Berkeley%20RAD%20Lab%20Center%20Proposal

1
Berkeley RAD LabCenter Proposal

• Armando Fox, Randy Katz, Michael Jordan, Dave
  Patterson, Scott Shenker, Ion Stoica
• RADS Retreat, June 2005

2
Setting the Context

• Academic research trains next generation of IT
  leaders AND helps expand IT market
• 19 1B industries from IT research (NAE)
• Long-term, experimental research been key
• Synergy of Industrial Academic research
• Academic govt (DARPA)/industry funded (gt80 /
  lt20)
• 3-6 faculty, 12-24 grad student, 2-3 staff, 4-5
  yrs
• Berkeley SW Spice/Magic CAD programs, BSD
  (Berkeley SW Distribution) Unix op. sys.,
  Ingres/Postgres Relational Databases,
• HW RISC (Reduced Instruction Set Computer),
  RAID (Redundant Array of Inexpensive Disks),
  NOW (Networks of Workstations),

3
Traditional Research Funding Today

• Avg. industry research shorter term today
• DARPA exiting long-term (exp.) IT research
• 03-05 BAAs 9 AI, 2 classified, 1 SW radio, 1
  sensornet, 1 reliability
• Academic led funding reduced 50 in 3 years
• Faculty consultants in consortia led by defense
  contractor, get grants support 1-2 students in
  return
• NSF swamped with proposals, conservative
• 2000 to 6500 proposals in 5 years
• IT has lowest acceptance rate at NSF (between 8
  to 16)
• Ambitious proposal is a negative review
• Even if get NSF funding, proposal reduced to
  stretch NSF e.g., ServRADS got 3 x ?
  faculty, 6 grad students, 0 staff, 3 years
• (To learn more, see www.cra.org/research)

4
Sustaining Innovation Engine?

• Replicate research centers based primarily on
  industrial funding to train next generation of IT
  leaders and expand IT market
• Berkeley Wireless Research Center (BWRC) 50 grad
  students
• Stanford Network Research Center (SNRC) ?? Grad
  students
• Exciting, long term technical vision
• Demonstrated by prototype(s)
• Industry largely funds, so industry helps set
  directions
• N companies, where N 6
• center proposes 1st draft of new standard
• Multiple interfaces Framework, OS, server,
  router,

5
Existing Center Models SNRC
FoundationMember
(http//snrc.stanford.edu/40.html)
Industrial Executive Council v
Place a Researcher-in-Residence v
CTO Roundtable/Forum v v
Technical Advisory Board v v
Initiate targeted Ph.D. student mentorship program v v
Industrial Review Committee v v v v
Annual Contribution 250k 250k 50k 50k
Senior Member
AffiliatePartner
Must also endow 1 Faculty Chair 2 Grad
Fellowships (1 X 1.5M 2 x 0.3M)
6
Existing Center Models BWRC

• IP put in public domain or offer no-fee,
  non-exclusive licenses

ParticipatingMember
(bwrc.eecs.berkeley.edu/Background/Charter.htm)
Member, Board of Advisors Member, Board of Advisors v
1 Provisional Industrial Researcher (lt 2 months, no UC IP agreement) 1 Provisional Industrial Researcher (lt 2 months, no UC IP agreement) v
1 Visiting Industrial Fellow ( 2 months, sign UC IP agreement) 1 Visiting Industrial Fellow ( 2 months, sign UC IP agreement) v
Attend 2 research reviews Attend 2 research reviews v v v v
Donations Product, equipment or technology Donations Product, equipment or technology v v v v
Annual Grant Contribution 150K 150K 75K 75K 75K
Associate Member
BWRC gets ½ funding via MARCO center, funded
50 industry 50 DoD
P.M. who is also a CITRIS Founding Corporate
Member called Platinum
7
Founding the RADLab Start Jan 06?

• 2.5M / yr (½ BWRC) 80 industry, 20 govt
• 5 years 25 grad students, 25 undergrads , 6
  faculty, 2 staff
• Looking for 3 founding companies to fund 3-5
  years _at_ cost of 2 employees / year 3
  founding companies to fund 3-5 years _at_ cost of
  1 employee / year
• Follow Berkeley Wireless Research Center model
• IP model, Board of Advisors, Participating/Associa
  te Members, Industrial Fellow, Executive Director
  for day-to-day operations,
• Add good ideas from SNRC Targeted Mentor,
• Your good ideas go here (Design review, brain
  storming, ?)
• Got leads or advice? Contact one of us fox,
  randy, jordan, pattrsn, shenker, stoica
  _at_cs.berkeley.edu

8
Overarching Mantra

• Enable a faster pace of network service
  innovationthrough new distributed system
  architectures that reduce operations cost by 2-3
  orders of magnitude
• The Challenge
• Software systems Too much information ? make
  sense of it through statistical learning
  control theory
• Network systems Too little information ?exploit
  better observation and monitoring in the network
  infrastructure to drive management processes
• The Payoff
• Single person can write, profile, deploy, operate
  the next-generation IT business (the Fortune 1
  million)
• Do for Internet apps what Web did for individual
  publishing