History of Internet

• J.C.R. Licklider proposed the "Galactic Network" in 1962.

A Brief History of the Internet

• Licklider's vision was a network where everyone could quickly access data through interconnected computers.
• Licklider led researchers at DARPA to experiment with connecting two computers.
• Lawrence G. Roberts, an MIT researcher, connected computers in Massachusetts and California with a low-speed dial-up line.
• The ARPANET was established in 1969.
• Early experiments showed time-shared infrastructure worked, but researchers needed packet switching technology.
• Roberts developed the first network, connecting four nodes by the end of 1969
• The four nodes included UCLA, Stanford Research Institute, UCSB, and the University of Utah.
• In 1970, the Network Control Protocol (NCP) was introduced as the initial ARPANET host-to-host protocol
• NCP allowed network users to start developing applications.
• Email was among the first applications, launching in 1972.
• In 1973, a DARPA team led by Bob Kahn introduced open-architecture networking.
• Open Architecture allows individual networks to be independently designed and developed based on specific needs.
• Kahn collaborated with Vint Cerf at Stanford and presented the original TCP paper in 1973.
• The Transmission Control Protocol / Internet Protocol (TCP/IP) was developed from open architecture networking idea
• The first version of TCP was split into IP for addressing and forwarding packets, and TCP for service features like flow control.
• Paul Mockapetris at USC introduced the Domain Name System (DNS) in 1983.
• DNS translates domain names to IP addresses using a scalable, distributed mechanism since the Internet's scale was rapidly increasing.
• Tim Berners-Lee led a team of researchers to introduce the World Wide Web (WWW) in 1990, it became one of the first and most popular applications

Internet Architecture Introduction

• Current architectural design involves connecting hosts running the same applications but located on different types of networks.
• Computer networks are complex systems built on multiple components with varying technologies and application types.
• Network protocols organize protocols into layers to provide structure to the network architecture

Architecture, Layers, and Functionalities.

• Functionalities in network architecture are implemented by dividing the architectural model into layers.
• Each layer offers different services.
• An analogy to explain layered architecture is the airline system. Purchasing a ticket, checking bags, going through gates, flying, and claiming baggage are different layers of the airline travel process.
• Every layer implements some functionality, works based on the service provided by the layer below, and also provides some service to the layer that is above

Layering and Functionality

• Advantages of layered architecture include scalability, modularity, and flexibility to add or delete components for cost-effective implementations.

The OSI Model

• The Internet architecture follows a layered model.
• Every layer provides some service to the layer above.
• The seven-layered OSI (Open Systems Interconnection) model was proposed by the International Organization for Standardization (ISO).
• The OSI model layers include application, presentation, session, transport, network, data link, and physical layers.
• The Internet architecture model has five layers.
• Application, presentation, and session layers are combined into a single application layer with sockets as the interface between application and transport layers.
• The application developer designs the overall application functionality

Application, Presentation, and Session Layers

• The application layer includes protocols such as HTTP (web), SMTP (e-mail), and FTP (file transfer).
• The DNS protocol translates domain names to IP addresses.
• Services offered by this layer depend on the application.
• At the application layer, the packet of information is referred to as a message.
• The presentation layer formats information received from the layer below and delivers it to the application layer.
• Presentation layer functionalities include formatting video streams or translating integers between big endian and little endian.
• There are no example protocols for the presentation layer.
• The session layer manages different transport streams belonging to the same session between end-user application processes.
• As an example the session layer ties together audio and video streams in a teleconference application.

Transport and Network Layer

• The transport layer is responsible for end-to-end communication between end hosts using TCP and UDP protocols.
• TCP offers a connection-oriented service, guaranteed delivery, flow control, and congestion control.
• UDP provides a connectionless best-effort service without reliability, flow, or congestion control.
• At the transport layer, a packet of information is called a segment.
• In the network layer, the information packet is referred to as a datagram.
• The network layer is responsible for moving datagrams from one Internet host to another.
• The IP Protocol is the glue that binds the Internet together and must be run on all hosts/devices with a network layer.
• IP protocol defines datagram fields and how hosts/routers use them so that the sent datagrams reach their destination.
• Routing protocols determine the routes datagrams can take between sources and destinations.

Data Link Layer and Physical Layer

• In the data link layer, packets of information are referred to as frames.
• Ethernet, PPP, and WiFi are Protocols used at the data link layer
• The data link layer moves frames from one node to the next using network layer routing.
• This layer offers services such as reliable delivery, covering data transmission across one link.
• The reliable delivery service offered is different from the TCP protocol, which offers reliability from source host to destination.
• The physical layer facilitates interaction with hardware, transferring bits within a frame between two nodes via a physical link.
• Protocols depend on the link and transmission medium, like Ethernet which has twisted-pair copper wire, coaxial cable, etc.
• At the physical layer, info is moved between two systems connected by a physical link and specifies how to send one bit.
• Coaxial cable, fiber optics, and radio frequency transmitters are examples of physical layer transmission.

Layers Encapsulation

• Layers and protocols communicate through encapsulation and de-encapsulation, referring to the data's physical path from sender to receiver.
• The sending host sends an application layer message (M) to the transport layer.
• The transport layer adds header information(Ht) to the application layer message, called a segment.
• Added header info, segment, helps the receiving host direct the message to the correct application and determine if there were errors.
• The segment gets forwarded to the network layer, which adds the network header information (Hn), and the combination is called a datagram.
• The datagram encapsulates the segment and the added header gives source and destination addresses of end hosts.
• The message at the link layer is called a frame, transmitted across the physical medium with its own header information (Hl).
• At each layer the message is a combination of the payload from the layer above with the new appended header
• At the receiving end, de-encapsulation reverses the process by stripping headers off each layer.
• Intermediate layer-3 devices like routers, and layer-2 devices like switches may be part of Path between sender and receiver
• Routers and layer-2 switches implement protocol stacks similarly to end-hosts.
• Routers implement layers 1 to 3, and layer-2 switches implement layers 1 to 2.
• The switch de-encapsulates data and encapsulates it again to send it forward to the next device.
• End-hosts implement all five layers, while intermediate devices don't which makes the network edges smarter, and the core simple which is called end-to-end principle.

The End-to-End Principle

• The end-to-end (e2e) principle suggests that specific application-level functions cannot be built into the network's lower levels or the core.
• The network core should be simple and minimal, and the end systems (edges) should carry the intelligence.
• Moving functions and services closer increases the application designer's autonomy to offer services specific to their application
• The question of the matter, is that the higher-level protocol layers are more specific to a particular end-to-end service needed by a particular app
• Where as the the lower level protocol are free to focus on the more minimal generic functions that benefit all different types of apps.
• Violations of the e2e principle typically refer to cases where a functionality cannot be implemented entirely by the end hosts.
• Examples of the E2E violation include firewalls and traffic filters, and NAT boxes

Violations of the End-to-End Principle and NAT Boxes

• Firewalls are examples of E2E principle Violations, because they are operated as intermediate devices between end hosts and drop certain traffic.
• Network Address Translation (NAT) boxes are also an E2E principle Violation.
• NAT boxes enable for multiple devices in a home network to connect to the Internet, NAT-enabled home routers are commonly setup
• An internet service provider typically assigns a single public IP address to a home router and specifically to the the globally network facing interface
• The other NAT-enabled router interface facing the home network, is assigned an IP address belonging to a private subnet eg 10.0.0/24
• Private and local IP addresses only have meaning to devices within that particular subnet.
• Private networks are always behind a NAT, which enables for communication between the local hosts on local networks and public internet hosts
• All the traffic leaving the house router to the public internet, uses the public facing interface of the NAT enabled router.
• When the home router receives network traffic, it has as destination address the IP of the public facing interface and plays a translator using NAT translation talbe.
• Translation table that provides mapping between public IP and private IP ranges, that belongs to the internal hosts.
• The hosts behind NAT boxes are not publicly addressable or routable.
• Due to the NAT configurations, not possible for others to create connections to these behind NAT boxes.
• Tools include STUN (to discover NATs) and UDP hole punching (to establish bidirectional UDP connections).

The Hourglass Shape of Internet Architecture

• The Internet protocol stack has a layered hourglass shape.
• Multiple network-layer protocols competed with IPv4 in the early 1990s.
• Innovations happen most frequently at the bottom and top layers.
• A model called the Evolutionary Architecture model, or EvoArch, that can help to study layered architectures and their evolution in a quantitative manner can explain the hourglass shape.

Evolutionary Architecture Model

• The EvoArch model considers an abstract model of the Internet's protocol stack components
• A protocol stack represented as a directed and acyclic network with L layers.
• Each network protocol is represented as a node.
• Protocol dependencies are represented as directed edges.
• If protocol at layer one uses servies from protocol W at layer I-1, then it is represented by by edge from W, to U
• Substrates node represent S(u),the set of nodes that U, is using their service, every node has a substrate except bottom layer.
• The outgoing edges from a node U, are U. products represented by P(U)
• Each Layer associated probability S(I), which refer to as layer generality.
• Node U at layer I+1 selects independently each node of layer I, with the substrate probability
• As we move to higher layers, layer generality decreases and lower layers protocols are most general
• For given of protocol node U, denoted the V(U), is recursively computed based on product U
• If competition threshold, or C, is shared, a large percentage of U's product
• In the the model, the most competive nodes are the U, nodes at that Layer
• The competition among nodes increases intenseley, and protocl U dies if it has comp higher value
• In the model protocol, one new basic birth place: New node is randomly assigned to the given layer.
• Model with "Toy Network with Four Layers" is considered, the value of each node is at a circle, the probability is at the left

EvoArch iterations

• A discrete-time model executed over rounds
• 1. introduce new nodes
• 2. placing randomly and examines by top to bottom
• 3.connect nodes that we introduced choosing substrata by choosing generality from that layer 1-1, then the product is I.
• 4. update each value to each layer I
• 5. We examine each layer decrease based on its current value, and remove any nodes.
• 6. Execution of the model is stopped.
• Model implies for the internet architecture, and future interenet arch. and suggests survivial of the TCP/IP at the time

Implications for the Internet Architecture and future Internet architecture

• EvoArch helps explain the survival of TCP/IP, because the TCP/IP stack wasn't competiting with telephone networks, because it handled differeent applications
• TCP/IP used for FTP, email and telnet, it managed to grow and increase without being threatened with telephone network first appeared.
• A large birth rate at a layer above the waist can cause death for the protocols at that waist. IF there are not chosen the substrates by the new nodes.
• The transer layer acts as an evolutionary shield, as more ports appear, are unkilely by new TCp adn UDPs
• Stability is added as new products are chosen and compete.
• EvAOch model predicts future brands and evolution of the shapes, which will lead to ossificied protocols
• The model suggests layer functionity and try to designers to each layer, wide and general service is not compete

What are some of the ramifications of the "hourglass shape of the internet"?

• Technologies not designed for the internet can be modified and have versions to communicate
• Trasitioning to IPV5 has been a slow process due to the internet structure
• Internet infrastructure uses ipV4, and has high transition cost which is reflects narrow waste consequence

Architecture Redesign

• Researchers believe is neccessary to rethink fundamentals and decisions in design, and that clear-state approach to that is necessary

The clean-slate design as a process

• An important aspect of new approach through clean-slate,is test deployment at appropriate facitility.
• A facitility is needed to function, with requriement infrastructure,and reackout users and traffic.
• In this context, clean-slate needs to viewde as deployment, in new service, and deployment for commerical interenet.
• It is possible to create enterily new network, and which may eventually remove current
• Most conservatice "that we current interen tis beast"
• ""radical outcome cna be what such experiment becoems bue preint future ineternet

Redesigning the Internet architecture to optimize for control and management

• 4D, started a clean late desigin based: netowrk objecitves, view and direct control with function and compnonnet
• In the 4 D deicison has network traffic, the direct over current operatison a no desioc loging hard in and elements
• The is and the data desin technology to ever improve reliablity and platmrs
• "Greenburg [ 2005 ]"

Redesigning the Internet architecture to offer better accountability

• Researchers proposed network “accountability” and provides IP network
• In order to establish foundation, against behaivos of the ability associated for each entity by the response
• Also to the addressed and improvement that could come interent AIP account
• The AIP is used tracing actions of misbehaving hosts.
• Describes address in unique AD; AID for network by gloablity host

Interconnecting Hosts and Networks

• Devices help provide connectivity between hosts in the same network
• Repeaters and Hubs operate in L1 - provide connectivity between hosts and Ethernet segments.
• Advantage: simple and inexpensive, plus arranged in hierarchy.
• Disadvantage: hosts can be arranged through same devices, and can compete access in link
• Bridges and L2 Switches enable communication between no directly host connected.
• Operate is based in MAC addresses, and forward reach reach the appropriates.
• Router and L3 switches = devices operaate on L2 Switch
• These routers run through routing protocols on the next protocols

Learning Bridges

• Bridge is a device that transfers frames from an input to one or multiple outputs, and doesn't have to receive traffic for the task.
• Learnings bridges have population and maintance table, which contaitns forwarding table, consults table forwards ports.
• This means forwarding only on spcesific ports versus ports that the network table needs.
• Is conderes the for topolofy. if bridge table with source Hosta and destination, it is a needed to forwaf it

Looping Problem in Bridges and the Spanning Tree Algorithm

• In the cases of using using bridges, its fails in networks and topologies, and bridges loop forever

• Problem: that includes "loops and problems ",by running tree spam. represanting netwrok topologoy graph

• Every node in the graph has unique graph ID.

• Algorithm runs in “rounds” and sends each node to neighbors with ID in "rounds "

• AT everey Round. each node in the table needs tthat has ever recie.

• To see more details about this tooplgyy.

• A node stops messaging config when. it receive that the indiciate a not root