There was an interesting article from SIGCOMM posted yesterday about in-center-inter-connect: http://cseweb.ucsd.edu/~vahdat/papers/portland-sigcomm09.pdf http://ucsdnews.ucsd.edu/newsrel/science/08-09PortLand.asp
http://rss.slashdot.org/~r/Slashdot/slashdot/~3/pTgFA2gqsEg/How-To-Build-a-100000-Port-Ethernet-Switch http://www.networkworld.com/news/2009/081709-portland.html There's going to be a separation between dense clouds, where you get excellent connectivity within the cloud and fast vm migration when nodes or arcs fail, and sparse clouds where everything sort of works like the internet, somewhere on a scale of excellent to not at all. -- rec -- On Thu, Aug 20, 2009 at 11:26 AM, Marcus G. Daniels <[email protected]>wrote: > Douglas Roberts wrote: > >> Interesting article about cloud computing on Slashdot today: >> >> >> http://tech.slashdot.org/story/09/08/20/0327205/Amazon-MS-Google-Clouds-Flop-In-Stress-Tests?art_pos=7 >> >> One nice thing about what Amazon does in contrast to most supercomputing > centers is to let you boot whatever kernel image you want. That can be > important for diagnosing and fixing some kinds of problems. > > I kind of doubt the commercial players pony up for the kind of > interconnects used on supercomputers though. Maybe the NMCAC could > distinguish itself from the commercial players by providing users total > control over hardware, while also providing an interconnect that can absorb > stress? > > Marcus > > > > ============================================================ > FRIAM Applied Complexity Group listserv > Meets Fridays 9a-11:30 at cafe at St. John's College > lectures, archives, unsubscribe, maps at http://www.friam.org >
============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org
