50 PERVASIVE computing Published by the IEEE CS n 1536-1268/09/$25.00 © 2009 IEEE
T he Web has revolutionized com-puting applications. Originally
designed to navigate static documents
using simple browsers and servers, the
Web evolved into a full-fledged appli-
cation platform backed by a myriad of
technologies. This transformation also
fundamentally changed the underlying
network. Without Web applications, we
might not have layer-7 switches, SSL,
or content distribution networks. The
Web’s need for complex, distributed
services similarly changed computer
systems, leading to the “computing
clouds” of large data centers.
The Web’s gradual evolution has
led to complex combinations of many
underlying technologies. While the
ability to flexibly stitch together data-
bases, Ajax, and other technologies has
been empowering, constituent parts
can interact in unforeseen and danger-
ous ways. SQL injection is a canonical
example: a Web service that does not
carefully filter user input is vulner-
able to users writing arbitrary queries
against back-end databases. A clean and
thought-out design would undoubtedly
have been simpler and safer.
We believe that one of the next major
application platforms will be three-
dimensional, online virtual worlds.
They provide a compelling substrate
for shared, networked environments
where people can communicate, shop,
socialize, collaborate, and learn. Appli-
cations of virtual worlds are already
gaining traction. Numerous multi-
player online games such as World of
Warcraft, Everquest, Lineage, and Eve
Online demonstrate that virtual worlds
are a lucrative and powerful platform
for entertainment. An ever-growing
list of blue-chip companies are deploy-
ing their own worlds, as evidenced by
Intel’s and IBM’s research on virtual
worlds, Sony’s Home, and Sun’s Project
Wonderland.
Unfortunately, the early evolution
of virtual worlds is as ad hoc as the
evolution of the Web. Systems have
completely independent constructions,
sharing few architectural aspects and
offering no interoperability. Users gen-
erate content with custom formats, new
world-specific programming languages
are created for programmable behav-
iors, and proprietary protocols run
each world. Systems today are closed,
limited, or do not scale. The problem
of designing open, programmable,
scalable, secure, and extensible vir-
tual worlds remains an open research
problem.
The Meru Project at Stanford Uni-
versity is designing and implementing
an architecture for the virtual worlds of
the future. The hope is that we can avoid
some of the complexities the Web has
encountered by learning how to build
applications and services before they
are subject to the short-term necessi-
ties of commercial development. While
Meru cannot compete with the content
creation of commercial virtual worlds,
it can, like the original World Wide
Web at CERN, investigate basic ques-
tions about system design. By doing so,
we can open the door to a future where
physical sensors in the real world seed
their virtual reflections, users can visu-
ally browse a sea of information, and
virtual avatars convey physical social
cues to bring distance interaction to the
level of actual presence.
Our research suggests that we can
achieve these properties by a careful
architecture of the underlying systems
and using the physical world as analogy
for communication within the virtual
world. By creating a special space zero
that reflects the physical world and by
constraining digital communication
based on real-world physics, we can
bridge virtual and physical environ-
ments at a planetary scale.
ComPonEntIzIng
a VIrtual World SyStEm
Virtual worlds today spend signifi-
cant effort to scale to large numbers of
users. Approaches range from sharding
(where different groups of users inhabit
different replicas of the entire world),
Scaling Virtual Worlds
with a Physical Metaphor
Daniel Horn, Ewen Cheslack-Postava, Tahir Azim,
Michael J. Freedman, and Philip Levis
The problem of designing
open, programmable,
scalable, secure,
and extensible virtual
worlds remains an
open research problem.
Spotlight
Editor: Brian Brannon n bbrannon@computer.org