4
Disclaimer (1) Disclaimer (1) Not exactly “ideas I’d like to see worked on” More “computing paradigms we ought to pay attention to”

Disclaimer (1) Not exactly ideas Id like to see worked on More computing paradigms we ought to pay attention to

Embed Size (px)

Citation preview

Page 1: Disclaimer (1) Not exactly ideas Id like to see worked on More computing paradigms we ought to pay attention to

Disclaimer (1)Disclaimer (1)

• Not exactly “ideas I’d like to see worked on”

• More “computing paradigms we ought to pay attention to”

Page 2: Disclaimer (1) Not exactly ideas Id like to see worked on More computing paradigms we ought to pay attention to

Disclaimer (2)Disclaimer (2)

• Expect to apply these paradigms in my own work

• “Tooting my future horn”

Page 3: Disclaimer (1) Not exactly ideas Id like to see worked on More computing paradigms we ought to pay attention to

Paradigm 1: Autonomic/AdaptiveParadigm 1: Autonomic/Adaptive

• Yes we’ve looked at it some, e.g.:– Database design advisors

– Adaptive query processing

• But we can be more sweeping– Discard configure-compile-run mentality

– Embrace run-monitor-adapt-continue

• Widely applicable– Traditional query optimization to web-scale

information brokering and lots in between

Page 4: Disclaimer (1) Not exactly ideas Id like to see worked on More computing paradigms we ought to pay attention to

Paradigm 2: Paradigm 2: Simple Local Decision-MakingSimple Local Decision-Making

• Provoked by a theory faculty candidate in distributed algorithms If each site makes a locally-optimal decision, how

far is it from the globally-optimal decision made by a know-it-all centralized site? (Answer: not far)

• Forget complicated and expensive distributed protocols– Simple independent decisions are good enough

Especially if you use Paradigm 1Paradigm 1 at the same time