Re: The rise of the hypervisor
- From: rst0wxyz <rst0wxyz@xxxxxxxxx>
- Date: Thu, 17 Jan 2008 15:14:31 -0800 (PST)
On Jan 17, 2:38 pm, PaPaPeng <PaPaP...@xxxxxxxxx> wrote:
Sounds impressive but I don't know enough about techology to feel
excited anymore or care.
Virtualisation
The rise of the hypervisor
Jan 17th 2008http://www.economist.com/business/displaystory.cfm?story_id=10534566
PALO ALTO
From The Economist print edition
Is this the most disruptive technology in business computing since the
internet?
HISTORY may not repeat itself, but it does sometimes rhyme. In 1980
when IBM asked Microsoft, then an unknown software firm, to provide
the operating system for its personal computer (PC), it made the
mistake of allowing its supplier to license the software, called DOS,
to other hardware firms. DOS quickly became the dominant computing
platform for the PC and the basis of Microsoft's might.
Two decades later Microsoft may have made a similarly seminal mistake.
In 2002 it balked at paying a high asking price for VMware, then also
an unknown start-up. VMware was later acquired by EMC, a big
data-storage supplier, but the Silicon Valley firm remained largely
independent. This allowed it to develop software that may yet emerge
as a dominant platform in its own right-not for single computers, but
for the vast warehouses full of machines, known as data centres, where
much computing will be done in future.
VMware's claim to fame is a technology called virtualisation,
originally developed for big computers such as mainframes. It allows
computers to split themselves into several "virtual machines", each of
which can run its own operating system and applications, in effect
separating software from hardware. To do this, VMware developed a
small program called a hypervisor, which controls how access to a
computer's processors and memory is shared.
Before VMware came along, virtualisation had lingered in obscurity.
Rather than splitting up big machines, firms found it easier to use
small ones for each new application. For a while, this was a rational
strategy. Servers were cheap. Machines that ran more than one
application were more likely to crash. Yet the approach led to "server
sprawl", turning data centres into complex warrens of understretched
hardware that required ever more people, space and power to keep them
going. So it is hardly surprising that virtualisation, which allows
multiple servers to be consolidated into a single machine, is one of
the fastest growing areas in the software industry (see chart). As the
software industry matures and consolidates-this week Oracle said it
would buy BEA Systems, and Sun said it would buy MySQL-the emphasis
has shifted away from fancy new technology and towards tools that cut
costs and allow firms to do more with less.
Server consolidation is only the most obvious merit of virtualisation,
however. Once computers have essentially become bits of software,
getting new ones up and running takes minutes, not weeks. Even more
important, servers can be moved around, even when in use. This allows
for clever tricks such as concentrating virtual machines on as few
computers as possible and switching off the rest to save energy. And
particular server configurations can be packaged into downloadable and
highly reliable "virtual appliances".
It is not just servers that can be disembodied. Desktop computers are
next on the list, because virtualisation allows them to be managed
centrally. Operating systems and applications need no longer run on
PCs on desks, but can run on virtual machines in the data centre that
can be accessed remotely-theoretically from any PC in the world.
Storage is also getting more and more virtualised, so that data can be
shunted around just as easily.
The ultimate goal of virtualisation is to make a data centre, or even
several of them, look like a single pool of computing, storage and
networking resources that can be allocated as needed. Thomas Bittman
of Gartner, a market-research firm, calls this "real-time
infrastructure". Yet for the vision to become reality, something is
still missing: a set of tools to manage all these computing
resources-an operating system for the data centre, if you will.
VMware is well on its way to building just that. Founded ten years ago
by a bunch of computer scientists and run by Diane Greene, it has
become one of the fastest growing software firms, a fact that explains
its blockbuster flotation last August (EMC still holds a majority
stake). Although VMware is best known for its hypervisor, it now makes
more money from a suite of other products, such as software to manage
virtual machines. And it has been successful at persuading other
computer firms to make their wares work well with its products, thus
building an ecosystem around its emerging platform.
How was it possible for a newcomer to leave incumbents such as HP,
Microsoft and Sun Microsystems in the dust? One reason was that VMware
had the right product at the right time. More important, VMware
delivered "a non-disruptive disruptive technology", as Ms Greene puts
it. Customers can install it without having to rejig their existing
set-ups. By contrast, other virtualisation efforts were too ambitious
and required extensive changes.
Now the industry's heavyweights are fighting back with a vengeance. In
recent months they have announced big virtualisation initiatives. HP
and Sun want to bring technologies developed for their high-end
servers to the rest of the data centre. Microsoft, meanwhile, will
integrate its own hypervisor, called Viridian, into the next version
of its Windows operating system, essentially giving it away-and
raising the spectre of yet another antitrust case. (On January 14th
the European Commission opened two new investigations to see whether
Microsoft abused its desktop monopolies to restrict competition.)
Yet the most interesting competitor is another newcomer: XenSource. It
distributes its hypervisor as free, open-source software but sells
related products. Simon Crosby, the firm's technology chief, likens
this to giving away an engine in order to sell a car around it. He
believes this approach will help to spread virtualisation more quickly
and prevent VMware from becoming another Microsoft. XenSource now has
the necessary backing, having recently been acquired by Citrix,
another software firm, for $500m. Some people think that if Microsoft
fails to catch up with VMware on its own, it will buy Citrix.
Will VMware be able to withstand the collective onslaught? It is
unlikely to crumble like Netscape, the most recent start-up to vie to
become a new platform, since hypervisors are much harder to replace
than browsers. And VMware has already reacted cleverly by slashing the
price for its hypervisor and persuading hardware firms to embed it in
their machines, putting it on a more equal footing with Microsoft when
it comes to distribution.
Whatever happens to VMware, however, the virtualisation technology it
has helped to popularise is here to stay-and will transform the
economics of computing in the years to come. In particular, computing
will be much easier to outsource. Perhaps the best way to understand
virtualisation is to view it as an electronic form of globalisation:
when borders disappear, everything is up for grabs.
The idea is not new. IBM has an operating system on the mainframe
called VM which can run IBM/DOS, MVS and VM of many copies and
versions as you like at the same time. I'm surprise IBM didn't do it
on the PC.
.
- Follow-Ups:
- Re: The rise of the hypervisor
- From: PaPaPeng
- Re: The rise of the hypervisor
- References:
- The rise of the hypervisor
- From: PaPaPeng
- The rise of the hypervisor
- Prev by Date: Re: The power of calligraphy
- Next by Date: Re: The power of calligraphy
- Previous by thread: The rise of the hypervisor
- Next by thread: Re: The rise of the hypervisor
- Index(es):
Relevant Pages
|
