1. CSE 5346 – Networks II: High Performance Networks and Quality of Service Spring 2016 Mike O’Dell Senior Lecturer 1

2. Introduction Performance and Quality of Service 2

3. CSE 5346 – Computer Networks II Course Objective: Understand the concepts critical to quality of service and understand how the Internet supports advanced classes of service. Focus on topics such as QoS architectures, queue management, traffic modeling, and relevant quality of service protocols deployed in IP networks. Understand current QoS practices such as Flow Aware Networking. Introduction 3

4. CSE 5346 – Computer Networks II Instructor: Mike O’Dell (odell@cse.uta.edu) Instructor: Mike O’Dell (odell@cse.uta.edu)odell@cse.uta.edu GTA: Sharad Velmajala (sharad.velmajala@mavs.uta.edu) GTA: Sharad Velmajala (sharad.velmajala@mavs.uta.edu)sharad.velmajala@mavs.uta.edu Class Web Site: http://ranger.uta.edu/~odell/ Class Web Site: http://ranger.uta.edu/~odell/http://ranger.uta.edu/~odell/ Textbooks: Textbooks: –General topics: High-Speed Networks and Internets - Performance and Quality of Service, Second Edition, William Stallings, Prentice Hall, 2002 (ISBN: 0-13- 032221-0) –Advanced topics: Guide to Flow-Aware Networking: Quality of Service Architectures and Techniques for Traffic Management, Second Edition, William Stallings, Prentice Hall, 2002 (ISBN: 0-13-032221-0) Introduction 4

5. Course Foundation Assumptions (pre-requisites) Good understanding of packet-switched networking concepts and principles of operation Good understanding of packet-switched networking concepts and principles of operation Good understanding of Internet protocols and architectures (e.g., IP protocol stack) Good understanding of Internet protocols and architectures (e.g., IP protocol stack) Solid foundation in computer operating systems fundamentals Solid foundation in computer operating systems fundamentals Ability to perform independent research, analyze findings and document results Ability to perform independent research, analyze findings and document results Introduction 5

6. Class Administration & Policy Web Site: http://ranger.uta.edu/~odell/ Web Site: http://ranger.uta.edu/~odell/http://ranger.uta.edu/~odell/ –Schedule, Syllabus, Class Materials/Information Email - will be used for time-critical info Email - will be used for time-critical info –I will use ONLY UTA email (mavmail) account for correspondence Schedule Schedule –Ambitious... may be modified... check web site frequently Attendance… expected, not measured Attendance… expected, not measured Late Submission and Make-Up Policy Late Submission and Make-Up Policy –Late Homework, Programs/Projects: 10%/day, max of 40%, then zero –Quizzes and Exams: NO make-ups. NO early exams/quizzes. Absence = zero grade. Introduction 6

7. Class Administration & Policy Grading Policy Grading Policy –Project (4 stages)30% –Quizzes (4)20% –Exam 120% –Exam 225% Final Grade Assignment (typical-guideline only) Final Grade Assignment (typical-guideline only) –Based on final numerical score out of 100% possible: A > 85% B85% – 75% C75% - 65% D65% - 55% F55% & below –Generally final grades will not be “curved”. Introduction 7

8. Class Administration & Policy Honesty… expected, dishonesty will not be tolerated Honesty… expected, dishonesty will not be tolerated –Discussions, brainstorming are encouraged, HOWEVER –Homework, Final Paper & Programming Assignments, Quizzes, Exams, etc. are to be YOUR individual work –See the syllabus and UTA references for definitions of Cheating Collusion Plagiarism Any suspected violation of this policy will be referred to the UTA Office of Student Conduct for disposition. Any suspected violation of this policy will be referred to the UTA Office of Student Conduct for disposition. Introduction 8

9. Class Administration & Policy Office Hours Office Hours –Individual grades or questions on grading of individual quizzes, exams, etc. are discussed only during office hours (i.e. NOT at the end of class period) –Mr. O’Dell’s Office Hours (in ERB 647) MWF: 12:30 PM – 1:30 PM –Mr. Velmajala’s Office Hours (in TBD) TBD Introduction 9

10. What will we cover? Networking Introduction/Review Networking Introduction/Review –This week… setting the stage Performance Modeling and Estimation Performance Modeling and Estimation Congestion, Network Traffic Analysis/Management and Routing Congestion, Network Traffic Analysis/Management and Routing Quality of Service Models in the Internet Quality of Service Models in the Internet –Queuing models, traffic and flow management –QoS Foundation: Intserv and Diffserv Flow Aware Networking Flow Aware Networking – Architectures, Techniques, Considerations Introduction 10 Some Review New Content

11. How will we cover it? Classroom lectures Classroom lectures Quizzes(4) and exams (2) Quizzes(4) and exams (2) –NO final exam Incremental network simulation project Incremental network simulation project –“DIY Simulator” Self-study research, and simulation analysis by students Self-study research, and simulation analysis by students Introduction 11

12. Introduction An Overview of Networking – Traffic Management and Quality of Service 12

13. What is an/the Internet? connected computing devices: hosts, end- systems connected computing devices: hosts, end- systems –PC’s, workstations, servers –PDA’s, phones, toasters, cars running network applications communication links communication links –fiber, copper, radio, satellite routers/switches: forward packets (chunks) of data thru network routers/switches: forward packets (chunks) of data thru network Introduction 13 router workstation server mobile local ISP companynetwork regional ISP

14. The Need for Speed! Scale Scale –growing number of hosts -> growing demands on bandwidth –new technologies result in new paradigms for device and connection types e.g. ?? e.g. ?? Introduction 14 Application Application –demand for large to huge file transfers –increasing critical nature of Internet use –demand for “real- time” performance characteristics –demand for guarantees of service levels e.g. ?? e.g. ?? User Expectations! User Expectations!

15. The Need for Improved (better) Levels of Service Internet Best- Effort Service Internet Best- Effort Service –all packets treated equally –designed for elastic traffic –no guarantees of bandwidth or throughput –no guarantees of delay –no guarantee of jitter (delay variation) Introduction 15 Applications Applications –often create inelastic traffic –often sensitive to delay –often sensitive to jitter –often critical in nature –generate elastic traffic as well User Requirements! User Requirements!

16. Delay Sensitivity & Criticality Introduction 16

17. IP-based Internets (aka TCP/IP networks) Internetworking: the dominant paradigm of computer networking Evolution - key internetworking technologies: packet switching: best effort, or specified service level packet switching: best effort, or specified service level Internet protocol stack Internet protocol stack –TCP: reliable end-to-end transport –UDP: best effort datagram transport –IP: internet routing and delivery (v4…v6) dynamic routing, load balancing, traffic management, congestion avoidance dynamic routing, load balancing, traffic management, congestion avoidance Introduction 17

18. Advancements in Internet Standards and Architecture Many recent advancements are driven by the need to support streaming multimedia and real-time traffic Many recent advancements are driven by the need to support streaming multimedia and real-time traffic Evolution of the Internet’s Integrated Services Architecture (ISA, or IntServ) and Differentiated Services (DS, or DiffServ) Evolution of the Internet’s Integrated Services Architecture (ISA, or IntServ) and Differentiated Services (DS, or DiffServ)  QoS Architectures/Frameworks are driving protocol changes: –IPv6 – includes specific features for QoS –Realization of enhancements to ISA/DS architectures –Flow Aware Networking models Introduction 18

19. Delays in Packet Switched (e.g. IP) Networks End-to-end delay (simplified) = End-to-end delay (simplified) = (d prop + d trans + d queue + d proc ) x Q (d prop + d trans + d queue + d proc ) x Q Introduction 19 B A Where:  Propagation delay (d prop )  Transmission delay (d trans )  Queuing delay (d queue )  Processing delay (d proc )  Number of links (Q)

20. Introduction 20 So what’s the problem? What makes this so hard? (I.e., what are we going to focus on in this course.)

21. Delays in Packet Switched (e.g. IP) Networks End-to-end delay (simplified) = End-to-end delay (simplified) = –(d prop + d trans + d queue + d proc ) … on each link Introduction 21 Much more later B A Where:  Propagation delay (d prop ) = d/s  Transmission delay (d trans ) = L/R ?  Queuing delay (d queue ) = ? ?  Processing delay (d proc ) = ? ?  Number of links (Q) = ?

22. The Router in IP Networks An Internet router has two functions: An Internet router has two functions: –Establish routes: routing protocols –Forward packets: move from input to appropriate output Routers play THE KEY ROLE in end-to- end performance (throughput) of network applications Routers play THE KEY ROLE in end-to- end performance (throughput) of network applications –Routing: establishing paths consistent with requirements –Forwarding: queue management at each hop on the path Introduction 22

23. Basic IP router architecture Two key router functions: Two key router functions: 1.run routing algorithms/protocol (RIP, OSPF, BGP) 2.forward datagrams from incoming link (queue) to outgoing link (queue) Introduction 23 high-speed switching fabric routing processor router input ports router output ports forwarding data plane (hardware – nanosecond speed) routing, management control plane (software – multi-millisecond speed) forwarding tables computed, pushed to input ports Input links Output links

24. Input port functions decentralized switching: given datagram destination, lookup output port using the forwarding table in input port memory (“match plus action”) given datagram destination, lookup output port using the forwarding table in input port memory (“match plus action”) goal: complete input port processing at ‘line speed’ goal: complete input port processing at ‘line speed’ queuing: if datagrams arrive faster than forwarding rate into switch fabric queuing: if datagrams arrive faster than forwarding rate into switch fabric Introduction 24 line termination link layer protocol (receive) lookup, forwarding queueing physical layer: bit-level reception data link layer: e.g., Ethernet switch fabric (bus, memory, etc.)

25. Input port queuing  fabric slower than input ports combined -> queueing may occur at input queues  queueing delay and loss due to input buffer overflow!  Head-of-the-Line (HOL) blocking: queued datagram at front of queue prevents others in queue from moving forward Introduction 25 output port contention: only one red datagram can be transferred. lower red packet is blocked switch fabric one packet time later: green packet experiences HOL blocking switch fabric

26. Introduction 26 Output port functions  buffering required when datagrams arrive from fabric faster than the transmission rate  scheduling discipline chooses among queued datagrams for transmission line termination link layer protocol (send) switch fabric datagram buffer(s) queueing

27. Output port queueing  buffering when arrival rate via switching fabric exceeds output line speed  queueing (delay) and loss due to output port buffer overflow! Introduction 27 at packet time t, multiple packets arrive for same output port at packet time t+1, output queue grows switch fabric switch fabric

28. Hence, our focus… We will study, primarily, the Internet’s network layer: We will study, primarily, the Internet’s network layer: –Traffic management –Queuing disciplines, queue management –Flow aware traffic policies –IP protocol and relevant standards (rfc’s) Demonstrated via a multi-phase, “build- your-own” network QoS routing simulator Demonstrated via a multi-phase, “build- your-own” network QoS routing simulator Introduction 28