Introduction: Part-11 CSCI 6781 Advanced Computer Networks Ram Dantu

Introduction: Part-12 Course Overview

Introduction: Part-13 Syllabus This class IS about... Recent advances in Computer Networks New developments in Physical Layer, Layer 2 (ATM/MPLS), Layer 3 (mobile IP, ForCES), Layer 4 (SCTP) and application layers like SIP, MIDCOM, MobileIP, etc. Vulnerability of these protocols, threat models Wireless networks and handoff issues Working together and learning together ….

Introduction: Part-14 Background r My background r Your background, thesis topic and why are you talking this course m Need some course to finish the degree m Want to be network security engineer m Want to be a hacker r Check your often (atleast twice a day) r Office Hours and appointment (available most of the week days, get an appointment)

Introduction: Part-15 Grading r See website

Introduction: Part-16 Issues in Internet r Scalability r QoS r Security r Availability

Introduction: Part-17 MPLS, Why r Full mesh of routers leads to O(NxN) flooding when a link goes down. This is way of expressing scalability of flooding mechanism. r Decoupling routing and forwarding r Better integration of the IP and ATM worlds r Basis for building next generation network applications and services such as provider-provided VPNs r Traffic engineering m Optimizing network utilization m Handling unexpected congestion m Handling link and node failures r Quality of Service r Any Transport over MPLS. MPLS supports applications that facilitates carrying Layer 2 traffic, such as Frame Relay, Ethernet and ATM over MPLS cloud.

Introduction: Part-18 Why SCTP SCTP preserves message boundaries TCP is byte-oriented. Applications must add their own record marking to delineate messages. concept of chunks security against SYN flooding attack multi-homing multi-streaming message fragmentation and bundling congestion control

Introduction: Part-19 Why SCTP SCTP PDU = 1 common header + 1 or more chunks ( control or data) Association setup = 4 way handshake (INIT, INIT-ACK, COOKIE-ECHO, COOKIE_ACK) COOKIE mechanism to prevent SYN flooding attack Graceful shutdown(SHUTDOWN, SHUTDOWN-ACK, SHUTDOWN-COMPLETE) no half-close as in TCP Separates reliability from ordered delivery Preserves message boundaries SACK chunks to ack cumulative TSN + gap ack blocks + duplicate TSNs Achieves link / path redundancy by supporting multi-homed hosts along with reachability check

Introduction: Part-110 Forces r Forwarding and Control separation r Scalability r Interoperability between network processors r Managing Middleboxes

Introduction: Part-111 Voice Over IP r lower costs per call, especially for long-distance calls  lower infrastructure costs: once IP infrastructure is installed, no or little additional telephony infrastructure is needed. Note that voice over IP traffic does not necessarily have to travel over the public Internet: it may also be deployed on private IP networks. r End user can manage services r Setup and rerouting r Voice and data convergence r Maintainability

Introduction: Part-112 What’s the Internet: “nuts and bolts” view r millions of connected computing devices: hosts, end-systems m pc’s workstations, servers m PDA’s phones, toasters running network apps r communication links m fiber, copper, radio, satellite r routers: forward packets (chunks) of data thru network local ISP company network regional ISP router workstation server mobile

Introduction: Part-113 What’s the Internet: “nuts and bolts” view r protocols: control sending, receiving of msgs m e.g., TCP, IP, HTTP, FTP, PPP r Internet: “network of networks” m loosely hierarchical m public Internet versus private intranet r Internet standards m RFC: Request for comments m IETF: Internet Engineering Task Force local ISP company network regional ISP router workstation server mobile

Introduction: Part-114 What’s the Internet: a service view r communication infrastructure enables distributed applications: m WWW, , games, e- commerce, database., voting, m more? r communication services provided: m connectionless m connection-oriented r cyberspace [Gibson]: “a consensual hallucination experienced daily by billions of operators, in every nation,...."

Introduction: Part-115 Perspective r Network users: Does the network support the users’ applications m Reliability m Error free service m Speed of data transfer r Network designers: Cost efficient network design m Good utilization of network resources m Cost of building the network m Types of services to be supported m Security

Introduction: Part-116 Perspective r Network providers: Network administration and customer service m Maximize Revenue m Minimize Operations Expenses m Survivability and Resiliency (Why) Discuss the difference! Give examples

Introduction: Part-117 A closer look at network structure: r network edge: applications and hosts r network core: m routers m network of networks r access networks, physical media: communication links

Introduction: Part-118 Building Blocks r Switches, Routers, Gateways m Special network components responsible for “moving” packets across the network from source to destination. r Network hosts, workstations, etc. m they generally represent the source and sink (destination) of data traffic (packets) r We can recursively build large networks by interconnecting networks via gateways and routers.

Introduction: Part-119 An Interconnection Network

Introduction: Part-120 The network edge: r end systems (hosts): m run application programs m e.g., WWW, m at “edge of network” r client/server model m client host requests, receives service from server m e.g., WWW client (browser)/ server; client/server r peer-peer model: m host interaction symmetric m e.g.: teleconferencing

Introduction: Part-121 Network edge: connection-oriented service Goal: data transfer between end sys. r handshaking: setup (prepare for) data transfer ahead of time m Hello, hello back human protocol m set up “state” in two communicating hosts r TCP - Transmission Control Protocol m Internet’s connection- oriented service TCP service [RFC 793] r reliable, in-order byte- stream data transfer m loss: acknowledgements and retransmissions r flow control: m sender won’t overwhelm receiver r congestion control: m senders “slow down sending rate” when network congested

Introduction: Part-122 Network edge: connectionless service Goal: data transfer between end systems m same as before! r UDP - User Datagram Protocol [RFC 768]: Internet’s connectionless service m unreliable data transfer m no flow control m no congestion control App’s using TCP: r HTTP (WWW), FTP (file transfer), Telnet (remote login), SMTP ( ) App’s using UDP: r streaming media, teleconferencing, Internet telephony

Introduction: Part-123 The Network Core r mesh of interconnected routers r the fundamental question: how is data transferred through net? m circuit switching: dedicated circuit per call: telephone net m packet-switching: data sent thru net in discrete “chunks”

Introduction: Part-124 Different Types of Switching r Different Types of Switching: m Circuit Switching (telephone network) dedicated circuit, sending and receiving bit streams m Message Switching m Packet Switching store and forward, sending and receiving packets m Virtual Circuit Switching m Cell Switching (ATM) r What are Packets? m Data to be transmitted is divided into discrete blocks

Introduction: Part-125 Network Core: Circuit Switching End-end resources reserved for “call” r link bandwidth, switch capacity r dedicated resources: no sharing r circuit-like (guaranteed) performance r call setup required

Introduction: Part-126 Cost-Effective Resource Sharing r Must share (multiplex) network resources among multiple users. r Common Multiplexing Strategies m Time-Division Multiplexing (TDM) m Synchronous TDM (STDM) m Frequency-Division Multiplexing (FDM) r Multiplexing multiple logical flows over a single physical link.

Introduction: Part-127 Network Core: Circuit Switching network resources (e.g., bandwidth) divided into “pieces” r pieces allocated to calls r resource piece idle if not used by owning call (no sharing) r dividing link bandwidth into “pieces” m frequency division m time division

Introduction: Part-128 Network Core: Packet Switching each end-end data stream divided into packets r user A, B packets share network resources r each packet uses full link bandwidth r resources used as needed, resource contention: r aggregate resource demand can exceed amount available r congestion: packets queue, wait for link use r store and forward: packets move one hop at a time m transmit over link m wait turn at next link Bandwidth division into “pieces” Dedicated allocation Resource reservation

Introduction: Part-129 Network Core: Packet Switching Packet-switching versus circuit switching: human restaurant analogy r other human analogies? A B C 10 Mbs Ethernet 1.5 Mbs 45 Mbs D E statistical multiplexing queue of packets waiting for output link

Introduction: Part-130 Network Core: Packet Switching Packet-switching: store and forward behavior

Introduction: Part-131 Packet switching versus circuit switching r 1 Mbit link r each user: m 100Kbps when “active” m active 10% of time r circuit-switching: m 10 users r packet switching: m with 35 users, probability > 10 active less that.004 Packet switching allows more users to use network! N users 1 Mbps link

Introduction: Part-132 Packet switching versus circuit switching r Great for bursty data m resource sharing m no call setup r Excessive congestion: packet delay and loss m protocols needed for reliable data transfer, congestion control r Q: How to provide circuit-like behavior? m bandwidth guarantees needed for audio/video apps still an unsolved problem (chapter 6) Is packet switching a “slam dunk winner?”

Introduction: Part-133 Packet-switched networks: routing r Goal: move packets among routers from source to destination m we’ll study several path selection algorithms (chapter 4) r datagram network: m destination address determines next hop m routes may change during session m analogy: driving, asking directions r virtual circuit network: m each packet carries tag (virtual circuit ID), tag determines next hop m fixed path determined at call setup time, remains fixed thru call m routers maintain per-call state

Introduction: Part-134 Access networks and physical media Q: How to connect end systems to edge router? r residential access nets r institutional access networks (school, company) r mobile access networks Keep in mind: r bandwidth (bits per second) of access network? r shared or dedicated?

Introduction: Part-135 Residential access: point to point access r Dialup via modem m up to 56Kbps direct access to router (conceptually) r ISDN: intergrated services digital network: 128Kbps all- digital connect to router r ADSL: asymmetric digital subscriber line m up to 1 Mbps home-to-router m up to 8 Mbps router-to-home m ADSL deployment: UPDATE THIS

Introduction: Part-136 Residential access: cable modems r HFC: hybrid fiber coax m asymmetric: up to 10Mbps upstream, 1 Mbps downstream r network of cable and fiber attaches homes to ISP router m shared access to router among home m issues: congestion, dimensioning r deployment: available via cable companies, e.g., MediaOne

Introduction: Part-137 Institutional access: local area networks r company/univ local area network (LAN) connects end system to edge router r Ethernet: m shared or dedicated cable connects end system and router m 10 Mbs, 100Mbps, Gigabit Ethernet r deployment: institutions, home LANs soon r LANs: chapter 5

Introduction: Part-138 Wireless access networks r shared wireless access network connects end system to router r wireless LANs: m radio spectrum replaces wire m e.g., Lucent Wavelan 10 Mbps r wider-area wireless access m CDPD: wireless access to ISP router via cellular network base station mobile hosts router

Introduction: Part-139 Physical Media r physical link: transmitted data bit propagates across link r guided media: m signals propagate in solid media: copper, fiber r unguided media: m signals propagate freelye.g., radio Twisted Pair (TP) r two insulated copper wires m Category 3: traditional phone wires, 10 Mbps ethernet m Category 5 TP: 100Mbps ethernet

Introduction: Part-140 Physical Media: coax, fiber Coaxial cable: r wire (signal carrier) within a wire (shield) m baseband: single channel on cable m broadband: multiple channel on cable r bidirectional r common use in 10Mbs Ethernet Fiber optic cable: r glass fiber carrying light pulses r high-speed operation: m 100Mbps Ethernet m high-speed point-to-point transmission (e.g., 5 Gps) r low error rate

Introduction: Part-141 Physical media: radio r signal carried in electromagnetic spectrum r no physical “wire” r bidirectional r propagation environment effects: m reflection m obstruction by objects m interference Radio link types: r microwave m e.g. up to 45 Mbps channels r LAN (e.g., waveLAN) m 2Mbps, 11Mbps r wide-area (e.g., cellular) m e.g. CDPD, 10’s Kbps r satellite m up to 50Mbps channel (or multiple smaller channels) m 270 Msec end-end delay m geosynchronous versus LEOS

Introduction: Part-142 Different Types of Links Sometimes you install your own!

Introduction: Part-143 Bigger Pipes! Sometimes leased from the phone company STS: Synchronous Transport Signal

Introduction: Part-144 Delay in packet-switched networks packets experience delay on end-to-end path r four sources of delay at each hop r nodal processing: m check bit errors m determine output link r queueing m time waiting at output link for transmission m depends on congestion level of router A B propagation transmission nodal processing queueing

Introduction: Part-145 Delay in packet-switched networks Transmission delay: r R=link bandwidth (bps) r L=packet length (bits) r time to send bits into link = L/R Propagation delay: r d = length of physical link r s = propagation speed in medium (~2x10 8 m/sec) r propagation delay = d/s A B propagation transmission nodal processing queueing Note: s and R are very different quantitites!

Introduction: Part-146 Bandwidth (throughput) r Amount of data that can be transmitted per time unit r Example: 10Mbps r link versus end-to-end r Notation m KB = 2 10 bytes m Mbps = 10 6 bits per second Performance

Introduction: Part-147 r Bandwidth related to “bit width” Performance a) b) 1 second

Introduction: Part-148 Latency (delay) r Time it takes to send message from point A to point B r Example: 24 milliseconds (ms) r Sometimes interested in in round-trip time (RTT) r Components of latency Latency = Propagation + Transmit + Queue + Proc. Propagation = Distance / SpeedOfLight Transmit = Size / Bandwidth

Introduction: Part-149 The “hard” limit r Speed of light m 3.0 x 10 8 meters/second in a vacuum m 2.3 x 10 8 meters/second in a cable m 2.0 x 10 8 meters/second in a fiber r Notes m no queuing delays in direct link m bandwidth not relevant if Size = 1 bit m process-to-process latency includes software overhead m software overhead can dominate when Distance is small

Introduction: Part-150 r Relative importance of bandwidth and latency m small message (e.g., 1 byte): 1ms vs. 100ms dominates 1Mbps vs. 100Mbps m large message (e.g., 25 MB): 1Mbps vs. 100Mbps dominates 1ms vs. 100ms r Consider two channels of 1Mbps and 100 Mbps respectively. For a 1 byte message, the available bandwidth is relatively insignificant given a RTT of 1 ms. The transmit delay for each channel is 8  s and 0.08  s, respectively.

Introduction: Part-151  Delay x Bandwidth Product e.g., 100ms RTT and 45Mbps Bandwidth = 560KB of data r We have to view the network as a buffer. This may have interesting consequences: m How much data did the sender transmit before a response can be received? Bandwidth Delay

Introduction: Part-152 r Application Needs m bandwidth requirements: burst versus peak rate m jitter: variance in latency (inter-packet gap) r Average Bandwidth Requirement is Not enough: m consider a source with an avg. BW-requirement of 2Mbps. If the application generates 1 Mbit in one second interval and 3Mbit in a second, a channel that can support 2 Mbps max. will have a tough time. r Other Quality of Service (QOS) Parameters: m max. and min. delay m max. and min. bandwidth demand m rates for dynamic increase of demands m Cell-Loss Rate

Introduction: Part-153 Queueing delay (revisited) r R=link bandwidth (bps) r L=packet length (bits) r a=average packet arrival rate traffic intensity = La/R r La/R ~ 0: average queueing delay small r La/R -> 1: delays become large r La/R > 1: more “work” arriving than can be serviced, average delay infinite!

Introduction: Part-154 What Goes Wrong in the Network? r Different types of Error: m Bit-level errors (electrical interference) 1 in in copper - 1 in in fiber m Packet-level errors (congestion) m Link and node failures r How should we deal with these types of error? r What are the consequences of errors?

Introduction: Part-155 Other types of problems in the Network? r Messages are delayed r Messages are deliver out-of-order r Third parties eavesdrop The key problem is to fill in the gap between what applications expect and what the underlying technology provides.