Is the Intercloud History Repeated?

February 2, 2009 - 5 Comments

“History doesn’t repeat itself, but it does rhyme.” – Mark Twain

Lately I’ve been seeing some interesting parallels between the nascent formation of the Intercloud and the formation of the Internet itself. Not one-to-one matches, by any means, but most definitely some of the same elements are appearing in the Cloud Computing ecosphere that once helped build the Internet. Specifically, I see three key initiatives that have an analog in the Internet’s past:

  • The rising importance of academia. Several initiatives are out there that show the increasing importance of the academic pursuit of cloud computing on the overall effort:

    • Eucalyptus – This is the leading academic effort for providing infrastructure to explore commercial cloud platform concepts. Initially focused on Amazon EC2 APIs, the underlying platform was designed to morph both at the API and at the cluster level to experiment with a variety of cloud models.
    • The UCSB/IBM Cloud Ontology – This very simple but most definitely useful entry into the taxonomy/ontology discussion gives academic communities a great starting place for categorizing cloud computing initiatives.
    • “Yahooptel” – a Yahoo, HP, and Intel initiative to make available a large scale cloud for computer science programs.
    • Google/IBM University Initiative – the Google/IBM competitor to Yahooptel.

    These intiatives are important, because similar “share the wealth” intiatives are at the heart of ARPANet and its evolution into the Internet. If these initiatives are highly successful, it is possible that more and more vendors and academic institutions will get involved. Over time, Yahooptel and Google/IBM may find themselves having to agree on standards and open pathways between the two clouds.

  • Increasing interest in interoperability among cloud vendors. Surprisingly, vendors that stand to gain somewhat from cloud lock-in are admitting that customers are hesitant to move to the cloud for just that reason. Thus, cloud vendors are pursuing interoperability options, despite the huge problems presented by differing infrastructure, impedence mismatches for workload images and even complete differences in market approaches. Several bodies are working on this, including:

  • Carrier interest in new service opportunities. The Internet represented huge business growth for telephone carriers in the early 90s, resulting in changing that designation to data network carriers. Worried about being squeezed into a “just provide the pipes” role, network service providers are eager to find new value added services that only they can provide. Hopes are strong that the Intercloud provides this opportunity.

Of course, the order in which these things are appearing, and the effect that they will ultimately contribute to the creation of an Intercloud–assuming one in fact is created–varies greatly from its ancestor. However, I recently saw a post that I thought was incredibly striking given these observations. The post was an open letter to Barak Obama’s team from newly appointed InfoWorld cloud blogger, William Hurley, calling for the creation of a public cloud infrastructure for academia and research.

Wow. Where have I heard that one before? It almost rhymes…

In an effort to keep conversations fresh, Cisco Blogs closes comments after 60 days. Please visit the Cisco Blogs hub page for the latest content.


  1. How do you configure a Cisco switch to work with Eucalyptus and is dynamic vlans?

  2. I think that you’ve hit upon an important paradigm shift – the ascendancy of connected applications and data and the mainstreaming of the network.In reality, applications and data have always been what really mattered, but the networking piece was still quite time- and resource-intensive to implement, and so a lot of the focus of technical folks was on connectivity. Now that there’s a somewhat mature (spotty, incomplete, somewhat creaky, heh), semi-ubiquitous connectivity model in place to leverage, we’re starting to see the focus shift away from the connectivity itself onto how apps/data can truly leverage that connectivity.

  3. Is cloud computing research as US centric as ARPAnet was? The list of universities and firms you give seems to suggest so, but that would be surprising. (I’m not saying you’re deliberately ignoring non-US players, just wondering who those would be).

  4. James, this reminds me of a TED talk by Kevin Kelly – Predicting the next 5,000 days of the web””. He predicts that the Internet will evolve into one giant cloud computer. When it is put this way, it is a win-win situation to work together on Cloud Computing standards, just as it was during the days of the ARPANet.”