Peer to Peer networks and Performance Hanoch Levy (hanoch at cs.tau.ac.il) Office: Kaplun 511 Office hours: by appointment 14.03.2016 H. Levy P2P+Performance 1 Course Information Lectures: Wed 9-12 Shreiber 8 Web site: http://www.cs.tau.ac.il/~hanoch/ And follow the link: Advanced topics in Computer-Networks – Peer-to-Peer Networks 14.03.2016 H. Levy P2P+Performance 2/23 Course Objective 1. Study P2P networks 2. Open the door for research in this area (Master/ Ph.D) 3. Cover both: Theoretical background and advanced material 14.03.2016 H. Levy P2P+Performance 3/23 Course Approach Prerequisites: Course in computer Networks/ Equivalent / teacher approval Requirements: Most material is New (articles) Requires some analytic models Approach: Theoretical background (performance: elementary queueing theory) – frontal lectures by teacher. Recent material: (articles) – lectures by students 14.03.2016 H. Levy P2P+Performance 4/23 Course Structure and requirements First half of semester: Teacher lectures 2nd half of semester – student lectures. Each student gets one paper – and covers one paper in 45 minutes. 14.03.2016 H. Levy P2P+Performance 5/23 Course requirements First half (theoretical): In Class Exam (5060% of grade) 2nd Half: Give a good lecture. (40-50% of grade) exam time: see Mazkirut HW assignments Active participation will benefit a bonus Overflow students (if any): lecture can be replaced by special assignment by lecturer. 14.03.2016 H. Levy P2P+Performance 6/23 Course Information Supporting Books (theoretical part) 1. Queueing Theory, L. Kleinrock, Vol I, Wiley (hard copy) 2. Online books: 1. Web site: http://web2.uwindsor.ca/math/hlynka/qonline.html (search “queueing hlynka) 2. Introduction to Queueing Theory (2nd edition). Robert B. Cooper. 1981. 347 pp. This classic book is available on line through Robert Cooper's home page. The link to the book is: http://www.cse.fau.edu/~bob/publications/IntroToQueueingTheory_ Cooper.pdf The solution manual (by Borge Tolt, 182 pages, 1981) is available online at http://www.cse.fau.edu/%7Ebob/publications/QueueingTheory_soln s.pdf 14.03.2016 H. Levy P2P+Performance 7/23 Course Information Supporting Books (theoretical part – cont ) 3. Queueing Theory. Ivo Adan and Jacques Resing. 2001. 180 pp. 4. Queues: –A Course in Queueing Theor. Moshe Haviv. October 1, 2009. New and complete. 5. more… 14.03.2016 H. Levy P2P+Performance 8/23 Motivation Last 10-15 years: communications revolution Internet + Computer communications Is a key factor of the Information revolution Implications A drastic change of some aspects of life Revolution is affected by life Technology drives applications Applications drive technology 14.03.2016 H. Levy P2P+Performance 9/23 Motivation (cont) Applications / technology / research rapidly change over time If want to stay in frontier: => Research material => Course material 14.03.2016 very dynamic very dynamic H. Levy P2P+Performance 10/23 Objectives Computer Networking course: Internet infra-structure 1 Introduction and Layering 2 Physical Layer, Data Link Layer, MAC Protocols 3 Hubs, Bridges, SwitchesData Link Layer 4 Switching UnitsSTP, Switching Fabric 5 Scheduling: Buffer Management Scheduling, WFQ example 6 Network Layer: RoutingRouting 7 Reliable Data TransferIP 8 End to End ProtocolsARQ 10 Flow Control, Congestion ControlTCP flow & congestion control 11 Network SecurityNetwork Sniffing (no slides) 12 DNS, HTTPTCP (state chart) 13 DDoS ALL – operations of network of networks. 14.03.2016 H. Levy P2P+Performance 11/23 Objectives (2) Advanced Material – network development following technology Peer to Peer (P2P): Bittorent, Skype Songs /movies / video-on-demand/video online 14.03.2016 H. Levy P2P+Performance 12/23 Internet Physical Infrastructure Residential access Cable Fiber DSL Wireless ISP Backbone ISP ISP The Internet is a network Campus access, e.g., Ethernet Wireless 14.03.2016 of networks Each individually administrated network is called an Autonomous System (AS) H. Levy P2P+Performance 13/23 13 Data Networks Set of interconnected nodes exchange information sharing of the transmission circuits= "switching". many links allow more than one path between every 2 nodes. network must select an appropriate path for each required connection. 14.03.2016 H. Levy P2P+Performance 14/23 Real Network 14.03.2016 H. Levy P2P+Performance 15/23 Peer to Peer – what is it? “Historical” Internet : send data from A to K. Client-server model: 14.03.2016 A = server = data source K = client data consumer If C wants too – get from A (unicast or broadcast) H. Levy P2P+Performance 16/23 Peer to Peer – what is it 14.03.2016 A (source) sends to K. K (client) may become now a server. K sends to C (another client). H. Levy P2P+Performance 17/23 Peer to Peer – what the diff? 14.03.2016 A K Server Client (user) Commercial Costs Private “Cost free” Charges “contributes” Legal obligations Legal?? H. Levy P2P+Performance 18/23 Peer to Peer – what the diff? A K Small number Huge number Reliable Non reliable planned unplanned Huge traffic 80%!!!! 14.03.2016 H. Levy P2P+Performance 19/23 Peer to Peer – How important • 0% of costs • O(0)% of revenues • 0% of planning • “Nothing” • BUT: 80% of traffic • cannot disregard… • If you can’t beat them, join them… 14.03.2016 H. Levy P2P+Performance 20/23 Peer to Peer – Historical View • Networks developed for 40+ years • Internet – started developing late 70’s early 80’s • Distributed, Semi Organized • ATM – developed throughout the 90’s • huge amount of money!!! • Very well organized network • “failed” • P2P – started in the 00’s • VERY unorganized • 80% of traffic (though SMALL % of money) 14.03.2016 H. Levy P2P+Performance 21/23 Peer to Peer – WHY?? Legal (this is how it started…) Broadcast is not really implemented A is bottleneck Resource Utilization: K is idle X% (95?)of the day Scalability Issues: 14.03.2016 Communications (costs!!) CPU BW cost? Free ride? Files? Video on demand? Stream (video Broadcast) H. Levy P2P+Performance 22/23 Questions for this course Is it viable? Does it have life of itself? Can it survive without economy? Can you carry 80% of traffic without make charges? Are users (clients) going to let it go? Are ISP’s going to be happy? Can 80% go unnoticeable without economic bodies “join the party” E.g – will ISP take advantage of it? Is it the right way to plan a net? 14.03.2016 H. Levy P2P+Performance 23/23 Theory Many network models – based on stochastic modeling Queueing systems Stochastic processes P2P models – included Objective: study basic stochastic / queueing models Elementary Queueing theory 14.03.2016 H. Levy P2P+Performance 24/23