[Air-l] e-science, the grid, and supercomputers - inter2?

elijah wright elw at stderr.org
Sun Sep 28 11:23:32 PDT 2003


> Ok, doesn't grid sound like inter2 all over again? Anyone out there
> working on project implementation with an inter2 link that is going
> well?? and I have continually heard that IPv6 is coming. Has it arrived
> in the US or anywhere else for that matter?

"grid" computing is being pushed as mostly cell-based - for ex, pods of
nodes with gigabit links between them, seperated by slower 10mbit or
100mbit lines - while I2-style research was characterized by its focus on
massive bandwidth consumption without much regard for the computation
happening at the peripheries.

so, paradigm 0: "gosh this network is slow..."

paradigm 1 (I2): "we need more bandwidth, we'll figure out what to
do with it later..."

paradigm 2 (grid-ish):  "golly, now that we have a lot more bandwidth, it
sure would help to be able to generate computationally intensive problems
that take advantage of that... whether they require massive data transfer
or just massive amounts of intercommunication...."



elijah




More information about the Air-L mailing list