Escolar Documentos
Profissional Documentos
Cultura Documentos
The Evolution
of Virtualization
Virtualization is moving out of the data center and making inroads
with mobile computing, security, and software delivery.
I
t’s no secret that virtualization,
a technology long associated
with mainframe computers,
has been transforming data
centers due to its ability to
consolidate hardware resources and
reduce energy costs. But in addition to
its impact on data centers, virtualiza-
tion is emerging as a viable technolo-
gy for smartphones and virtual private
networks, as well as being used to re-
conceive agile and cloud computing.
Over the past decade there has
been a great deal of work on improv-
ing the performance, enhancing the
flexibility, and increasing the manage-
ability of virtualization technologies.
Developments in the past five years
alone, for example, include the abil-
ity to move a running virtual machine, An iMac computer, with VMware Fusion, which enables it to run Windows XP Pro on the left
along with its live operating system screen, Windows Vista Home on the right, and Mac OS X Leopard in the background.
and applications, to a physical host
without major downtime. The indus- virtualization is being used; they want prior to shipping. Virtualization, he
try has also recently witnessed the access to their applications, and they says, will let handset manufactur-
ability of virtualization to log the ac- want the very rich media experiences ers test once and deploy on different
tions of a virtual machine in real time, that many modern applications offer.” handsets. For the carriers, Herrod pre-
with the purpose of being able to roll Arguably, one of the most interest- dicts that virtualization will enable a
back an entire system to an arbitrary ing and novel uses of the technology new set of services, such as allowing
point and then roll it forward for de- is on mobile devices, where virtual- users to deploy a virtual copy of their
bugging or auditing. These and other ization enables several new use-cas- mobile data to a newly purchased
recent developments have positioned es, such as isolating work and home handset. And for businesses, he says
virtualization as a core technology in smartphones on a single physical that those who want a single handset
cloud computing and have facilitated handset. Gartner predicts that more for home and work will be able to use
the technology’s move to the desktop. than 50% of new smartphones will different virtual phones. “Their work
“It’s clear that virtualization is here have a virtualization layer by the year phone could be restricted to very spe-
to stay,” says Steve Herrod, chief tech- 2012. The need for virtualization on cific applications and corporate data
nology officer at VMware. “In the fu- smartphones is strong, says Herrod, that is secure and completely isolated
ture, we’ll look back at the nonvirtual- particularly as these devices become from their home phone, where they
ized compute models as we look back more powerful, as mobile applica- may have personal information and
at the phonograph and bulky CRTs.” tions become more advanced, and as games,” he says. “The more we talk
Screensh ot by Ch ristoph er Glisso n
But Herrod also says that the industry is security becomes a bigger issue. “Just with people about this new area, the
far from realizing the full benefits that as in the early days of our x86 desktop more use-cases we find.”
virtualization can bring to desktops, virtualization efforts, we see many dif-
laptops, and smartphones. “Virtual- ferent benefits that will come with this Enhanced Security
ization is picking up steam rapidly for virtualization,” says Herrod. The notion that one of the strengths
desktop users, but it has certainly not As one example, Herrod cites the of virtualization is its ability to isolate
achieved ubiquity yet,” he says. “End substantial testing procedures that data and applications corresponds
users don’t want or need to know that every new handset must undergo to another aspect of the technology
ma r c h 2 0 0 9 | vo l. 52 | n o. 3 | c om m u n ic at ion s of t he acm 19
news
Lef t: Photog ra ph courtesy of Caroline Self ridge, Righ t P hotograph co urt esy o f Informat i on sd i e n st W i s s e n s chaft
Obituaries
In Memoriam
The world of computer science definition and selection, being handcrafted ahead of time metaheuristics, and his
recently lost two esteemed dependencies among variables, and fixed, the agents and their conviction that optimization
members: Oliver G. Selfridge, and unsupervised learning— networks of communication algorithms based on
who died at 82, and Ingo learning without explicit access could evolve with experience. metaheuristics, like evolutionary
Wegener, 57. to signals about success versus “For decades, Oliver algorithms and simulated
Selfridge, whose career failure.” communicated an exciting vision annealing, should be studied
included positions at MIT, In 1956, Selfridge, with where computers would one day with the methods from
BBN, and GTE four colleagues, organized a learn to infer human intentions the theory of
Laboratories, is conference at Dartmouth College and act to assist people without efficient
widely regarded as that led to the creation of the the need for detailed expression algorithms and
a leading pioneer field of artificial intelligence. And of problems,” says Horvitz. “Such complexity theory.
in the field of his 1958 paper, “Pandemonium: a vision has evolved to be central Wegener’s new,
artificial A Paradigm for Learning,” is a in research on human-computer theoretical
intelligence and classic AI treatise that essentially interaction.” approach
the father of machine perception. provides a blueprint for machine Ingo Wegener, a professor of produced a profound
“In prescient research in the learning research. computer science at the understanding of the limitations
1950s,” says Eric Horvitz, “The Pandemonium work Technical University of of such metaheuristics.
president of the American introduced a distributed model Dortmund, is well known for his Wegener was appointed a
Association of Artificial for pattern recognition, where groundbreaking work in member of the German Council
Intelligence, “he introduced and a community of interacting complexity theory. He wrote a of Science and Humanities,
tackled key problems that are ‘demons’ or agents with different pair of important monographs, the leading scientific advisory
now well known to machine competencies and functions The Complexity of Boolean Functions committee to the German
learning researchers, including perform different subtasks (1987) and Branching Programs government, in 2004, and
the challenges of search and that are then combined into and Binary Decision Diagrams won the Konrad-Zuse-Medal,
optimization over large final answers or behaviors,” (2000). In the early 1990s, he Germany’s most prestigious
parameter spaces, feature Horvitz notes. “Rather than worked in the formal analysis of computer science award, in 2006.