Title: Planning for Cyberinfrastructure CI Development in EPSCoR States
1Planning for Cyberinfrastructure (CI) Development
in EPSCoR States
- Discussion breakoutsNSF EPSCoR CI workshopMay
10-12 2006 - Compiled byTerry Moore and Mel Ciment
2Q1. Drivers and enabling technologies
- What are the main opportunities, applications and
enabling technologies driving the development of
CI plans? - What are the perceived benefits of
Cyber-infrastructure deployment in your State? - What are the major RD Applications Drivers that
require advanced Cyber-infrastructure? - What are the key enabling technologies for
advancing your CI plans?Facilitator Micah
BeckScribe Rich Sincovec
3Key enabling Technology Networking
- Networking is important but cyber infrastructure
is much more so - RON
- Regional optical networks are being formed with
examples in EPSCoR states. - Costs and users still are issues.
- Top down or bottom up have driven the deployment
of these RONs
4Non-networking Tech Enablers
- Storage
- Computing
- Instruments
- Applications
- Visualization
- NSF Centers (probably underused by EPSCoR states)
- Why?
5Main Benefits of Cyberinfrastructure?
- Inside or outside of state?
- Education opportunities
- Economic drivers?
- Example of obstacle
- Willingness to share data
6Applications
- Medical resources multi-state collaborative
project that could access might enable
opportunities. - Solve locally or in a broader sense.
- Education and training
- More than just bandwidth
- Drought risk assessment
- Multiple data bases
- Sharing of data
- Natural resource management
- Agriculture decision support
- Synchronous and asynchronous approaches
7Applications
- Application driven approaches
- CI driven approaches
- Co-funding is appropriate, e.g., CISE plus
application area - NSF is looking for mechanisms so CI driving
applications or applications driving CI my depend
on NSF program. Co-funding is a high priority. - TN 110 plan reaching rural parts of state also
included high performance backbone. Education,
electronic library, large caches at high end
computing, health care. - Medical imaging
- Resource sharing of instruments
8Applications
- Multiple resolution imaging
- Access to data, discover new results in the data
- Lot of important applications need better tools
so need people for survivability. - People may be harder to get than equipment.
Equipment rich and people poor. - Underlying issue is always sustainability.
9Requirements
- People, people, people
- Gaming to get the people and yet have the
resources - Renovation costs are important but source of
funds is problematic - Access Grid support was problematic
10Key Enabling Technologies
- Campus infrastructure networking
- Firewalls create barriers
- Costs per port may restrict innovation
- Other network operating restrictions
- Separate research network on campus
- Question Are these substantial obstacles?
- Need to work and collaborate with campus IT folks
11CI in General
- Some things can be done in the lab
- Desire to deploy global infrastructure
- Put stuff out there and see what happens
- Social thing is a driver
- Create separate networks
- Technology to minimize IT folks issues
12CI for Other Disciplines
- What are our responsibilities and role?
- Do we have the tools?
- Hartford Life Ins has implemented CI
- CI is real What are the benefits?
- Spread CI around or the digital divide will
become even larger - Money to be made so companies like Microsoft
should have a role in addressing the digital
divide - We cannot wait for the commercial folks since CI
enables research which commercial may not have as
a high priority
13Q2 Requirements
- What are the requirements for successful
implementation of Cyber-infrastructure in your
state? - How do you plan and implement CI advances?
- What are the impacts of regional and statewide
initiative in deploying network capacity? To what
extent are your advances limited by lack of such
plans? - Are you able to attract qualified personnel and
leaders in IT departments? - Facilitator Mary Fran Yafchak
- Scribe Seth Lilly
14Initial Themes
- Networking
- Dark Fiber available, but restricted due to
political or financial reasons. - Wide range of connectivity options cable. T1,
fiber, microwave - Alaska has telcos negotiating the majority of the
higher-speed connections. - Canada is telco friendly much nicer regulation
therefore, more connectivity. - Organizational inconsistencies exist and make CI
planning difficult
15Required Homework
- Understand and address the states climate
(political, carriers, K-12, higher ed) - Understand existing infrastructure
right-of-way, dark fiber, etc. - Research leadership One or two dedicated
researchers can make a difference in
participating in a grid/connectivity environment.
- Identify which resources/stakeholders
(government/education/carriers/businesses) are
useful in solving each step of implementation - Tennessee contributors included UT System, Board
of Regents (about 20 universities) and Economic
Development - Learn from the experiences of other successful
implementations, through networking opportunities
(Internet2, EPSCoR, SURA)
16Required Goals and attitudes
- Regionalism (looking outside state boundaries)
- Be persistent with your funding requests.
- Have a purpose for implementation (application
examples) - Organize for good internal communications
17Impediments and Facilitators
- Impediments for successful CI implementation
- Carrier support
- State support
- Facilitators of successful CI implementation
- Additional assets
- Innovative thinking
18How to Plan and Implement CI advances
- Identify a neutral group to develop a viable
demonstration and model. - Groups that begin partisan end partisan.
- Regular and persistent communication to core
groups political, constituents (targeted
audiences) and those who have existing
implementations. - Finally identify your champions political,
educational, business and utilize them
strategically.
19Can you get the people you need?
- We are not able to attract the qualified
personnel and leaders in IT That we need. - This is a universal issue, and a strong
recommendation is given to train your own.
Cultivation and retention are issues which need
to be addressed.
20Q3. Organizational Mechanisms and Key Players
What are the most constructive organizational
mechanisms and associated key individuals to help
develop CI plans?Facilitator Bonnie
NeasScribe Dorette Kerian
21a) Stakeholders and how do they communicate?
- Need common vision and goal for CI.
- Stakeholders are those who believe in visions and
goals and meet regularily to advance goals. - In planning stage, stakeholders are researchers,
educators, public safety and business leaders,
etc.
22b) What degree of State financial commitment is
necessary for development?
- Need high level of state commitment - 50 or more
but less than 100. - Need to develop business plan and find state
funding long enough to get through initial
development. - States should commit to CI analogously to the
multiple funding mechanisms used for state
highway systems
23c) Are there advocates for cross-institutional
and statewide CI development?
- Yes, Stakeholders listed in A plus national
agencies (such as NSF EPSCoR and Oak Ridge
National labs) as well regional and national
educational affiliations. - There is potential for catalyst but, potential
for obstructionists. Theres strength in numbers
and value in a champion.
24d) The most effective state-wide regional
collaborations for promoting CI
- Most effective statewide and regional
collaboration. States are relaxing policy to
enable collaboration. - States have different examples of how this works
but recent events from 9/11 and Katrina have
forced some to focus on more pressing needs. - Indiana benefits from having a state foundation
with funds, - SURA regional organization addresses common
visions and issues - Louisiana collaboration was brought together by
key researcher champion.
25Other
- Stakeholders differing needs/levels of
expectations result in varying demands on CI
network - There is a need for a national vision.
- Local jurisdictions can also assume a role in
developing CI, example local sales tax.
26Q4. Challenges and Case Studies What was
required for successful plan implementation?
Adapted from John Connollys Notes
- What mechanisms have been successful in promoting
buy-in from various stakeholders? - Are cost-benefits analyses of application drivers
or critical components helpful? - What leadership is needed for the transition from
plan development to CI implementation? - What can we learn from other successful CI
implementations?
27Successful University Sector buy-in mechanisms
- Research must drive justification
- robust cyberinfrastructure is necessary to
compete for national programs - examples GEON, NEON, Teragrid, Petascale
facilities, , QUASI (hydrology) CERN Tier 2 ,
SciDAC, etc - Forming a open community of science such as OSG,
SURAgrid, GridChem - Other applications like teleconferencing and
telemedicine. - Counterargument If your average use is 10 why
do you need more capacity? - Answer The bursty nature of research and other
examples.
28Successful Private Sector Buy-in Mechanisms
- Some companies use CI to create new jobs, BUT
some, e.g., telecommunications business, regard
public programs as competition. - The challenge is to convince industry that
partnering with the research community is to
their advantage. - Recall that the original internet wasnt of any
interest to private industry until it had been
seed funded by the Federal Government.
29Successful State Government Buy-in Mechanisms
- Many functions of state government,- connections
to county courthouses, drivers license , tax
collection can be enhanced by partnership with
academic grids. - Most state governments need to be convinced that
CI will provide economic development , and it is
a challenge to convince them hat there is a
connection.. the private sector needs to help
make the argument that CI can provide JOBS! - One of the obstacles is the EPSCoR Dilemma
Most of the EPSCoR jurisdictions are not heavily
industrialized so there are less advocates from
the private sector, and the costs of networking
are high due to being far from the central part
of the country (AK,HI and PR) are far from major
urban centers where the gigapops are. Both
factors increase the cost/benefit ratio.
30Proposed Federal Government Role
- The NSF has in the past subsidized network
connections to universities (e.g.Internet2) but
doesnt do so now. - This policy should be updated in the light of the
EPSCoR opportunity. - The essence of EPSCoR is to create the level
playing field, but the field in
cyberinfrastructure is becoming less level.. the
gap between CA and KY is widening. Thomas
Friedman says that the world is flat, but perhaps
the US is not? (actually Friedman also admits
that the world is not level in certain cases,-
think of the EPSCoR part of the world like
sub-Saharan Africa Grid-wise.) - The Federal Government should promote flatness by
starting a Rural Gridification Program to promote
research, development, education and training - Analogous to the Rural Electrification
Administration of the 1930s - This could also include rural parts of non-EPSCoR
31Needed State Leadership
- In many states, the EPSCoR committee plays a key
coordinating role, since it is often one of the
only entities which cares about state wide
issues. - Kentucky EPSCoR has special programs for the more
rural areas in the state, and is part of a larger
structure KSTCKentucky Science and Technology
Council which looks at development, outreach and
educational issues, all of which are affected by
having a robust and advanced cyberinfrastructure. - State governments can play key roles in trying to
secure Federal government support from a variety
of agencies, not just NSF - Hawaii, one of the furthest of the EPSCoR states
gets help from the Defense Department in
networking to the mainland. Cyberinfrastructure
will be very important to the Department of
Homeland Security which must ensure that the
nations critical data structure be resilient to
malicious attack.
32Lessons Learned from CI Implementations
- Of course, many of the non-EPSCoR states
(California for example) have gargantuan
cyberinfrastructure, and those would not be
suitable models. - One EPSCoR state which has lead the way is
Louisiana which has invested substantial funding
into cyberinfrastructure in all its aspects, -
including big iron hardware, wide band networks,
and software and training activities. - Another plan, which we hope will be successful,
is the Northern Tier project in North Dakota and
neighboring states. - Outside the US, such EPSCoR-like entities as
Canada and Korea have provided very advanced
cyberinfrastructure (more advanced than many
parts of the US in fact) and they, and their
national justifications, could be used as very
useful role models.