Title: Grids: How do we define success
1Grids How do we define success?
- Vicky White, Fermilab
- April 27, 2004
- ISGC-2005 , Taiwan
2Views of a Straddler
- Verb to straddle
- To stand or sit with a leg on each side of
bestride straddle a horse. - To be on both sides of extend over or across a
car straddling the centerline. .
3Hopefully not a Straggler
- Strag"gler1. One who straggles, or departs from
the direct or proper course, or from the company
to which he belongs one who rambles without any
settled direction.
4What is a Grid?
- So many definitions e.g.
- Coordinated resource sharing and problem solving
in dynamic, multi-institutional virtual
organizations. Ian Foster - So many analogies but most are not very useful
- Power Grid
- The next big thing after the WWW ?
- Seamless virtual environment
- So many technologies and types of Grid
- Globus almost a standard but not universal
- Data Grid, Computational Grid, Campus Grid,
National Grid, Climate Grid, and so on. - So much hype and such high expectations !
- Confusion with
- Public Resource Computing e.g. SETI_at_home
- Plain old distributed computing, commercial
enterprise computing
5Global GridCommunity
6The (Dubious) Power Grid Analogy
- Do we ship work to the power source?
On-demand access to and integration of resources
services, regardless of location
Source Ian Foster
7Grid Vision
Federico.Carminati_at_cern.ch
8Grid Vision
- Useful, able to support/promote new science
- Usable (accessible, robust, easy-to-use)
- High-capacity (rich in resources)
- High capability (rich in options)
- Evolutionary (able to adapt to new technologies
and uses) - Persistent (usable by community in the long term)
- Stable (usable by community in the short term)
- Scalable (growth must be a part of the design)
- Integrative (promoting end-to-end solutions)
Source Fran Berman, SDSC
9Grid Vision CyberInfrastructure?
10Why do we want to use Grids for Science?
- Cost ?
- Reliability ?
- Ease of use ?
- Scalability with growing datasets ?
- Attracts funding?
- ?
11Large Hadron Collider(LHC) at CERN
12Is Grid really about costs and technology?
- The way to get massive computing cycles?
- The CMS detector alone cost several hundred
million CHF and the LHC Accelerator several
billion CHF - The total estimated cost of the Tier0 and Tier 1
center at CERN for LHC is in the tens of millions
of CHF - Total estimated cost for computing might be 3
times this - Power and cooling for all this is a problem but.
- The only way to store multi-petabyte datasets ?
- Today at Fermilab we have robotic storage that
houses 2.5 PB of data - Potential for 13 PB with todays tape technology
and probably at least double that by 2008 - The way to reasonably share the costs ?
13Is Grid about scavanging unused computing cycles?
- NOT Public Resource Computing
- seti_at_home
- Einstein_at_home
- folding_at_home
- LHC_at_home
- You-name-it_at_home
14Why do companies jump on Grid
- To make money
- To control markets
- To leverage open source into market leadership
- For Science and Computer Science these are not
our goals or the ways in which we determine
success.
15The Value of Grid ComputingIBM Perspective
Increased Efficiency
Higher Quality of Service
Increased Productivity ROI
Reduced Complexity Cost
Improved Resiliency
16Grids HP Perspective
computing utility or GRID
virtual data center
value
programmable data center
grid-enabled systems
UDC
Tru64, HP-UX, Linux
clusters
Open VMS clusters, TruCluster, MC ServiceGuard
today
shared, traded resources
17Grid - its really about collaboration!
- Its about sharing and building a vision for the
future - And its about getting connected
- Its about the democratization of science
18Cathedral and Bazaar
- Linux overturned much of what I thought I knew.
I had been preaching the Unix gospel of small
tools, rapid prototyping and evolutionary
programming for years. But I also believed there
was a certain critical complexity above which a
more centralized, a priori approach was required.
I believed that the most important software
(operating systems and really large tools like
Emacs) needed to be built like cathedrals,
carefully crafted by individual wizards or small
bands of mages working in splendid isolation,
with no beta to be released before its time. - The Linux community seems to resemble a great
babbling bazaar of different agendas and
approaches.. out of which a coherent and stable
system could seemingly emerge only by a
succession of miracles - Eric S. Raymond, The Cathedral and the
Bazaar
19The Grid Bazaar
- Many Grids and Grid environments
- An evolutionary process
- An environment for innovation and prototyping
ways of achieving the grand vision - A place where seemingly miracles are needed to
make it all work together - But really it is smart people and the will to do
so - Interoperability of seemingly divergent efforts
will become the norm - Standards must only come later and not suppress
the possibility of achieving the grander goals
and visions
20For Success Define a Destination?
or look for a pot of gold?
To travel hopefully is a better thing than to
arrive. Robert Louis Stevenson
21Borromean Rings
These rings have the interesting feature that,
while no two rings are linked, the set of rings
is linked. Remove any one ring, and the system
falls apart.
Computer and Information Science
Behavioral and Organizational Sciences, (Policy,
Government, Economics)
Scientific Research and Education
22Collaboration
- EMSL Collaboratory
- The logo of the Collaboratory Project at the
Environmental Molecular Sciences Laboratory uses
the rings to symbolise collaboration between the
researchers. Their website expains - The Borromean Rings are three symmetric loops.
Although the rings are interwoven and cannot be
pulled apart, no two rings are interlinked -
removing (breaking) one allows the other two to
slide apart. This is the nature of collaborative
work as well - people working together create
something that is more than the simple sum of
their individual efforts. This synergism is the
principle behind the Collaboratory concept.
23Pasteurs Quadrant
Creation of knowledge basic, curiosity-driven
research
Application of knowledge
Classic Linear Research Model
Pasteurs Quadrant Research Model
Pasteur
Yes
Bohr
Focus on New Knowledge Creation?
No
Edison
No
Yes
Focus on Application?
Source Dan Atkins
24Many possible Grid Quadrants
- Information Technology v. Science Applications
- Local economic, network and IT development v.
Global Science collaborations and projects - Physics v. Biology
- Earth Observation v. ?
- Education v. Scientific Research
25Fermilab and Collaboration
- Active experiments from the 2003 Research
Program Workbook - Of 213 Institutions involved 114 of them are
non-US - Of 1916 physicists, 753 are non-US
- Of 699 students, 234 are non-US
26Taiwan-Fermilab Collaboration on physics
- Incredibly valuable contributions to the CDF
experiment over many years - Now we travel the Grid road together also
27Open Science Grid a Collaboration
28Working towards the Grid vision- a building
process, not a deployment activity
- How many physicists does it take to change a
light bulb - Answer One to hold the lightbulb, but more than
100 to turn the building - How many physicists does it take to BUILD a Grid
- Two typical types of answer
- (1) None, because they all know that someone else
is deploying it for them - (2) Three times as many as it would take to get
their work done without using a Grid
29Grids Metrics for Success
- How many jobs are run ?
- With what level of errors?
- How much data is stored/moved/read/written?
- How many people are using a specific Grid ?
- How many press releases are there about a Grid?
- How much money do governments allocate to a Grid?
- How easy to use is a Grid?
- How much more science gets done from use of a
Grid? - How effectively are resources being shared and
well used?
30Grids Metrics for Success
- All of the previous metrics are great indicators,
but not the real measure for success - Too many spins and interpretations
31Grid Interoperable set of cyber infrastructures
(Grids)
A few problems still to solve
- Economy
- Payment
- Value of contributions
- Currencies
- Policy
- Who can use?
- Who is responsible?
- Who can decide?
- Security (of course)
32If we can work together
- To make a contribution to all of the following
- Science and Scientific Research
- Society and its prosperity and harmony
- Education of the next generation
- Creating a dynamic collaborative environment for
problem solving and innovation - Then we will have succeeded
33Contributing and Communicating
- If we can in some small way contribute to peace
and prosperity in the world - If we can help to close the digital divide and
make contributing to science and technology and
innovation a more democratic process (between the
haves and the have-nots) - If we can speak to our governments and have
them better understand the Grid Quadrants and
the Borromean rings I spoke of earlier - Then we will have succeeded
34Success on a Worldwide scale
- If we can bring together people from all over the
world (whether they be physicists, biologists,
computer scientists, climate researchers or .)
and they - Want to be part of building the cyber
infrastructure or Grid environments or
e-science environments for the future - Actively participate
- Get benefit from the collaboration
- Then we will be succeeding
35Success on a Worldwide scale
- The work will get done
- Scalability issues will get addressed
- Robustness will come
- Interoperability will be made to work
- Even more diversity than we see today
- Because people working together can and will make
things work if they are involved and getting
benefit for themselves, their institutions, their
science, their nation in a broad way
36Success is when we clearly communicate
- To each other
- To governments and funders
- To the public
- To our communities
- And the message we give is clear and positive
37The Message
- Collaboration is essential on a worldwide scale
- To carry out many of the ambitious endeavors we
want to undertake such as LHC and LHC
experiments and after that the International
Linear Collider and of course many other large
international science endeavors in fields other
than physics - Diversity is the norm and healthy
- Networks are fundamental and high speed
connectivity is essential - Data is pervasive and ever growing and we need to
analyze it, organize it, mine it, catalog it,
curate it, understand it for all purposes ranging
over - Wallmart inventory tracking
- Climate and Earthquake predictions
- Finding the Higgs? And understanding the
fundamental nature of matter - Mapping the human genome
- Fighting cancer
- And much more..
38Success
- When Grid the vision and the way of working -
helps us get this message out - Without finger pointing
- Without complaining about technology and its
imperfections it will be gone before its
perfect - With balance between what must work today and
what might help us tomorrow
39ISGC-2005
- This workshop is one of the very signs of success
- The program is really excellent and diverse with
strong participation from Asia-Pacific, Europe
and the Americas - Lets shut down our email and listen to the
speakers and hear how their solutions to problems
(in the small) fit in and each play a part in
building a Grid environment (in the large) - Lets communicate by listening
40Thank you
- To Academia Sinica for hosting us and to Simon
Lin and all of his staff for organizing this
workshop - To all of you who have been and will be
collaborators on both Science and Grid projects - Thank you for your energy and willingness to work
together despite financial and cultural pressures
that make collaboration hard - To all of you for listening