its Server Template selections, with users needing to bother with few of the details.
Crandell from an early stage has been a frequent speaker at cloud events, showing a strong sense of humor. He's been outspoken on the need for hybrid cloud when others were reluctant to venture forth, and he's been an advocate of maintaining mobility between clouds as a way to avoid lock-in. For the value it's brought to cloud computing, Rightscale was named by the World Economic Forum a 2013 technology pioneer "committed to improving the state of the world."
Crandell maintained a cutting edge team to build the company that includes co-founders CTO Thorsten Von Eicken, a former professor of computer science from Cornell University; and VP of engineering Rafael Saavedra. The RightScale platform, among other things, serves as the management interface for online-gaming company Zynga, which has 75 million users at peaks connecting to their activities through Rightscale.
In February, Rightscale became the first management platform other than Google to support and manage the Google Compute Engine.
Sean O'Toole, the founder of ForeclosureRadar, said Crandell and RightScale helped his company build a scalable Web platform in just a few weeks "despite no prior experience with Amazon's EC2. We could not have done it without them."
Will VMware Challenge Amazon Head On?
John Keagy co-founded ServePath with David Hecht. ServPath in 2001, in the first days of cloud computing, became GoGrid, one of the earliest IaaS providers. That was a long time ago, but in many ways GoGrid has matched the basic steps of Amazon Web Services in generating self-provisioning IaaS.
Co-founder Keagy became CEO, Hecht became chief marketing officer. Hecht left to market CloudKick, which graphed a customer's network bandwidth in the cloud and other usage metrics. It was acquired by Rackspace in December 2010. Hecht currently markets the product line of CloudAmp, bringing Google analytics to Salesforce.com application users.
Having previously founded and run InReach Internet, one of the leading ISPs in the Western U.S., Keagy sold it in 1999 and two years later founded ServePath, which became a leader in the dedicated server hosting market, a predecessor to the virtualized, self-provisioning cloud-hosting service.
Keagy has founded 13 companies that have achieved breakeven status; seven have become profitable or were sold. In 2008, when ServePath became GoGrid, an IaaS provider, Amazon was still offering its IaaS as a beta service.
After 10 years at GoGrid, Keagy felt the company was beyond the startup phase with its own CIO, CFO and CMO. So in June 2011 he left it in the hands of the board's executive chairman, Warren Heffelfinger, to work more with Upstream Networks, where he was president. Keagy remained a member of the GoGrid board. Eight months later, he was called back to the GoGrid CEO's seat.
"Sitting on the sidelines when I have the opportunity to move the needle, when I have so much more leverage on how I invest my career, I've got to take advantage of that," he told The West Host Industry Review. "There are a few things I can be doing that are transcendental and that are exciting. But when you're talking about changing the future of computing? Those are big words and the best way is taking the helm as CEO at a pioneer cloud infrastructure."
Gartner has previously placed GoGrid in the "visionary" segment of its cloud industry quadrant; this year it labeled GoGrid a "challenger," along with Joyent and close VMware partner Bluelock.
Richard Donaldson, director of global managed services at eBay, calls Kreagy "a tireless entrepreneur with great vision, excellent motivational skills and an ability to create an atmosphere of success … which is critical to any ongoing business. Bright, insightful and fun to work with."
Chandra Krintz labors in the same academic setting as her better-known fellow faculty member, Rich Wolski, at the University of Callifornia at Santa Barbara. He is founder of Eucalyptus and cited in InformationWeek's original 10 Cloud Computing Pioneers slideshow.
Wolski focused on duplicating the Amazon Web Services API set; Krintz has focused on duplicating as open-source code the APIs used by Google AppEngine. She is the founder of the AppScale open-source project, which has accomplished that goal. She is a computer language specialist who has won several awards, including IBM's Open Collaboration Research Award, a University of California at Santa Barbara distinguished teaching award, and an innovation award in the use the of X10 research language as an alternative to Python and Java.
Krintz has pioneered use of X10 as a language for producing applications doing parallel processing or distributed processing. She used it as a language for implementing MapReduce in the AppScale project and as the implementation language for parallel and concurrent activities in AppScale itself. (Google's preferred language for App Engine and MapReduce is Python.) She has helped develop AppScale APIs to work with Ruby on Rails and other popular dynamic languages. She initially had no plans to commercialize AppScale but now serves as CTO of AppScale Systems, a company founded in 2009 to productize and support AppScale software.
She continues to explore how to program load-balancing and parallel task execution techniques into applications running on large clusters or in cloud settings. She is also committed to developing applications that can run in both Google App Engine and Microsoft Azure or other, multiple cloud settings. She is a professor of computer science at the UC-Santa Barbara.
She described herself, in a 2010 interview, as "a researcher at heart, and I love computer science for the scientist."