Transcript What’s new?
“Background to eScience” Mark Hayes Technical Director - Cambridge eScience Centre A practical introduction to eScience - Wednesday 26th May 2004 In the beginning… "The collection of people, hardware, and software... will become a node in a geographically distributed computer network…. Through the network... all the large computers can communicate with one another. And through them, all the members of the community can communicate with other people, with programs, with data, or with a selected combination of those resources.” J.C.R.Licklider, “The Computer as a Communication Device” Science and Technology, April 1968 The ARPAnet in 1970 International connectivity - 1991 International connectivity - 1997 International bandwidth From “3D geographic network displays” - Cox et al, ACM Sigmod Record - December 1996 What does the Internet look like? http://www.cybergeography.org/ The World Wide Web Invented at CERN by Tim Berners-Lee in 1989 as a tool for collaboration and information sharing in the particle physics community. The Grid - 1998 Editors: Foster & Kesselman 700 pages 22 chapters 40 authors Analogy with the electrical power grid - just plug in. The Grid - 2003 Editors: Berman, Hey, Fox 1000 pages 43 chapters 116 authors Applications, data sharing and virtual communities. 4 types of Grid • CPU intensive cycle scavenging (SETI@home) • Data sharing • Application provision • Human-human interaction (e.g. Access Grid) Early distributed computing 1.2 million CPU years so far... Brute force attempt to crack strong encryption Protein folding It’s not just compute cycles... An exponential growth in data from many areas of science. The data explosion - some big numbers • CFD turbulence simulations - 100TB • BaBar particle physics experiment - 1TB/day • CERN LHC will generate 1GB/s or 10PB/year • VLBA radio telescope generates 1GB/s today • NCBI/EMBL database is “only 0.5TB” but doubling each year • brain imaging - 4TB/brain at full colour, 10mm resolution (4PB/brain at 1mm i.e. cellular resolution) • Pixar - 100TB/movie FTP and GREP are not adequate (Jim Gray) Application provision • Google - 10K cpus, 2PB database (2 years ago) • free email services - HotMail, Yahoo! 2-10PB storage • netsolve - numerical algorithms on demand with Matlab & Mathematica plugins • renderfarm.net - graphics rendering on demand The Access Grid High end video conferencing and collaboration technology. O(100) nodes world wide. Presenter mic Presenter camera Ambient mic (tabletop) Audience camera “...one of the most compelling glimpses into the future I’ve seen since I first saw NCSA Mosaic.” Larry Smarr The Grid in the UK Pilot projects in particle physics, astronomy, medicine, bioinformatics, environmental sciences... Contributing to international Grid software development efforts 10 regional “eScience Centres” Some UK Grid resources • Daresbury - loki - 64 proc Alpha cluster • Manchester - green - 512 proc SGI Origin 3800 • Imperial - saturn - large SMP Sun • Southampton - iridis - 400 proc.Intel Linux cluster • Rutherford Appleton Lab - hrothgar - 32 proc Intel Linux • Cambridge - herschel - 32 proc Intel Linux cluster • ... • National Grid Service: 4x >64 CPU JISC clusters, HPC(X) Applications on the UK Grid Ion diffusion through radiation damaged crystal structures (Mark Calleja, Earth Sciences, Cambridge) • Monte Carlo simulation lots of independent runs • small input & output • more CPU -> higher temperatures, better stats • access to ~100 CPUs on the UK Grid • Condor-G client tool for farming out jobs Applications on the UK Grid GEODISE - Grid Enabled Optimisation & Design Search for Engineering (Simon Cox, Andy Keane, Hakki Eres, Southampton) • Genetic algorithm to find the best design for satellite truss beams • Java plugins to MATLAB for remote job submission to the Grid • Used CPU at Belfast, Cambridge, RAL, London, Oxford & Southampton Applications on the UK Grid Reality Grid (Stephen Pickles, Robin Pinning - Manchester) • Fluid dynamics of complex mixtures, e.g oil, water and solid particles (mud) • Used CPU at London, Cambridge • Remote visualisation using SGI Onyx in Manchester (from a laptop in Sheffield) • Computational steering Applications on the UK Grid GENIE - Grid Enabled Integrated Earth system model (Steven Newhouse, Murtaza Gulamali - Imperial) • Ocean-atmosphere modelling • How does moisture transport from the atmosphere effect ocean circulation? • ~1000 independent 4000year runs (3 days real time!) on ~200 CPUs • Flocked condor pools at London & Southampton • Coupled modelling The European DataGrid • Tiered structure: Tier0=CERN • Lots of their own Grid software •Applications: particle physics, earth observation, bioinformatics http://www.eu-datagrid.org/ NASA Information PowerGrid • First “production quality” Grid • Linking NASA & academic supercomputing sites at 10 sites • Applications: computational fluid dynamics, meteorological data mining, Grid benchmarking http://www.ipg.nasa.gov/ TeraGrid • Linking supercomputers through a high-speed network • 4x 10GBps between SDSC, Caltech, Argonne & NCSA • Call for proposals out for applications & users http://www.teragrid.org/ Asia-Pacific Grid • No central source of funding • Informal, bottom-up approach • Lots of experiments on benchmarking & bio apps. http://www.apgrid.org/ £1 buys... • • • • • • • 1 day of cpu time 4 GB ram for a day 1 GB of network bandwidth 1 GB of disk storage 10 M database accesses 10 TB of disk access (sequential) 10 TB of LAN bandwidth (bulk) How do you move a terabyte? Source: Terascale SneaketNet, Jim Gray et al Context Speed Mbps Rent $/month $/Mbps $/TB Sent Time/TB Home phone 0.04 40 1,000 3,086 6 years Home DSL 0.6 70 117 360 5 months T1 1.5 1,200 800 2,469 2 months T3 43 28,000 651 2,010 2 days OC3 155 49,000 316 976 14 hours OC 192 9600 1,920,000 200 617 14 minutes 100 Mpbs 100 1 day Gbps 1000 2.2 hours Some consequences Compute cycles are (almost) free... by comparison with network costs. -The cheapest and fastest way to move 1TB of data out from CERN is still by FedEx. Though this considers only bandwidth, low latency networks are even more expensive! (MPI over WAN doesn’t work well.) What makes a good Grid application? A distributed community of users. Tiny network input & output, huge compute requirement. Database access & storage is also expensive, therefore put the computation near the data. Questions?