Joe Weinman| Wired
A group of interrelated technologies is redefining how we live and work: cloud computing, big data, mobility, and the Internet of Things.
The cloud is at the epicenter of all this activity: big data migrates to the cloud to be sliced and diced; today’s tablets, smartphones, and phablets rely on the cloud for services and entertainment ranging from social networking and microblogs to streaming video; and the hyperconnected world of smart grids, biosensors and connected vehicles will rely on the cloud to collect data and then turn down thermostats, alert physicians, or avoid collisions.
As a result, cloud computing is growing exponentially. The industry is generally viewed as having three major segments: “infrastructure as a service,” where computers and storage are available on an on-demand, pay-per-use rental basis, while residing in a cloud service provider’s data center; “software as a service,” which includes applications accessible over a network such as the Internet, e.g., customer relationship management and billing but also voice and video communications and gaming; and a layer in between, referred to as “platform as a service,” which enables developers to rapidly develop and deploy new applications.
A broad variety of companies offer cloud products and services, and just to list the largest ones would use up all of the space in this article. Suffice it to say that there are hundreds of cloud service providers of various sizes, and a rich ecosystem of additional companies that provide chips, networking gear, servers and storage, power distribution equipment, colocation and interconnection facilities, software stacks, management and monitoring tools, consulting services, and so forth.
This rich set of options offers plenty of choice to customers, whether they be consumers or businesses. And, as the cloud moves from early adopters to ubiquity, customers will leverage not just one cloud provider, but many, to solve needs across various business units and functions as well as one-off strategic initiatives.
However, a plethora of providers has its downsides: cost and risk associated with the management of complexity, search costs associated with provider selection, and transaction costs associated with everything from migration to operations.
And, while there will no doubt be some consolidation and organic growth, causing large providers to become even larger, there will also likely be a long tail of mid-sized and small providers that will compete on other factors, such as location, specialized expertise, or customer intimacy. By way of analogy, Walmart, Target and Albertsons may have a substantial share of the retail market, but there is no shortage of “Mom and Pop” fruit stands. Marriott, Starwood, and Hilton may have a large portion of the global hotel market, but there are plenty of roadside inns and bed and breakfasts.
The IEEE (Institute of Electrical and Electronics Engineers) is helping to ameliorate problems associated with incompatibility across multiple providers. The IEEE is a global organization with a charter to “foster technological innovation and excellence for the benefit of humanity.” Among other things, the IEEE has developed hundreds of critical industry standards, including one that we are all familiar with, IEEE 802.11, better known as Wi-Fi. In 2012, the IEEE launched the Cloud Computing Initiative, chaired by Steve Diamond, to facilitate and coordinate cloud computing and big data activities across the breadth of the IEEE. An early focus of the Initiative was the formation of the IEEE P2302™ Intercloud standards working group, chartered to develop standards for cloud computing interoperability and federation.
In the same way that proprietary networks were made interoperable and thus easy to use by the Internet, the vision of this effort is for clouds, including proprietary clouds, to be made interoperable and easy to use via the Intercloud. The Intercloud does not supplant today’s cloud providers; in fact, it should benefit current and emerging providers as well as business and consumer customers.
Analogies from the air travel industry and the Internet can help explain this. While there are hundreds, if not thousands, of airlines and millions of websites around the world, they’ve agreed to jointly follow a number of protocols and procedures.
Airlines often need to transfer bags between carriers and the Internet needs to transfer data packets between ISPs. The Intercloud will need to reliably transfer applications and data between providers.
Airlines needed to agree on what codes to use for airports and the Internet needed a standard way of using URLs such as “http://insights.wired.com/.” The Intercloud will need mechanisms to uniquely identify cloud service providers.
Travelers must present valid identification documents, payment methods and boarding passes; the Internet and Intercloud both need mechanisms for identification, authentication, and payment.
Airlines want the ability to advertise their available seats and the prices for those seats on a variety of exchanges such as Travelocity and Expedia; websites want to be found in search results; cloud providers will want to be able to advertise available resources on Intercloud exchanges.
The similarities are apparently endless, because abstractly, any service industry — airlines, Internet, retail, hotel, rental cars, telephony, and so on — faces similar challenges, and there are clear benefits to solving these kinds of challenges collaboratively while continuing to compete in other areas. An airline can compete on price, on-time arrival, service quality, and legroom. It doesn’t need to compete based on a better scheme for naming airports.
And, in the same way that a shared mechanism for interline baggage transfer doesn’t obviate the need for airlines, nor a standard mechanism for IP addressing put any web sites out of business, the Intercloud will facilitate continued growth of the industry, while increasing customer satisfaction.
The Intercloud will benefit service providers because they will be able to more easily partner to share each other’s resources, advertise services to customers, participate in markets and exchanges, and help each other out during emergencies such as outages. It will benefit customers, because they will be able to more easily select from multiple providers, switch providers if one ceases service, offer improved user experiences thanks to a more dispersed footprint for highly interactive applications, and more easily build complex solutions from individual service components. Moreover, the Intercloud will take the onus for solving these problems away from users, helping to accelerate cloud adoption.
To help make the Intercloud standards a reality, the IEEE just announced the Intercloud Testbed initiative to help evolve and validate the P2302 standard. Emerging standards need to be tested in the real world on real physical networks, servers, storage, and software. And interoperability standards require even more testing, as various combinations of components are evaluated for things like functionality, reliability, and performance. Almost two dozen companies and academic institutions have come together as founding members of this initiative (Disclosure: my employer, Telx, an interconnection and colocation company is one of them), and the testbed and related IEEE-led efforts are open to all who wish to contribute — not only service providers, but equipment vendors, subject matter experts, and end-customers.
The idea is that testing — both abstractly and against emerging customer scenarios — will identify needed improvements in the emerging standard which can be addressed. These improvements will then be further tested. The Internet took decades to go from a vision of packet switching to where it is today. Between the IEEE, industry, and academia, one can hope that the vision of an Intercloud, conceived as early as 2009 by testbed initiative founder and chief architect David Bernstein, cloud computing initiative chair Steve Diamond, and their colleagues, is now getting the attention it deserves.
