The Data Avalanche Talk at National Youth Leadership Forum on Technology, aka nerd camp July Jim Gray Microsoft Research [email protected] http://research.microsoft.com/~Gray.
Download ReportTranscript The Data Avalanche Talk at National Youth Leadership Forum on Technology, aka nerd camp July Jim Gray Microsoft Research [email protected] http://research.microsoft.com/~Gray.
The Data Avalanche Talk at National Youth Leadership Forum on Technology, aka nerd camp July 2004 Jim Gray Microsoft Research [email protected] http://research.microsoft.com/~Gray Numbers TeraBytes and Gigabytes are BIG! • • • • Mega – a house in san francisco Giga – a very rich person Tera – ~ The Bush national debt Peta – more than all the money in the world • A Gigabyte: the Human Genome • A Terabyte: 150 mile long shelf of books. Outline Historical trends imply that in 20 years: 1. we can store everything in cyberspace. The personal petabyte. 2. computers will have natural interfaces speech recognition/synthesis vision, object recognition beyond OCR Implications 1. The information avalanche will only get worse. 2. The user interface will change: less typing, more writing, talking, gesturing, more seeing and hearing 3. Organizing, summarizing, prioritizing information is a key technology. Yotta Zetta Exa Peta We are here Tera Giga Mega Kilo How much information is there? Yotta • Soon everything can be recorded and indexed • Most bytes will never be seen by humans. • Data summarization, trend detection anomaly detection are key technologies See Mike Lesk: How much information is there: Everything ! Recorded All Books MultiMedia Zetta Exa Peta All books (words) .Movi e Tera Giga http://www.lesk.com/mlesk/ksg97/ksg.html See Lyman & Varian: How much information http://www.sims.berkeley.edu/research/projects/how-much-info/ 24 Yecto, 21 zepto, 18 atto, 15 femto, 12 pico, 9 nano, 6 micro, 3 milli A Photo A Book Mega Kilo Things Have Changed 1956 • IBM 305 RAMAC • 10 MB disk • ~1M$ (y2004 $) The Next 50 years will see MORE CHANGE ops/s/$ Had Three Growth Curves 1890-1990 1890-1945 Mechanical Relay 7-year doubling 1945-1985 Tube, transistor,.. 2.3 year doubling 1985-2004 Microprocessor 1.0 year doubling Combination of Hans Moravac + Larry Roberts + Gordon Bell WordSize*ops/s/sysprice 1.E+09 ops per second/$ doubles every 1.0 years 1.E+06 1.E+03 1.E+00 1.E-03 doubles every 7.5 years doubles every 2.3 years 1.E-06 1880 1900 1920 1940 1960 1980 2000 Constant Cost or Constant Function? • 100x improvement per decade • Same function 100x cheaper • 100x more function for same price Mainframe SMP Constellation Cluster Constant Price SMP Constellation Graphics/storage Camera/browser Growth Comes From NEW Apps • The 10M$ computer of 1980 costs 1k$ today • If we were still doing the same things, IT would be a 0 B$/y industry • NEW things absorb the new capacity The Surprise-Free Future in 20 years. • 10,000x more power for same price – Personal supercomputer – Personal petabyte stores • Same function for 10,000x less cost. – Smart dust --the penny PC? – The 10 peta-op computer (for 1,000$). 10,000x would change things • Human computer interface – Decent computer vision – Decent computer speech recognition – Decent computer speech synthesis • Vast information stores • Ability to search and abstract the stores. How Good is HCI Today? • Surprisingly good. – Demo of making faces http://research.microsoft.com/research/pubs/view.aspx?pubid=290 – Demo of speech synthesis • Daisy, Hal • Synthetic voice – Speech recognition is improving fast, – Vision getting better – Pen computing finally a reality. – Displays improving fast (compared to last 30 years) Outline Historical trends imply that in 20 years: 1. we can store everything in cyberspace. The personal petabyte. 2. computers will have natural interfaces speech recognition/synthesis vision, object recognition beyond OCR Implications 1. The information avalanche will only get worse. 2. The user interface will change: less typing, more writing, talking, gesturing, more seeing and hearing 3. Organizing, summarizing, prioritizing information is a key technology. Yotta Zetta Exa Peta We are here Tera Giga Mega Kilo How much information is there? Yotta • Almost everything is recorded digitally. • Most bytes are never seen by humans. • Data summarization, trend detection anomaly detection are key technologies See Mike Lesk: How much information is there: Everything ! Recorded All Books MultiMedia Zetta Exa Peta All books (words) .Movi e Tera Giga http://www.lesk.com/mlesk/ksg97/ksg.html See Lyman & Varian: How much information http://www.sims.berkeley.edu/research/projects/how-much-info/ A Photo A Book Mega Kilo Low rent min $/byte Shrinks time now or later Shrinks space here or there Automate processing knowbots Immediate OR Time Delayed And >90% in Cyberspace Because: Point-to-Point OR Broadcast Locate Process Analyze Summarize MyLifeBits The guinea pig • Gordon Bell is digitizing his life • Has now scanned virtually all: – – – – – – – • • • • Books written (and read when possible) Personal documents (correspondence, memos, email, bills, legal,0…) Photos Posters, paintings, photo of things (artifacts, …medals, plaques) Home movies and videos CD collection And, of course, all PC files Recording: phone, radio, TV, web pages… conversations Paperless throughout 2002. 12” scanned, 12’ discarded. Only 30GB Excluding videos Video is 2+ TB and growing fast Capture and encoding I mean everything 25Kday life ~ Personal Petabyte Lifetime Storage 1PB 1000. 100. 10. TB 1. 0.1 0.01 0.001 Msgs web pages Tifs Books jpegs 1KBps sound music Videos Will anyone look at web pages in 2020? Probably new modalities & media will dominate then. Challenges • • • • • • • • Capture: Get the bits in Organize: Index them Manage: No worries about loss or space Curate/ Annotate: atutomate where possible Privacy: Keep safe from theft. Summarize: Give thumbnail summaries Interface: how ask/anticipate questions Present: show it in understandable ways. Memex As We May Think, Vannevar Bush, 1945 “A memex is a device in which an individual stores all his books, records, and communications, and which is mechanized so that it may be consulted with exceeding speed and flexibility” “yet if the user inserted 5000 pages of material a day it would take him hundreds of years to fill the repository, so that he can be profligate and enter material freely” Too much storage? Try to fill a terabyte in a year Item Items/TB Items/day 300 KB JPEG 3M 9,800 1 MB Doc 1M 2,900 1 hour 256 kb/s MP3 audio 1 hour 1.5 Mbp/s MPEG video 9K 26 290 0.8 Petabyte volume has to be some form of video. How Will We Find Anything? • Need Queries, Indexing, Pivoting, Scalability, Backup, Replication, Online update, Set-oriented access • If you don’t use a DBMS, you will implement one! • Simple logical structure: – Blob and link is all that is inherent – Additional properties (facets == extra tables) and methods on those tables (encapsulation) • More than a file system • Unifies data and meta-data SQL ++ DBMS Photos Searching: the most useful app? • Challenge: What questions for useful results? • Many ways to present answers • Detail view Resource explorer Ancestor (collections), annotations, descendant & preview panes turned on Synchronized timelines with histogram guide Value of media depends on annotations • “Its just bits until it is annotated” System annotations provide base level of value • Date 7/7/2000 Tracking usage – even better • Date 7/7/2000. Opened 30 times, emailed to 10 people (its valued by the user!) Get the user to say a little something is a big jump • Date 7/7/2000. Opened 30 times, emailed to 10 people. “BARC dim sum intern farewell Lunch” Getting the user to tell a story is the ultimate in media value • • • • A story is a “layout” in time and space Most valuable content (by selection, and by being well annotated) Stories must include links to any media they use (for future navigation/search – “transclusion”). Cf: MovieMaker; Creative Memories PhotoAlbums Dapeng was an intern at BARC for the summer of 2000 We took him to lunch at our favorite Dim Sum place to say farewell At table L-R: Dapeng, Gordon, Tom, Jim, Don, Vicky, Patrick, Jim Value of media depends on annotations “Its just bits until it is annotated” • Auto-annotate whenever possible e.g. GPS cameras • Make manual annotation as easy as possible. XP photo capture, voice, photos with voice, etc • Support gang annotation • Make stories easy Dapeng was an intern at BARC for the summer of 2000 We took him to lunch at our favorite Dim Sum place to say farewell At table L-R: Dapeng, Gordon, Tom, Jim, Don, Vicky, Patrick, Jim 80% of data is personal / individual. But, what about the other 20%? • Business – Wall Mart online: 1PB and growing…. – Paradox: most “transaction” systems < 1 PB. – Have to go to image/data monitoring for big data • Government – Government is the biggest business. • Science – LOTS of data. Instruments: CERN – LHC Peta Bytes per Year Looking for the Higgs Particle • Sensors: 1000 GB/s (1TB/s ~ 30 EB/y) • Events 75 GB/s • Filtered 5 GB/s • Reduced 0.1 GB/s ~ 2 PB/y • Data pyramid: 100GB : 1TB : 100TB : 1PB : 10PB CERN Tier 0 Information Avalanche • Both – better observational instruments and – Better simulations are producing a data avalanche • Examples Image courtesy of C. Meneveau & A. Szalay @ JHU – Turbulence: 100 TB simulation then mine the Information – BaBar: Grows 1TB/day 2/3 simulation Information 1/3 observational Information – CERN: LHC will generate 1GB/s 10 PB/y – VLBA (NRAO) generates 1GB/s today – NCBI: “only ½ TB” but doubling each year, very rich dataset. – Pixar: 100 TB/Movie Q: Where will the Data Come From? A: Sensor Applications • Earth Observation – 15 PB by 2007 • Medical Images & Information + Health Monitoring – Potential 1 GB/patient/y 1 EB/y • Video Monitoring – ~1E8 video cameras @ 1E5 MBps 10TB/s 100 EB/y filtered??? • Airplane Engines – 1 GB sensor data/flight, – 100,000 engine hours/day – 30PB/y • Smart Dust: ?? EB/y http://robotics.eecs.berkeley.edu/~pister/SmartDust/ http://www-bsac.eecs.berkeley.edu/~shollar/macro_motes/macromotes.html The Big Picture Experiments & Instruments Other Archives Literature questions facts facts ? answers Simulations The Big Problems • • • • • • Data ingest Managing a petabyte Common schema How to organize it? How to reorganize it How to coexist with others • Query and Vis tools • Support/training • Performance – Execute queries in a minute – Batch query scheduling FTP - GREP • Download (FTP and GREP) are not adequate – – – – You can GREP 1 MB in a second You can GREP 1 GB in a minute You can GREP 1 TB in 2 days You can GREP 1 PB in 3 years. • Oh!, and 1PB ~3,000 disks • At some point we need indices to limit search parallel data search and analysis • This is where databases can help • Next generation technique: Data Exploration – Bring the analysis to the data! The Speed Problem • Many users want to search the whole DB ad hoc queries, often combinatorial • Want ~ 1 minute response • Brute force (parallel search): – 1 disk = 50MBps => ~1M disks/PB ~ 300M$/PB • Indices (limit search, do column store) – 1,000x less equipment: 1M$/PB • Pre-compute answer – No one knows how do it for all questions. Next-Generation Data Analysis • Looking for – Needles in haystacks – the Higgs particle – Haystacks: Dark matter, Dark energy • Needles are easier than haystacks • Global statistics have poor scaling – Correlation functions are N2, likelihood techniques N3 • As data and computers grow at same rate, we can only keep up with N logN • A way out? – Relax notion of optimal (data is fuzzy, answers are approximate) – Don’t assume infinite computational resources or memory • Combination of statistics & computer science Analysis and Databases • Much statistical analysis deals with – – – – – – – – – Creating uniform samples – data filtering Assembling relevant subsets Estimating completeness censoring bad data Counting and building histograms Generating Monte-Carlo subsets Likelihood calculations Hypothesis testing • Traditionally these are performed on files • Most of these tasks are much better done inside a database • Move Mohamed to the mountain, not the mountain to Mohamed. Outline Historical trends imply that in 20 years: 1. we can store everything in cyberspace. The personal petabyte. 2. computers will have natural interfaces speech recognition/synthesis vision, object recognition beyond OCR Implications 1. The information avalanche will only get worse. 2. The user interface will change: less typing, more writing, talking, gesturing, more seeing and hearing 3. Organizing, summarizing, prioritizing information is a key technology. Yotta Zetta Exa Peta We are here Tera Giga Mega Kilo Information Avalanche • In science, industry, government,…. – better observational instruments and – and, better simulations producing a data avalanche Image courtesy C. Meneveau & A. Szalay @ JHU • Examples – BaBar: Grows 1TB/day 2/3 simulation Information 1/3 observational Information – CERN: LHC will generate 1GB/s .~10 PB/y – VLBA (NRAO) generates 1GB/s today – Pixar: 100 TB/Movie BaBar, Stanford P&E Gene Sequencer From http://www.genome.uci.edu/ • New emphasis on informatics: – Capturing, Organizing, Summarizing, Analyzing, Visualizing Space Telescope The Evolution of Science • Observational Science – Scientist gathers data by direct observation – Scientist analyzes data • Analytical Science – Scientist builds analytical model – Makes predictions. • Computational Science – Simulate analytical model – Validate model and makes predictions • Data Exploration Science Data captured by instruments Or data generated by simulator – Processed by software – Placed in a database / files – Scientist analyzes database / files e-Science • Data captured by instruments Or data generated by simulator • Processed by software • Placed in a files or database • Scientist analyzes files / database • Virtual laboratories – Networks connecting e-Scientists – Strong support from funding agencies • Better use of resources – Primitive today The Big Picture Experiments & Instruments Other Archives Literature questions facts facts ? answers Simulations The Big Problems • • • • • • Data ingest Managing a petabyte Common schema How to organize it? How to reorganize it How to coexist with others • Query and Vis tools • Support/training • Performance – Execute queries in a minute – Batch query scheduling e-Science is Data Mining • There are LOTS of data – people cannot examine most of it. – Need computers to do analysis. • Manual or Automatic Exploration – Manual: person suggests hypothesis, computer checks hypothesis – Automatic: Computer suggests hypothesis person evaluates significance • Given an arbitrary parameter space: – – – – – – Data Clusters Points between Data Clusters Isolated Data Clusters Isolated Data Groups Holes in Data Clusters Isolated Points Nichol et al. 2001 Slide courtesy of and adapted from Robert Brunner @ CalTech. Data Analysis • Looking for – Needles in haystacks – the Higgs particle – Haystacks: Dark matter, Dark energy • Needles are easier than haystacks • Global statistics have poor scaling – Correlation functions are N2, likelihood techniques N3 • As data and computers grow at same rate, we can only keep up with N logN • A way out? – Discard notion of optimal (data is fuzzy, answers are approximate) – Don’t assume infinite computational resources or memory • Requires combination of statistics & computer science TerraServer/TerraService http://terraService.Net/ • US Geological Survey Photo (DOQ) & Topo (DRG) images online. • On Internet since June 1998 • Operated by Microsoft Corporation • Cross Indexed with – Home sales, – Demographics, – Encyclopedia • A web service • 20 TB data source • 10 M web hits/day USGS Image Data • Digital OrthoQuads – 18 TB, 260,000 files uncompressed – Digitized aerial imagery – 88% coverage conterminous US – 1 meter resolution – < 10 years old • Digital Raster Graphics – 1 TB compressed TIFF, 65,000 files – Scanned topographic maps – 100% U.S. coverage – 1:24,000, 1:100,000 and 1:250,000 scale maps – Maps vary in age User Interface Concept Display Imagery: 316 m 200 x 200 pixel images 7 level image pyramid Resolution 1 meter/pixel to 64 meter/pixel Concept: User navigates an ‘almost seamless’ image of earth Navigation Tools: 1.5 m place names “Click-on” Coverage map Longitude and Latitude search U.S. Address Search External Geo-Spatial Links to: USGS On-line Stream Gauges Home Advisor Demographics Home Advisor Real Estate Encarta Articles Steam flow gauges Click on image to zoom in Buttons to pan NW, N, NE, W, E, SW, S, SE Links to switch between Topo, Imagery, and Relief data Links to Print, Download and view meta-data information Terra Service New Things • A popular web service – Exactly the map you want. • Dynamic Map Re-projection – UTM to Geographic projection – Dynamic texture mapping? • New Data – 1 foot resolution natural color imagery – Census Tiger data • Lights Out Management – MOM – Auto-backup / restore on drive failure New “Urban Area” Data Microsoft Campus at 4 meter resolution “Redundant Bunch 1” Ball field at .25 meter resolution TerraServer Becomes a Web Service TerraServer.net -> TerraService.Net • Web server is for people. • Web Service is for programs – The end of screen scraping – No faking a URL: pass real parameters. – No parsing the answer: data formatted into your address space. • Hundreds of users but a specific example: – US Department of Agriculture TerraServer Web Services Terra-Tile-Service • Get image meta-data • Query TS Gazetteer • Retrieve TS ImageTiles • Projection conversions Landmark-Service • Geo-coded data of wellknown objects (points), e.g. Schools, Golf Courses, Hospitals, etc. • Polygons of well-known objects (shapes), e.g. Zip Codes, Cities, etc Sample Apps • Web Map Client – OpenGIS “like” – Landmarks layered on TerraServer imagery • Fat Map Client – Visual Basic / C# Windows Form – Access Web Services for all data http://terraservice.net Web Services • Web SERVER: – Given a url + parameters – Returns a web page (often dynamic) Your program Web Server • Web SERVICE: – Given a XML document (soap msg) – Returns an XML document – Tools make this look like an RPC. • F(x,y,z) returns (u, v, w) – Distributed objects for the web. – + naming, discovery, security,.. • Internet-scale distributed computing Your program Data In your address space Web Service TerraServer Hardware • Storage Bricks – “White-box commodity servers” – 4tb raw / 2TB Raid1 SATA storage – Dual Hyper-threaded Xeon 2.4ghz, 4GB RAM • Partitioned Databases (PACS – partitioned array) – 3 Storage Bricks = 1 TerraServer data – Data partitioned across 20 databases – More data & partitions coming • Low Cost Availability – 4 copies of the data • RAID1 SATA Mirroring • 2 redundant “Bunches” – Spare brick to repair failed brick 2N+1 design – Web Application “bunch aware” • Load balances between redundant databases • Fails over to surviving database on failure • ~100K$ capital expense. KVM / IP Research Objectives User/App Goals • Public: Access to remote sensing data with no GIS expertise required • Ubiquitous: No special hw/sw required by client • Delivery: All OnLine/Internet Based, no tape or CD distribution • Simple: Designed to be used by a “6th grade geography student” Technology Goals • Test/show scalability • Test/show availability: • Test/show lights out: – all operations & maintenance occurs remotely – Minimal ops and dev staff • “web service” poster child Virtual Observatory http://www.astro.caltech.edu/nvoconf/ http://www.voforum.org/ • Premise: Most data is (or could be online) • So, the Internet is the world’s best telescope: – – – – It has data on every part of the sky In every measured spectral band: optical, x-ray, radio.. As deep as the best instruments (2 years ago). It is up when you are up. The “seeing” is always great (no working at night, no clouds no moons no..). – It’s a smart telescope: links objects and data to literature on them. Why Astronomy Data? IRAS 25m •It has no commercial value –No privacy concerns –Can freely share results with others –Great for experimenting with algorithms 2MASS 2m •It is real and well documented –High-dimensional data (with confidence intervals) –Spatial data –Temporal data •Many different instruments from many different places and many different times •Federation is a goal •The questions are interesting DSS Optical IRAS 100m WENSS 92cm NVSS 20cm –How did the universe form? •There is a lot of it (petabytes) ROSAT ~keV GB 6cm Time and Spectral Dimensions The Multiwavelength Crab Nebulae Crab star 1053 AD X-ray, optical, infrared, and radio views of the nearby Crab Nebula, which is now in a state of chaotic expansion after a supernova explosion first sighted in 1054 A.D. by Chinese Astronomers. Slide courtesy of Robert Brunner @ CalTech. SkyServer.SDSS.org • A modern archive – Raw Pixel data lives in file servers – Catalog data (derived objects) lives in Database – Online query to any and all • Also used for education – 150 hours of online Astronomy – Implicitly teaches data analysis • Interesting things – – – – – – Spatial data search Client query interface via Java Applet Query interface via Emacs Popular -- 1% of Terraserver Cloned by other surveys (a template design) Web services are core of it. Demo of SkyServer • • • • • Shows standard web server Pixel/image data Point and click Explore one object Explore sets of objects (data mining) Data Federations of Web Services • Massive datasets live near their owners: – – – – Near the instrument’s software pipeline Near the applications Near data knowledge and curation Super Computer centers become Super Data Centers • Each Archive publishes a web service – Schema: documents the data – Methods on objects (queries) • Scientists get “personalized” extracts • Uniform access to multiple ArchivesFederation – A common global schema Federation: SkyQuery.Net • Combine 4 archives initially • Just added 10 more • Send query to portal, portal joins data from archives. • Problem: want to do multi-step data analysis (not just single query). • Solution: Allow personal databases on portal • Problem: some queries are monsters • Solution: “batch schedule” on portal server, Deposits answer in personal database. SkyQuery Structure • Each SkyNode publishes – Schema Web Service – Database Web Service • Portal is – Plans Query (2 phase) – Integrates answers – Is itself a web service Image Cutout SDSS SkyQuery Portal FIRST 2MASS INT SkyQuery: http://skyquery.net/ • Distributed Query tool using a set of web services • Four astronomy archives from Pasadena, Chicago, Baltimore, Cambridge (England). • Feasibility study, built in 6 weeks – Tanu Malik (JHU CS grad student) – Tamas Budavari (JHU astro postdoc) – With help from Szalay, Thakar, Gray • Implemented in C# and .NET • Allows queries like: SELECT o.objId, o.r, o.type, t.objId FROM SDSS:PhotoPrimary o, TWOMASS:PhotoPrimary t WHERE XMATCH(o,t)<3.5 AND AREA(181.3,-0.76,6.5) AND o.type=3 and (o.I - t.m_j)>2 SkyNode Basic Web Services • Metadata information about resources – Waveband – Sky coverage – Translation of names to universal dictionary (UCD) • Simple search patterns on the resources – Cone Search – Image mosaic – Unit conversions • Simple filtering, counting, histogramming • On-the-fly recalibrations Portals: Higher Level Services • Built on Atomic Services • Perform more complex tasks • Examples – – – – – Automated resource discovery Cross-identifications Photometric redshifts Outlier detections Visualization facilities • Goal: – Build custom portals in days from existing building blocks (like today in IRAF or IDL) MyDB added to SkyQuery • Moves analysis to the data • Users can cooperate (share MyDB) • Still exploring this • Let users add personal DB 1GB for now. • Use it as a workbook. • Online and batch queries. INT Image Cutout SDSS SkyQuery Portal MyDB FIRST 2MASS The Big Picture Experiments & Instruments Other Archives Literature questions facts facts ? answers Simulations The Big Problems • • • • • • Data ingest Managing a petabyte Common schema How to organize it? How to reorganize it How to coexist with others • Query and Vis tools • Support/training • Performance – Execute queries in a minute – Batch query scheduling