Kwak, Chris; Fagin, Robert; Internet: Internet Infrastructure & Services - Introducing Internet 3 - Thick clients in the era after the web-browser 2002.

The Edge: The Undiscovered Country. In Internet 2.0, the Web era, everything has collapsed into the network center. Servers dominate the network with content, processing, and storage; PCs are largely presentation-level devices whose participation is consumption. In Internet 3.0, the center begins t...

THEMES: Kwak, Chris | Fagin, Robert | Internet
YEAR: 2002
 
Key Assumptions
The Edge: The Undiscovered Country. In Internet 2.0, the Web era, everything
has collapsed into the network center. Servers dominate the network with
content, processing, and storage; PCs are largely presentation-level devices
whose participation is consumption. In Internet 3.0, the center begins to unravel
and diffuse. The edge becomes the Internet, and devices do more with what have
to date been dormant native resources. PCs become dominant, and all clients are
thick. The domain name system is no longer the only addressing system utilized
by networks. The browser is no longer the only gateway to the Internet.

§ Faster, Stronger, More. We are ruled by laws that produce faster, stronger, and
more sophisticated technologies. While innovation drives these technologies, we
are also guided by prevailing laws that are seemingly deterministic in their
predictability and accuracy. The interplay and consequence of these laws suggest
fantastic developments to come.

§ XML: Objectifying Everything. Since 1990, the language of the Web has been
HTML, the HyperText Markup Language. We are entering a new era where
XML is quickly becoming the lingua franca not just for the Web, but the entire
Internet, across all platforms and applications. XML enables anything on the
Internet to be treated as an object. This has important ramifications for resource
utilization and end-to-end communications.

§ Virtualization. As a corollary, much of what is happening is predicated on the
virtualization of systems resources. We define virtualization as the method by
which computing and network systems resources, which may be dispersed and
fragmented, are harnessed and utilized in aggregate, to create a virtual pool of
resources, and managed seamlessly. While this kind of centrally managed system
that creates a virtual cluster is not new (mainframes have done this for years), we
believe the scope and reach of the next generation of virtualization is even
broader. StorageNetworks, for example, has been virtualizing storage, as have
private storage services providers; Zambeel, a privately-held provider of
distributed storage systems, is focused on the distributed storage opportunity.
LoudCloud has been attempting to virtualize the datacenter, as have private
companies including Mimecom and Logictier. Privately held Ejasent and
Parabon have recognized the value in virtualizing processing power. There is no
shortage of interest from public companies in virtualizing assets, and certainly no
shortage of emerging private companies continuing to challenge public
companies. Most of these companies use meshed network topologies and XML
to overcome geographic limitations.

§ Public Companies Who Understood This. More than a handful of companies
recognized the trend toward distributed and decentralized architectures. Early on,
companies like DoubleClick, Exodus, and Inktomi took advantage of the need to
distribute systems resources. More recently, Akamai and others have begun to
capitalize on distribution and decentralization. We believe even newer entrants
(e.g., StorageNetworks) are taking advantage of distributed architectures to better
utilize systems resources.

§ A Stake in the Ground. Microsoft, Intel, and Sun Microsystems have each
claimed a plot of land on the distribution and decentralization frontier —
Microsoft with .NET, Intel with the Peer-to-Peer Working Group, and Sun
Microsystems with Project Juxtapose (JXTA, pronounced jux·ta), the third leg of
its Java-Jini-JXTA triumvirate which we describe below. Each has recognized
and begun to articulate the significance of distributed and decentralized resources
to the next generation of the Internet. Each of these companies has dedicated
major resources to executing its vision. Microsoft is aiming to become the
dominant platform for the next generation of Web services; Intel is attempting to
create a set of standards for peer-to-peer computing; and Bill Joy of Sun is
drafting JXTA, a platform for distributed computing