Peer-to-Peer Information Systems Week 1: IntroductionClass OverviewGradingClass Project(s)Class SessionsSlide 6ReadingsClient-ServerIs P2P really new?History: UUCPHistory: Usenet, NNTPHistory: DNSHistory: xFSComputers and PeopleSlide 15Slide 16What is P2P?What is P2P?P2P Litmus Test (p. 22)Slide 20Hybrid: C-S & P2PFull P2PContent at the EdgesBetter? Or Just Different?Does Content Influence Design?“Users reward simplicity” (p. 27)P2P - the cure for what ails you?“The P in P2P is People” (p. 49)It’s a Small WorldGraphs of Social NetworksZipf’s LawZipf’s LawPower LawWhy Do We Care?Peer-to-Peer Information SystemsWeek 1: IntroductionOld Dominion UniversityDepartment of Computer ScienceCS 495/595 Fall 2003Michael L. Nelson <[email protected]>09/26/03Class Overview•class homepage:–http://www.cs.odu.edu/~mln/teaching/cs595-f03/•class email list–http://list.odu.edu/listinfo/cs595-f03/Grading•Undergraduates (CS 495):–60% - class projects–20% - quizzes–20% - final exam•Graduates (CS 595):–50% - class projects–10% - class project presentations–20% - quizzes–20% - final examClass Project(s)•multi-stage P2P development project–more details in the next session•2-person groups–1 grad + ugrad–use email list to find partners if you don’t accomplish this after classClass Sessions•Tuesdays–lectures, case studies•Thursdays–discussion, “lab”, demo of assignments, quizzesGrading•4 quizzes–on any Thursday during the semester–based on assigned readings•open notes, in class, ~10 minutes•If the class average is > 85%, then there will be no final–your quiz grade will double as your final grade–please do us all a favor -- do your readings!!!Readings•required text:–Andy Oram (ed.), Peer-to-Peer: Harnessing the Power of Disruptive Technologies, O’Reilly, 2001•page numbers in the course slides are relative to this book unless otherwise stated–other readings as assignedClient-Serverimage from: http://www.kfki.hu/~arthp/html/m/michelan/3sistina/1genesis/6adam/index.htmlIs P2P really new?•or is this really distributed computing –repackaged and repainted?–infused with new technology and social paradigms?•let’s consider some other projects before we get to a definition…History: UUCP•uucp - unix-unix copy–how email and news was copied from server-to-server in the 1980’s–http://www.uucp.org/–RFC 976•http://www.rfc-editor.org/rfc/rfc976.txtHistory: Usenet, NNTP•does anyone remember Usenet news?–decentralized control between news servers–archive at: http://www.google.com/grphp?hl=en&ie=UTF-8–Network News Transfer Protocol•RFC 977, http://www.rfc-editor.org/rfc/rfc977.txtHistory: DNS•Prior to DNS, all host names existed in /etc/hosts.txt–this was periodically copied between sites•Domain Name Servers (DNS) implemented to dynamically resolve host names to addresses–P2P flavored communication between DNS servers–take Ajay Gupta’s courses!History: xFS•Figures 1 & 2 from: Wang & Anderson, “xFS: A Wide Area Mass Storage File System”–http://citeseer.nj.nec.com/161339.htmlComputers and People•historically, num(computers) << num(users)image from: http://136.162.32.160/company/h_systems.htmlComputers and People•starting in the mid-late 1980s, num(computers) ~ num(users)image from: http://www.black-cube.net:8000/Sun/?cat=cat_1025273806&picture=image_1025302668Computers and People•starting in the mid 1990s, num(computers) > num(users)•IPv6 notwithstanding, num(computers) > num(IP addrs)images from: http://www.apple.com/What is P2P? •“Peer-to-peer is a class of applications that takes advantage of resources -- storage, cycles, content, human presence -- available at the edges of the Internet.”–Clay Shirky, p. 22What is P2P?•“Because accessing these decentralized resources means operating in an environment of unstable connectivity and unpredictable IP addresses, peer-to-peer nodes must operate outside DNS and have significant or total autonomy from central servers”–Clay Shirky, p. 22P2P Litmus Test (p. 22)1. Does it allow for variable connectivity and temporary network addresses?2. Does it give the nodes at the edges of the network significant autonomy?Client-Serverquery=DJ Shadowquery=Veloce query=Thievery Corporation query=RadioheadHybrid: C-S & P2Pquery=Yo La Tengoquery=Spinal TapFull P2Pquery=Cut Chemist query=Cut Chemist query=Cut ChemistContent at the Edges•Content at the center…–model for most client-server applications–including most DLs•Content at the edges–content == “stuff” on your own personal PC, laptop, PDA, etc.Better? Or Just Different?•What scenarios would you prefer C-S?•What scenarios would you prefer P2P?•Hybrids?–cf. caching, work flow, preservation, provenance, authenticity…•take my “Intro to Digital Libraries” course!Does Content Influence Design?Which can you personally verify? Which would you accept from a stranger?“Users reward simplicity” (p. 27)•HTML was widely adopted not because it was the best approach … –in fact, there were/are many “better” approaches•…rather, HTML was adopted because it was simple–“follow the users”•great theory does nothing if it is not adopted…P2P - the cure for what ails you?•“decentralization is a tool, not a goal” (p. 28)•if P2P doesn’t make sense for a project, don’t use it…•the “disruptive technology” will be assimilated, as are all disruptive technologies“The P in P2P is People” (p. 49)•Understanding how people operate is critical to knowing how P2P systems operate–P2P is, after all, a reflection of people’s desire to communicate with each other–social networks…It’s a Small World•Stanley Milgram’s “small world” experiment –Stanley Milgram: The Small World Problem, Psychology Today 1(1), 60-67 (1967) –range = 2-10, median = 5•http://backspaces.net/PLaw/•Oracle of Bacon–http://www.cs.virginia.edu/oracle/Graphs of Social Networksimage from: http://backspaces.net/PLaw/citation of original work: Duncan J. Watts, Steven H. Strogatz. Collective Dynamics of 'Small-World' Networks. Nature 393, 440-442 (1998)Zipf’s Law •Pk ~ Ck-a where1. Pk = frequency of kth ranked item; 2. C = a corpus-wide constant;3. a ~ 1–Zipf was a linguist who sought to describe the frequency of words in language•http://linkage.rockefeller.edu/wli/zipf/Zipf’s Lawfrom: