Special Features:
THE GRID: DEFINING THE FUTURE OF THE INTERNET (Part 3 of 3)
By Wolfgang Gentzsch, MCNC Grid Computing & Networking Services
Benefits Of Grid Technology
In the past, different departments have developed different market-specific
solutions, each within their own home-grown departmental HPC environments.
From an enterprise perspective, this IT infrastructure is very inefficient. An
enterprise Grid, on the other hand, offers economies of scale, access to one
common HPC service for all departments, reliability and quality of service,
reduced hardware and software costs, reduced operational cost, and increased
productivity.
In most cases, an enterprise Grid can be built out in two phases. The first
phase is to optimize the resources that already exist within the departments.
This phase may take a couple of weeks, and does not require any additional
hardware. The next phase optimizes the overall enterprise environment through
central management by adding central services, based on the concept of server
consolidation.
Recent articles about Grid computing benefits have primarily focused on better
utilization of under-utilized computing resources. This alone often provides
dramatic cost savings, but this is not the only benefit a Grid can provide.
Other important benefits of an enterprise or a research Grid are:
- Access: Seamless, transparent, remote, secure, wireless access to
computing, data, experiments, instruments, sensors, etc.
- Virtualization: Access to compute and data services, not the servers
themselves, without concern about the infrastructure.
- On Demand: Access to the required resources, when they are needed most, at
the required quality levels.
- Sharing: Enabled collaboration of (virtual) teams, over the Internet, to
jointly work on a complex task.
- Failover: In case of system failure, applications can be migrated and
restarted automatically.
- Heterogeneity: In large and complex Grids, resources are heterogeneous
(platforms, operating systems, devices, software, etc.). Users can choose the
best-suited system for their specific application, or the Grid software will
transparently choose the best resource.
- Utilization: Grids are known to increase average utilization from
approximately 20 percent to 80 percent and more. For example, Sun
Microsystems's internal Enterprise Grid, with currently more than 8,000
processors in three different locations to design next-generation processors,
is utilized at over 95 percent, on average.
These elemental benefits translate into high-level value propositions which
are especially interesting to upper management when considering whether to
adopt and implement a Grid architecture within the enterprise. These values or
macro-level business benefits include:
- Enabling of Innovation: New capabilities, driven by the ability to do
things previously not possible.
- Increased Agility: Shorter time to market, outpacing the competition.
- Reduced Risk: Improved quality and innovation, better business decisions,
increased return on investment and reduced total cost of ownership.
- Increase Agility: Shorter time to market, improved quality and innovation,
reduced cost, increased return on investment and reduced total cost of
ownership.
- Reduce Risk: Better business decisions, faster than competition.
- Enable Innovation: Develop new capabilities, do things previously not
possible.
An Example: MCNC Grid Computing & Networking Services
One example of Grid computing already in practice is MCNC's Grid Computing &
Networking Services (disclosure: I am its managing director). MCNC is an
independent, nonprofit, advanced technology research and service center that
develops, tests and deploys Grid computing and advanced networking solutions
in testbed environments and in production to serve education, research,
government and commercial organizations. MCNC helped create one of the
country's leading Grids, the North Carolina Bioinformatics Grid testbed. MCNC
is currently developing one of the nation's first statewide production Grid
services networks.
Since the mid-1980s, MCNC has operated the state's North Carolina Research &
Education Network (NCREN), a production-level Internet Protocol network that
interconnects over 180 research, education, government and commercial
locations. This high-performance, high-speed communications and computing
network serves as the backbone for future technology growth and is the
foundation for North Carolina's statewide Grid.
In 2001, MCNC and North Carolina universities, in partnership with Cisco
Systems, IBM and Sun Microsystems, launched the North Carolina BioGrid -- one
of the nation's first Grid testbeds for life sciences research. This Grid
offers a reference platform for developing the high-performance computing,
data storage and networking resources needed for bioinformatics and
cheminformatics applications. The testbed currently involves resources from
the University of North Carolina, North Carolina State University, Duke
University and MCNC.
MCNC launched its Enterprise Grid in 2003 to address the needs of a broader
range of scientific disciplines and to provide resources for the N.C. BioGrid
and the statewide Grid. In addition, MCNC is formalizing its early work in
Grid computing with the establishment of the Grid Technology and Evaluation
Center (GTEC). The GTEC is a collaborative Grid deployment testbed for
applications, infrastructure and systems architecture that supports
interoperability, integration, experimentation, development and training.
Through its research and development initiatives, MCNC is addressing the
various challenges of deploying, operating and scaling a Grid infrastructure.
Research efforts underway include Grid-based information retrieval systems,
monitoring and tracking tools, joint collaboration in virtual environments,
on-demand cluster partitioning, high performance network provisioning and
addressing security throughout the Grid.
Agents of Change
From the perspectives of both science and industry, the exponential forces of
the Net Effect demand that we rethink the network -- the clients, the data
centers, the applications and the services. Everything is changing.
Clients will be lightweight appliances with secure Internet or wireless access
to any kind of resources. Data centers will be extremely safe, reliable and
virtually always available -- to anyone, anywhere, using virtually any device.
Applications will be part of a wide spectrum of services delivered over the
network. Such services will include: compute cycles; tools for data
processing; accounting; monitoring, coupled with customized consulting; and
additional information and communication tools, along with software that
allows you to sell or trade your results over the Internet.
With applications added to the Grid, we will be able to build any kind of
information or computing or Web service, delivered in the form of utility
computing, or application service provisioning, or peer-to-peer computing, or
complex, federated Web Services. Everything will be seen in the context of
Grids -- departmental Grids, campus Grids, enterprise Grids, global Grids,
research Grids, education Grids, access Grids, startup Grids, kids' Grids,
entertainment Grids, community Grids, health care Grids and many more.
What Comes After The Grid?
Even though so much of our thinking continues to evolve, I don't expect to see
any real, disruptive, immediate change in information technology or its usage.
We have so many new technologies on the table that we will be busy enough
improving, implementing and deploying them. This takes time. After we have
interconnected all kinds of digital devices through Grid infrastructures, the
next step will be to embed these devices anywhere they can serve our personal
or professional needs -- in our houses, cars, airplanes, clothes and even our
bodies. Tiny injected sensors could be used to monitor the blood-sugar level
of a diabetic, for example, and warn him when another bite of chocolate cake
could send him into insulin shock, or send this information to a datacenter
where it will be evaluated against other similar probes to predict its impact,
automatically inform the doctor and arrange for an appointment.
We are still far away from the global Grid infrastructure we envision today.
We need to continuously work on one common, unifying architecture for Grid
computing (OGSA: Open Grid Services Architecture), and on standards and
protocols for making all the Grid building blocks interoperable. The good news
is, of course, that hundreds of computer scientists in the GGF's
standardization working groups are concentrating on these most important
tasks.
We still have to develop much better software for managing, monitoring,
billing and trading the kinds of computing services discussed in this letter.
And we have to make the computing environments more user-friendly, fully
secure and reliable. And certainly, some non-technical organizational barriers
may still impact widespread adoption of Grid computing on an enterprise scale.
This takes time. I am confident, however, that with the contribution of the
Grid community represented by the Global Grid Forum, and with the ever growing
need in computing power in research and industry, we will have solved most of
the technology issues by 2010.
How Will the Grid Impact Our Future?
In assessing the potential of Grid computing, I'm reminded of some of
history's major milestones:
- In the 19th century, the steam engine ushered in the Industrial Revolution,
enabling mass production and dramatically shortening production cycles.
- In the 20th century, the combustion engine -- in cars, boats and airplanes
enabled us to travel between destinations much faster.
Both were, in essence, time machines.
The Grid (or the Grid Engine, if you like) will be the time machine of the
21st century. It will help young businesses to start up more easily by
remotely accessing the resources required without having to buy them. It will
provide a platform for distant learning and doing, taking higher education
into every corner of the world. It will enable new research and shorten time
to market for many products, enlarging computing space (or cyberspace)
seamlessly. The Grid will be responsible for bringing information, knowledge,
education and health care to the masses, and thus shrink the world and create
an increasingly global education, economy and society: The Grid, the
foundation for e-science, e-business and e-life.
That's another way to define The Grid.
About Wolfgang Gentzsch
Wolfgang Gentzsch is the managing director at MCNC Grid Computing & Networking
Services, an independent, nonprofit, advanced technology research and service
center that develops, evaluates and deploys Grid computing and advanced
networking solutions. Gentzsch joined MCNC in April 2004. He was formerly
senior director of Grid computing at Sun Microsystems Inc. In July 2000, Sun
acquired his company, Gridware, a spin-off of Genias Software, which he
founded in 1990. From 1985 to 2000, he was a professor of Mathematics and
Computer Science at the University of Applied Sciences in Regensburg, Germany.
|
