Internet Protocols and Networking Fundamentals PDF

1. Definition & Overview The Internet is a global computer network connecting billions of devices. It consists of hardware and software components that enable communication. The Internet can be viewed as a networking infrastructure providing services to applications. 2. Devic...

1. Definition & Overview The Internet is a global computer network connecting billions of devices. It consists of hardware and software components that enable communication. The Internet can be viewed as a networking infrastructure providing services to applications. 2. Devices & End Systems Traditionally included desktops, servers, and workstations. Now includes smartphones, tablets, TVs, IoT devices (thermostats, cars, home appliances, etc.). These devices are called hosts or end systems in network terminology. 3. Network Components End systems are connected through communication links and packet switches. Communication links use coaxial cable, fiber optics, copper wire, or wireless signals. Data transmission rate is measured in bits per second (bps). Data is broken into packets, sent through the network, and reassembled at the destination. 4. Packet Switching & Routing Packet switches forward packets across the network. Two main types of switches: Routers (used in the network core). Link-layer switches (used in access networks). The path taken by packets from sender to receiver is called a route or path 5. Internet as a Transportation Network Analogy Packets = Trucks transporting goods. Communication links = Highways & roads. Packet switches = Intersections. End systems = Buildings (factories & warehouses). 6. Internet Service Providers (ISPs) Provide Internet access to individuals, businesses, and content providers. Offer different access types: DSL, cable, fiber, mobile networks, and WiFi. Lower-tier ISPs connect to upper-tier ISPs (e.g., AT&T, Sprint, NTT) through fiber-optic networks. 7. Internet Protocols Govern the sending and receiving of data. The two most important protocols: IP (Internet Protocol) – Defines packet structure and addressing. TCP (Transmission Control Protocol) – Ensures reliable communication. The Internet’s protocol suite is called TCP/IP. 8. Internet Standards Standardized to ensure interoperability between different networks and devices. Standards are developed by IETF (Internet Engineering Task Force). Defined in Requests for Comments (RFCs) – technical documents outlining protocols. Other standards bodies include IEEE 802 (WiFi, Ethernet standards). Services View of the Internet 1. Two Views of the Internet The Internet can be described in two ways: Nuts-and-bolts view: Hardware and software components. Services view: Infrastructure that provides services to applications. 2. Internet as a Service Provider Supports applications like: Email, web browsing, and social networks. Streaming services, online gaming, and video conferencing. Location-based recommendation systems. 3. Distributed Applications Applications involve multiple end systems exchanging data. Internet applications run on end systems, not within network switches. Packet switches facilitate data exchange but do not run applications. 4. Developing an Internet Application Programs must be written in languages like Java, C, or Python. Programs on different end systems need a way to send data to each other. Internet is a platform that facilitates this communication. 5. Socket Interface End systems use a socket interface to communicate over the Internet. The interface defines rules for sending data from one program to another. 6. Analogy with Postal Service Sending data over the Internet is like mailing a letter: A sender must follow postal service rules (addressing, stamping). Similarly, a program must follow Internet socket rules to send data. 7. Internet Services for Applications Just like postal services offer express delivery, tracking, etc., The Internet provides multiple communication services. Developers must choose the right service for their applications. PROTOCOL Definition & Importance A protocol defines the rules governing communication between two or more entities in a network. It specifies message formats, order of exchange, and actions taken upon transmission/reception. Human Analogy Communication between people follows certain rules (e.g., greetings before asking a question) Misalignment in communication styles can cause misunderstandings, similar to network protocols. Example of a Human Protocol Asking for the time requires an exchange of greetings before proceeding. If no response is received, the conversation ends, just like in network protocols. Another Example: A Classroom Scenario A teacher asks, “Any questions?” A student raises their hand (request). The teacher acknowledges (response), and the student asks their question. The teacher then answers, completing the exchange. Network Protocols Definition A network protocol governs communication between devices like computers, routers, and smartphones. Every activity involving multiple entities in a network follows a protocol. Examples Hardware protocols control data flow between network interface cards. Congestion control protocols regulate data transmission rates. Routing protocols determine packet paths across networks. Example: Web Browsing Protocol The browser sends a connection request to the web server. The server responds with a connection reply. The browser sends a GET request for a webpage. The server sends the requested webpage. Key Takeaways Protocols structure communication in networks. They ensure seamless and standardized data exchange. Understanding protocols is essential to mastering computer networking. The Network Edge Introduction Explores the components at the edge of a computer network, focusing on devices like computers and smartphones. Moves from the network edge to the network core, covering switching and routing. End Systems (Hosts) Devices connected to the Internet are called end systems (or hosts) as they sit at the edge of the network. Examples include desktops, laptops, smartphones, tablets, servers, and IoT devices. Hosts & Applications End systems run applications like web browsers, web servers, email clients, and email servers. The terms hosts and end systems are used interchangeably. Clients vs. Servers Clients: Desktop PCs, mobile devices, and other user-end machines. Servers: Powerful machines that store and distribute content like web pages, emails, and videos. Data Centers Large-scale storage and computing centers house thousands of servers. Example: Google operates 50–100 data centers, with some containing over 100,000 servers. Case History: Internet of Things (IoT) Concept of IoT Envisions a world where almost everything is wirelessly connected to the Internet. Includes people, vehicles, eyewear, watches, home sensors, hospital equipment, classrooms, and even pets. Existing IoT Devices Smartphones: Track geolocation and send data to ISPs and applications. Wearables: Smartwatches and Internet-connected glasses that can upload and share experiences in real time. Smart Home Devices: Internet-connected thermostats, body scales, and security systems controlled remotely. IoT Toys: Interactive dolls that recognize and respond to speech. Benefits of IoT Provides convenience and automation in daily life. Enhances communication, monitoring, and data collection for better decision-making. Security and Privacy Risks IoT devices can be hacked, leading to potential misuse. Attackers could hijack smart toys, compromise personal health data, or manipulate connected devices. Security threats may hinder consumer trust and limit widespread adoption. Access Networks 1. Introduction to Access Networks An access network connects an end system to the edge router, the first router in the path to another end system. Different types of access networks are used in various settings: home, enterprise, and wide-area mobile wireless networks. 2. Home Access Networks: DSL (Digital Subscriber Line) DSL provides broadband Internet via existing telephone lines (twisted-pair copper wires). DSL modem translates digital data into high-frequency tones for transmission over telephone wires. DSL Access Multiplexer (DSLAM) at the central office (CO) separates data and voice signals and connects to the ISP. Uses frequency division multiplexing to separate: High-speed downstream (50 kHz – 1 MHz). Medium-speed upstream (4 kHz – 50 kHz). Ordinary telephone channel (0 – 4 kHz). DSL is asymmetric (higher download than upload speeds). Speed depends on distance from the CO, wire gauge, and interference. 3. Home Access Networks: HFC (Hybrid Fiber-Coaxial Cable) Uses the cable TV network for Internet access. Fiber optic cables connect head-end to neighborhood junctions, and coaxial cables connect homes (serving 500-5,000 homes per junction). Requires a cable modem, which connects to the PC via an Ethernet port. Cable Modem Termination System (CMTS) at the head-end converts analog signals to digital. Downstream and upstream channels are separate but shared among users. 4. Differences Between DSL and HFC DSL provides a dedicated line per user, while HFC is a shared broadcast medium (bandwidth is divided among active users). HFC provides higher speeds but suffers from congestion during peak usage. Both DSL and HFC are asymmetric (higher download than upload speeds). 5. Challenges in HFC Networks Shared upstream and downstream channels may lead to congestion. If multiple users download large files simultaneously, speeds drop significantly. Requires a distributed multiple access protocol to avoid data collisions in the upstream channel. Home Access Networks: FTTH Fiber to the Home (FTTH) provides high-speed internet access using optical fiber. Direct fiber: Each home gets a dedicated fiber from the Central Office (CO). More commonly, a single fiber is shared by multiple homes and split near residences. Two main optical distribution network architectures: Active Optical Networks (AONs) Passive Optical Networks (PONs) PON Architecture: Verizon FiOS uses PON. Optical Network Terminator (ONT) at each home connects via fiber to a neighborhood splitter. The splitter combines multiple homes (up to 100) onto a shared fiber. The shared fiber connects to an Optical Line Terminator (OLT) at the CO. The OLT converts optical signals to electrical signals and connects to the Internet via a telco router. Performance: Potential for gigabit-speed internet access. Different speed offerings based on pricing. Average U.S. FTTH speed in 2011: 20 Mbps (higher than cable and DSL). Enterprise Access Networks: Ethernet and WiFi Local Area Network (LAN): Used in corporate, university, and home settings. Ethernet: Most common wired LAN technology. Uses twisted-pair copper wire to connect to an Ethernet switch. Users typically have 100 Mbps or 1 Gbps access; servers may have 1–10 Gbps access. WiFi (Wireless LAN): Devices (laptops, smartphones, tablets) connect to an Access Point (AP) linked to the Ethernet network. IEEE 802.11-based WiFi is common in homes, businesses, airports, and public places. Provides shared speeds of 100+ Mbps. Home Networks: Combine broadband access (cable/DSL) with WiFi for flexible connectivity. Typical setup: A wireless router connects wired/wireless devices to a broadband modem. Enables seamless connectivity for users moving around the home. Wide-Area Wireless Access: 3G and LTE Used by smartphones for internet access, social media, streaming, and messaging. Works via base stations operated by cellular providers. Coverage Differences: WiFi: Range of a few tens of meters. Cellular (3G/LTE): Range of a few tens of kilometers. 3G Networks: Provides speeds above 1 Mbps for packet-switched internet access. 4G/LTE Networks: LTE evolved from 3G and offers 10+ Mbps speeds. Some commercial LTE deployments report tens of Mbps speeds. Physical Media Network Access Technologies: HFC (Hybrid Fiber-Coaxial) uses a combination of fiber and coaxial cables. DSL and Ethernet rely on copper wire. Mobile networks use the radio spectrum. Life of a Bit A bit travels from one end system to another through various links and routers. The bit is transmitted through multiple transmitter-receiver pairs. Propagation occurs through electromagnetic waves or optical pulses across physical media. Types of Physical Media 1. Guided Media: - Signals are transmitted through a solid medium, such as: 1. Fiber-optic cable 2. Twisted-pair copper wire 3. Coaxial cable 2. Unguided Media: - Signals propagate freely in the air or space, such as: 1. Wireless LAN 2. Satellite radio spectrum Cost Considerations Material costs (copper, fiber-optic cables) are relatively low compared to labor costs for installation. Builders often install multiple types of physical media (e.g., twisted pair, fiber, coaxial) during construction to avoid future rewiring costs and to prepare for potential upgrades. Twisted-Pair Copper Wire Least expensive and most commonly used guided transmission medium. Historical use: Primarily used in telephone networks for over 100 years. Structure: Consists of two insulated copper wires, each about 1 mm thick, twisted together to reduce electrical interference. Usage: More than 99% of wired connections from telephone handsets to local telephone switches use twisted-pair copper wire. Commonly used for LANs within buildings. Data Rates: Range from 10 Mbps to 10 Gbps depending on wire thickness and distance between transmitter and receiver. Can be bundled together in a cable for protection. Coaxial Cable Construction: Composed of two concentric copper conductors, providing better insulation and shielding. High data transmission: Achieves higher data rates than twisted-pair. Common use: Widely used in cable television systems and for cable modems to provide Internet access (up to tens of Mbps). Shared Medium: Can connect multiple end systems that receive the same signal. Signal transmission: Digital signals are shifted to a frequency band and transmitted as analog signals. Fiber Optics Structure: A thin, flexible medium that conducts light pulses, where each pulse represents a bit. High Data Rates: Supports tens to hundreds of gigabits per second. Advantages: Immune to electromagnetic interference. Very low signal attenuation (effective up to 100 km). Very hard to tap, ensuring security. Usage: Preferred for longhaul transmission, including international links and the backbone of the Internet. Cost: High cost of optical devices (transmitters, receivers, switches) limits use for short-haul transport like LANs or residential access networks. Link Speeds: Standards range from 51.8 Mbps to 39.8 Gbps, referred to as OC-n (e.g., OC-1, OC-3, OC-12, OC-24, OC-48, OC-192). Terrestrial Radio Channels Electromagnetic spectrum: Radio channels carry signals without needing physical wires, providing mobility and long-distance connectivity. Advantages: No need for physical installation. Can penetrate walls. Provides connectivity to mobile users. Potential for long-distance signal transmission. Channel characteristics: Dependent on propagation environment and distance. Path loss: Signal strength decreases over distance. Shadow fading: Signal strength decreases around/through obstacles. Multipath fading: Signal reflection causes distortion. Interference: External transmissions can cause signal disruption. Classification: Short distance: Devices like wireless headsets (1-2 meters). Local area: Wireless LANs (10 meters to a few hundred meters). Wide area: Cellular networks (tens of kilometers). Satellite Radio Channels Satellite communication: Uses Earth-based microwave transmitters/receivers (ground stations) linked via satellites. Satellite receives and regenerates signals before transmitting them on different frequencies. Types of Satellites: Geostationary Satellites: Positioned 36,000 km above Earth. Remain above the same spot on Earth. Introduces a 270 ms signal propagation delay due to the large distance. LEO (Low Earth Orbit) Satellites: Positioned much closer to Earth. Do not remain fixed above one spot (rotate around Earth). Can communicate with each other and ground stations. Require multiple satellites for continuous coverage. Satellite Link Speed: Can operate at hundreds of Mbps. LEO Satellites: Development of low-altitude systems for future Internet access. The Network Core The network core is the mesh of packet switches and links that interconnects the Internet's end systems. Packet Switching End systems exchange messages that can contain data or control information. Messages are broken into smaller chunks called packets. Packets travel through communication links and packet switches (routers and link layer switches). The transmission time for a packet is calculated as Time=L/R​, where: L= packet size (bits) R = transmission rate (bits per second) Store-and-Forward Transmission Store-and-forward: The packet switch must receive the entire packet before transmitting it onto the outbound link. Example: If a router receives a packet of size L bits from the source, it stores the packet until it has received the whole packet. After receiving the entire packet, the router starts transmitting it to the destination. Transmission time: For a single packet, the delay is L/R​seconds to send from source to router. After the router receives the packet, the delay to send from router to destination is also L/R​seconds. Total delay: Total delay=2L/R​. General Case for Multiple Link s For a path with N links and N−1 routers, the total end-to-end delay is: End-to-end delay = dend-to-end=NL/R​ where N is the number of links and routers, and R is the transmission rate of each link.  Output Buffers: Each packet switch has output buffers for each Queuing Delays and Packet Loss attached link, storing packets before transmission. Queuing Delay: If a packet arrives when the link is busy, it must wait in the output buffer, causing delays. Variable Delays: Queuing delays depend on network congestion. Packet Loss: If the buffer is full, an arriving packet or an already queued packet will be dropped. Packet Switching: Queuing Delay and Loss Example Scenario: Hosts A and B send packets to Host E through a router. The router has a 100 Mbps link from A/B and a 15 Mbps output link to E. Congestion: If the incoming packet rate exceeds the output link’s capacity (15 Mbps), packets queue in the router’s buffer. Forwarding Tables and Routing Protocols IP Addresses: Every end system has an IP address, which is used for routing packets. Hierarchical Structure: The IP address has a hierarchical structure, similar to postal addresses. Router Function: A router forwards packets by examining the destination IP address and using a forwarding table to find the appropriate outbound link. Forwarding Table: Maps destination addresses to outbound links. Routers consult this table to determine where to send a packet. Routing Protocols Automatic Configuration: Forwarding tables are set automatically using routing protocols, not manually configured in each router. Routing Protocols: Determine the best path for data, often using algorithms to find the shortest path to each destination. Circuit Switching Traditional Use: Common in telephone networks. Connection Establishment: Before data transmission, a dedicated connection (circuit) is established between the sender and receiver. Connection State: Switches maintain a connection state during the communication. Reserved Transmission Rate: A specific transmission rate is reserved for the connection, ensuring a constant data transfer rate. Guaranteed Data Transfer: The sender can transfer data at the reserved, constant rate throughout the connection. Alternative Core: Circuit Switching Network Network Example: The network consists of circuit switches interconnected by links, with each link supporting multiple circuits. Dedicated End-to-End Connection: A dedicated circuit is reserved for the communication between two hosts. Link Reservation: For communication, the network reserves circuits on each link, dividing the link’s transmission capacity among active connections. Example: If each link has a transmission rate of 1 Mbps and four circuits, each circuit gets 250 kbps. Circuit Switching vs. Packet Switching Packet Switching: Unlike circuit switching, no resources are reserved for packet transmission. Each packet is transmitted without guaranteed bandwidth. Congestion: If a link is congested due to other packets, the packet may need to wait in a buffer, resulting in delays. Best-Effort Delivery: The Internet aims to deliver packets promptly but does not provide guarantees on delivery or timing. Multiplexing in Circuit-Switched Networks Multiplexing Methods: Frequency-Division Multiplexing (FDM): Divides the frequency spectrum of a link into bands, assigning each connection a dedicated frequency band. Time-Division Multiplexing (TDM): Divides time into fixed-duration frames and assigns each connection a time slot in each frame. Frequency-Division Multiplexing (FDM): Each connection gets a dedicated frequency band for the duration of the connection. Example: In telephone networks, each band typically has a width of 4 kHz. FM radio uses FDM to allocate frequency bands between 88 MHz and 108 MHz to different stations. Time-Division Multiplexing (TDM): Time is segmented into frames, each containing a fixed number of time slots. Each connection is allocated one time slot in every frame. Example: If a link transmits 8,000 frames per second and each time slot is 8 bits, the transmission rate per circuit would be 64 kbps. Illustration of FDM and TDM: FDM: Frequency spectrum is divided into multiple bands (e.g., 4 x 4 kHz bands for 4 circuits). TDM: Time is divided into frames, each with slots assigned to different connections (e.g., 4 time slots per frame).  Example of File Transmission Over Circuit-Switched Network Given Data: 640,000 bits to be sent. 24 slots in TDM and 1.536 Mbps bit rate. Circuit establishment time = 500 ms. Transmission Calculation: Transmission rate per circuit: (1.536 Mbps)/24 = 64kbps. Time to send 640,000 bits: (640,000 bits)/(64 kbps)=10 seconds. Total time (including circuit setup): 10+0.5=10.5  Packet Switching vs Circuit Switching Criticism of Packet Switching: Real-Time Services: Critics argue that packet switching is unsuitable for real-time services (e.g., phone calls, video calls) due to variable and unpredictable delays, mainly caused by queuing. Advantages of Packet Switching: Better Capacity Sharing: More efficient in utilizing available transmission capacity. Cost and Simplicity: Easier and cheaper to implement compared to circuit switching. Efficiency in Packet Switching (Example 1): Circuit Switching: In a 1 Mbps link, each active user requires 100 kbps, limiting the link to 10 simultaneous users (1 Mbps / 100 kbps). Packet Switching: Users are active only 10% of the time. With 35 users, the probability of having more than 10 active users simultaneously is very low (0.0004), allowing efficient link utilization. When fewer than 10 users are active, packet switching operates without delay, similar to circuit switching, but can support more than three times the number of users.  Example of Data Transmission (Example 2): TDM Circuit Switching: For 10 users, if one user generates 1,000 packets of 1,000 bits each, they can only transmit at their allocated time slot, taking 10 seconds to transmit 1 million bits. Packet Switching: The active user can transmit at the full 1 Mbps link rate, completing the transmission in 1 second. Key Difference in Performance: Circuit Switching: Pre-allocates bandwidth regardless of demand, leading to unused bandwidth when not needed. Packet Switching: Allocates bandwidth on demand, only when users have data to send, leading to better utilization of the link. Trend Toward Packet Switching: The shift is toward packet switching, even in traditional circuit-switched networks. Example: Telephone networks increasingly use packet switching for the international portion of calls. Internet Structure: a "Network of Networks" End Systems and Access ISPs: End systems (PCs, smartphones, servers, etc.) connect to the Internet via access ISPs. Access ISPs provide wired or wireless connectivity through technologies like DSL, cable, FTTH, Wi-Fi, and cellular. The access ISP may not be a telecom or cable company; it could be a university or company providing access. Interconnecting Access ISPs: Connecting end users to access ISPs is just one part of the Internet's structure. To enable all end systems to communicate, access ISPs must be interconnected. This creates a "network of networks," which is key to the structure of the Internet. Challenges of Connecting Access ISPs: With millions of access ISPs, connecting them all directly (using a mesh design) would be very costly. A direct link between each pair of access ISPs would require O(N²) connections, which is not scalable.  Global Transit ISP: The first network structure involves connecting access ISPs to a global transit ISP, which has a worldwide network of routers and communication links. The global transit ISP would have routers near each access ISP to handle traffic. Building such an extensive network is costly for the global transit ISP. Economic Model: The global transit ISP would charge access ISPs for connectivity based on the amount of traffic exchanged. The access ISP is considered the customer, and the global transit ISP is the provider. Cost and Scalability: Connecting each access ISP to a single global transit ISP is a more scalable approach than direct mesh connections between access ISPs.  Emergence of Multiple Global Transit ISPs: If one company builds a profitable global transit ISP, other companies may build their own global transit ISPs to compete. This leads to Network Structure 2, which consists of multiple global transit ISPs interconnecting with hundreds of thousands of access ISPs. Access ISPs benefit by choosing between competing global transit providers based on pricing and services. Global Transit ISPs Interconnection: Global transit ISPs must interconnect to allow communication between access ISPs connected to different transit ISPs. Without interconnection, access ISPs would be isolated based on their chosen global transit provider. Two-Tier Hierarchical Network (Network Structure 2): Global transit ISPs form the top tier, and access ISPs form the bottom tier. In practice, no single global transit ISP has coverage in every city. Regional ISPs exist in different regions, connecting access ISPs to tier-1 ISPs.  Tier-1 ISPs: There are about a dozen tier-1 ISPs (e.g., Level 3, AT&T, NTT). Tier-1 ISPs do not pay anyone since they sit at the top of the hierarchy. Regional ISPs connect to tier-1 ISPs for global interconnectivity. Regional ISPs and Multi-Tier Hierarchy (Network Structure 3): Access ISPs connect to regional ISPs, which in turn connect to tier-1 ISPs. Some regions may have a larger regional ISP (e.g., a national ISP in China) that connects smaller regional ISPs to tier-1 ISPs. This multi-tier structure approximates the current Internet setup. Enhancing the Network Structure (Network Structure 4): To resemble today’s Internet, Points of Presence (PoPs), multi-homing, peering, and Internet Exchange Points (IXPs) are added. PoPs: Locations where customer ISPs connect to provider ISPs' routers. Multi-homing: ISPs connect to multiple providers to ensure connectivity even if one provider fails. Peering: ISPs at the same hierarchical level can directly connect, reducing costs by avoiding upstream intermediaries. Peering is often settlement-free. IXPs: Third-party locations where multiple ISPs can peer, reducing reliance on  Customer-Provider Relationships: Customer ISPs pay provider ISPs based on the amount of traffic exchanged. Settlement-free peering between ISPs at the same level helps reduce costs and increase efficiency. Internet Exchange Points (IXPs): IXPs are facilities where multiple ISPs can connect and peer, facilitating direct traffic exchange. There are over 400 IXPs worldwide, helping to reduce costs for ISPs and improve overall Internet efficiency. Network Structure 4: The final ecosystem involves access ISPs, regional ISPs, tier-1 ISPs, PoPs, multi-homing, peering, and IXPs. This structure reflects the modern Internet’s complex, multi-tier network of networks.  Network of Networks: The Internet consists of multiple interconnected ISPs, forming a "network of networks." Competition in Global Transit ISPs: If one company builds a successful global transit ISP, others will emerge, creating competition and multiple global transit ISPs. Access ISPs' Preferences: Access ISPs prefer having competing global transit ISPs, as they can choose based on pricing and services. Interconnection of Global Transit ISPs: For communication between access ISPs connected to different global transit ISPs, these global ISPs must interconnect. Network Structure 2: In this model, global transit ISPs are at the top, and access ISPs are at the bottom, forming a two-tier hierarchy. Regional ISPs: In some regions, regional ISPs connect access ISPs to global transit ISPs, forming a more complex structure (Network Structure 3). Tier-1 ISPs: These ISPs have extensive coverage, but no single ISP has a presence in every city. Around a dozen tier-1 ISPs exist (e.g., AT&T, Level 3 Communications).  Multi-Tier Hierarchy: Network Structure 3 may include larger regional ISPs that connect smaller regional ISPs to tier-1 ISPs. PoPs and Multi-Homing: Points of Presence (PoPs) allow ISPs to connect, and ISPs can multi-home, connecting to multiple providers for redundancy. Peering and IXPs: ISPs can peer directly with each other, often settlement-free, to exchange traffic without using higher-tier ISPs. Internet Exchange Points (IXPs) enable this peering. Network Structure 4: This structure adds PoPs, multi-homing, peering, and IXPs to Network Structure 3, creating a more sophisticated ecosystem. Content Provider Networks: Content providers like Google and Facebook run private networks to connect their data centers and bypass upper-tier ISPs. Network Structure 5: Today's Internet includes content provider networks, with major providers like Google running their own global networks. Google's Private Network: Google connects its data centers globally through a private network, bypassing tier-1 ISPs where possible, but still connecting to them for reachability. Advantages for Content Providers: Running private networks allows content providers to reduce payments to upper-tier ISPs and have more control over service delivery. Summary of Today's Internet: The modern Internet is made up of tier-1 ISPs, regional ISPs, access ISPs, and content provider networks, all interconnected through various means including IXPs and direct peering.  DELAYS Packet Journey: A packet travels from the source to the destination, passing through various routers along the way. Types of Delays: Nodal Processing Delay: Time to examine the packet’s header and detect errors. Queuing Delay: Time spent in the queue before the packet can be transmitted. Transmission Delay: Time taken to push the packet’s bits into the link. Propagation Delay: Time taken for the packet to travel through the physical link. Total Nodal Delay: The sum of processing, queuing, transmission, and propagation delays: dnodal=dproc+dqueue+dtrans+dprop ​ Packet Loss: Loss occurs when the packet arrives at a full queue, which cannot store more packets. Packet loss increases as traffic intensity approaches the link capacity (i.e., when the arrival rate exceeds the transmission rate). Processing Delay: Involves checking the packet’s header, detecting errors, and routing the packet. Typically in microseconds. 1. Transmission Delay: 1. Calculated as: L/R, where L is the packet length and R is the transmission rate. 2. Typically on the order of microseconds to milliseconds. 2. Propagation Delay: 1. Calculated as d/s, where d is the distance and s is the propagation speed of the medium (around 2×108 to 3x108m/s). 2. Propagation delay is typically in the range of milliseconds in wide-area networks. 3. Queuing Delay: 1. Depends on the packet arrival rate and the transmission rate. 2. If the traffic intensity (arrival rate / transmission rate) exceeds 1, the queuing delay can become very large, leading to packet loss. 4. Traffic Intensity: 1. Denoted as La/R, where a is the average packet arrival rate, L is the packet length, and R is the transmission rate. 2. If traffic intensity exceeds 1, the queue will grow unbounded, and delay will approach infinity. 5. Packet Loss and Retransmission: Packet loss occurs when a packet arrives at a full queue, resulting in dropped packets. Lost packets may be retransmitted by the source or other systems. 11.Impact of Traffic Intensity: If La/R ≤ 1, queuing delay is small, but if La/R → 1, queuing delay increases rapidly. When La/R > 1, the system becomes overloaded and delays increase significantly. Real Internet Delays and Routes 1. Traceroute Program: Measures delay from source to router on the path towards the destination. Sends three packets with a Time-to-Live (TTL) value that increases with each hop. Routers reply, and the sender measures the time between transmission and reply. 2. Traceroute Example: Example traceroute from gaia.cs.umass.edu www.eurecom.fr to with several hops. Shows round-trip delays (measured in milliseconds) for each hop. Some probes may fail, indicated by "*", suggesting packet loss or router failure to reply. 3. Observations on Delays: Delays decrease as the path progresses, likely due to fewer hops in the path, means packet loss or no response from a router. 4.Traceroute Testing: You can run traceroutes from different countries using platforms like www.traceroute.org. Protocol Layers and Reference Models 1. Network Complexity: 1. Networks consist of hosts, routers, links, protocols, hardware, and software. 2. The organization of the network and its protocols can be complex. 2. Protocol Layering: 1. Protocols are organized in layers to simplify the design of network systems. 2. Each layer has its own services and uses the services of the layer below it. 3. Layer Functionality: 1. Each layer may provide certain services like reliable message delivery, often by adding additional functionality to an unreliable service from the layer below. 2. Example: Transport layer provides reliable delivery, using network layer’s unreliable message delivery. 4. Implementation of Layers: 1. Some layers are implemented in software (e.g., application, transport), while others (physical, data link) are implemented in hardware. 2. Protocols at each layer are distributed across end systems and network components. 5. Merits and Drawbacks of Layering: 1. Merits: Modularity and ease of updates. 2. Drawbacks: Duplication of functionality across layers and the need for information from other layers may break the goal of separation. 6. Protocol Stack: 1. A stack of protocols in the network is referred to as the protocol stack. 2. Internet protocol stack consists of five layers: physical, link, network, transport, and application. Internet Protocol Stack Application Layer: Supports network applications (e.g., HTTP, SMTP, FTP). Handles human-friendly names (e.g., DNS for domain name resolution). Transport Layer: Responsible for transferring data between application endpoints. Two main protocols: TCP (reliable, connection-oriented) and UDP (unreliable, connectionless). Network Layer: Handles routing of datagrams from source to destination. Key protocol: IP (Internet Protocol) which defines fields and behavior of datagrams. Link Layer: Transfers data between network elements over a specific link. Examples: Ethernet, WiFi, PPP. Delivers packets from one node to the next along the path. Physical Layer: Handles the transmission of raw bits over the physical medium (e.g., copper wire, fiber optics). Specific physical-layer protocols depend on the medium used.