Internet Protocol Networks: Topologies and Security

• The following are key topics covered in a lecture on Internet Protocol Networks

Overview

• Packet-switched network principles are a foundational concept
• IP, including routing, ICMP, and mobile considerations, is an elemental topic
• IPv6, TCP, SNMP, and DNS are important protocols and systems in IP networks
• Network security to protect IP Networks is an absolute must

WAN Topologies

• WANs (Wide Area Networks) cover large geographical areas, necessitating complex designs unlike smaller LANs
• Mesh topology involves multiple paths between locations with added redundancy
• Redundancy ensures data delivery even if one path fails, often implemented via triangular meshes
• Hierarchical Tree Topology: a tree structure that connects the main network ("trunk") to smaller sub-networks ("branches")
• Mesh and hierarchical tree topologies are two general designs
• Triangulated meshes provide redundancy

Hierarchical Networks

• At the lowest level, hierarchical networks consist of local star network clusters
• Traffic is concentrated as it moves to higher levels in the hierarchy
• This network type is typical of national and international telephone networks
• Redundant routes and ring or mesh networks improve reliability and performance
• Data networks employ a more mesh-like topology

Switching Types

• Calls/connections in WANs are routed through switching nodes, forming connections via multiple links
• Circuit switching establishes a dedicated path, commonly used for telephone calls
• Problems of circuit switching include: setup time, resource waste, blocked calls
• Message switching stores messages at intermediate nodes and forwards them when circuits are available, with header info for control
• Message switching is unsuitable for interactive communications due to delays
• Packet switching breaks messages into smaller units (packets) with header information
• Multiple routes can be set up if throughput requirements demand

Virtual Circuits and Datagrams

• Both are basic approaches to packet switching
• Virtual Circuit Transmission
• Has a set-up phase that establishes the route for all packets in a session
• Packets need only short headers to identify the virtual circuit
• Provides error control and guaranteed delivery ensuring correct sequencing
• Datagram Transmission
• Each packet is treated separately and may follow different routes
• Relies on the user (Transport Layer) to implement error control, guaranteed delivery, and resequencing
• Requires that only simple protocols are used between intermediate nodes
• If connection is desirable, but error control and resequencing are of minimal importance than minimal delay, a compromise may be best

Comparison of Switching Techniques

• Message switching sees rare use, unless the message size is kept small, since there is no utility for segmenting messages into packets
• Packet switching offers numerous benefits for efficient bandwidth use, retransmissions, and minimized blocking compared to circuit switching
• Circuit switching offers minimal delays with little delay variability once a call has been established

Fast Packet Switching

• Fast Packet switching is also known cell switching.
• Cell Switching
• Developed for future high-speed digital networks, specifically broadband ISDN (B-ISDN)
• Employs fixed-length packets called cells
• Delays must be minimized, especially for types of traffic like digital TV, which do not allow for retransmissions
• End-to-end error correction and retransmission facilities must be applied by others that can accept delays

Public WANs

• Several types of public WANs exist like packet-switched public data networks, circuit-switched public networks, and integrated services digital network (ISDN).
• Most public data networks have converged to interconnection over the Internet
• There the TCP/IP protocol suite is applied as is standard procedure
• Most fixed telephone networks continue to be based on circuit switching, but are slowly moving to all-IP-based
• Integrated Services Digital Network (ISDN): a proposal towards a complete network for all services
• A key element to recognize is that IP based networking has become the the standard

Network Layer Protocols

• WANs require true routing, a Network layer function in the OSI model
• The object of the Network layer to provide an end-to-end communications capability to the Transport layer
• Transport layer is unconcerned with whether circuit or packet switching is implemented
• Network Layers should mask the technology that is used from the Transport Layer
• The Network layer demands cooperation between peer processes at all nodes
• WANs often cross or interconnect over national boundaries, there is a need to agree to internationally agreed standards

Connection-oriented and Connectionless Internetworking

• Virtual circuit packet switching can be used if all networks in provide this service in an internet for reliable connection to internetworking
• Datagram packet switching can be used when the only requirement from each network is to forward te packet to the following node along the route with relatively high probability of success

Connectionless Approach Advantages

• Flexibility
• An internet can consist of many kinds of networks
• Networks simply need to reliably send packets
• Robustness
  • Packets are routed away from areas with congestion
• Reduced Delays
  • Delay is reduced by waiting for correct sequencing at every "hop"
    • End-to-end Reliability -Provided by the Transport Layer if needed

Packet Routing

• The structure of the internet includes a hierarchy
• The Internet consists of large internets itself (ARPANET from which Internet derived, JANET academic network in the UK) and are referred is autonomous systems.a
• Packets sent between autonomous systems via exterior or border gateways
  • Within autonomous systems, networks connected with interior gateways
  • Hosts on the Internet connect to these networks needing global identification.
• Hosts (stations) are identified by unique network layer address (IP address)
• Data is transmitted to stations using a point-of-attachment address (usually a MAC address)
• Address Resolution Protocol (ARP) translates between MAC and IP addresses

Routing Protocols

• Routing protocols apply at the gateway level
• Generally dynamic protocols route single autonomous systems( can differ
• Aim to discover the next hop on a shorter or better path assuming all routers do the same
  • Shortest Path algorithms typically used
    • Routers discover network topology and use that data.
    • The most popular is the Open Shortest Path First OSPF
      • Routers must learn topology
      • Shortest path first algorithm is used
  • Distance vector algorithms
    • Use RIP
    • derive lookup tables determining the next hop by comparing the shortest distance to the destination
• Exterior (Border) routing protocols are protocols between of gateways (routers)
• The best is Border gateway protocols
  • Manage Autonomous systems
  • Each gateways nominate themselves and acquire their own identity -exterior Gateways acquire adjacent Gateways - this acquisition confirmed periodically through confirmation.
  • Gateways exchange information about reachable routes through routing update messages.

Internetwork Protocol Considerations

• Routing must be performed at different levels corresponding to hierarchy (between exterior gateways, between interior and exterior gateways on each autonomous system, and between hosts and gateways)
• Large packet sizes improve ratio between header and data size, leading to more effective bandwidth usage
• Internetwork-wide protocols must cope with varying packet sizes of networks
• Public WANs may have maximum packet sizes of 128 bytes but LANs may have packet sizes of 8000+ bytes
• Gateways/routers could get packets from a number of networks at once, with all perhaps destined for the the same out.
• The packets will require buffering- as well as queuing if slower transmission speeds are applied
• Overly congested packets may get discarded

Fragmentation

• Segmentation and reassembly handle varying network maximum packet sizes
  • Packets are broken up into smaller segments and reassembled. -Intranet fragmentation fragments and reassembles at each router between networks. - Uses networks most efficiently with long reassembly times and high network usage -Internet fragmentation only fragments and the receiving end reassembles the pieces -Reduces reassembly delays with some reduction in efficiency The best approach is do does away with the Routers fragmenting at all- but must have knowledge of shortest route length and transmit in this size. IPv6 removes fragmentation from routers to minimize delay.

Flow/congestion control

• Network layers cannot implement this independently- if they can there won't be identifiable flow
• The most effective method of congestion control is as the congestion will simply naturally delay transport acknowledgements and slow down traffic from hosts
• when congestion becomes a hurdle- routers will absolutely discard packets.
• With routers discarding packets, they need to accurately generate error reports using ICMP "Integral" to IP: Internet control message protocol -Used by gateways and endpoints- has multiple functions
  • The most common use for ICMP:
    • Error Reporting (Discarding packets)
    • Reachability testing (Echoing)
    • Congestion control (Source sending reduction rate) -Route change notifications
    • Performance monitoring
    • Subnet addressing
• The "Integral" ICMP data is sent inside IP packets

Explore the core principles of packet-switched networks, including IP routing, ICMP, and mobile considerations. Understand WAN topologies like mesh and hierarchical designs for redundancy and efficient data delivery. Learn about network security measures to protect IP networks.