Dynamic Routing Protocols

Dynamic Routing Protocols
  • Listen for subnet information from other routers
    • Sent from router to router
  • Provide subnet information to other routers
    • Tell other routers what you know
  • Determine the best path based on the gathered information
    • Every routing protocol has its own way of doing this
  • When network changes occur, update the available routes
    • Different convergence process for every dynamic routing protocol
Which routing protocol to use?
  • What exactly is a route?
    • Is it based on the state of the link?
    • Is it based on how far away it is?
  • How does the protocol determine the best path?
    • Some formula is applied to the criteria to create a metric
    • Rank the routes from best to worst
  • Recover after a change to the network
    • Convergence time can vary widely between routing protocols
  • Standard or proprietary protocol?
    • OSPF and RIP are standards, some functions of EIGRP are Cisco proprietary
Distance-vector routing protocols
  • Information passed between routers contains routing tables
    • How many "hops away is another network?
    • The deciding "vector" is the "distance"
  • Usually automatic
    • Very little configuration
  • Good for smaller networks
    • Doesn't scale well to very large networks
  • RIP, RIPv2, EIGRP
Link-state routing protocols
  • Information passed between routers is related to the current connectivity
    • If it's up, you can get there.
    • If it's down, you can't.
  • Consider the speed of the link
    • Faster is always better, right?
  • Very scalable
    • Used most often in large networks
  • OSPF - Large, scalable routing protocol
Hybrid routing protocols
  • A little link-state, a little distance-vector
    • Not many examples of a hybrid routing protocol
  • BGP (Border Gateway Protocol)
    • Determines route based on paths, network policies, or configured rule-sets


Comments

Post a Comment

Popular posts from this blog

Spanning Tree Protocol

Image
Loop protection Connect two switches to each other They'll send traffic back and forth forever There's no "counting" mechanism at the MAC layer This is an easy way to bring down a network And somewhat difficult to troubleshoot Relatively easy to resolve IEEE standard 802.1D to prevent loops in bridged (switched) networks (1990) Switch operation Forwarding decisions made by MAC address Keeps a big table of MAC address that have been seen All forwarding decisions are filtered through this list If the destination MAC is unknown, the frame is flooded Sent to every switch port in the local subnet/VLAN Hopefully the destination station will respond Flooding is hopefully a temporary process Directed traffic resumes when the MAC is seen STP port states Blocking - Not forwarding to prevent a loop Listening - Not forwarding and cleaning the MAC table Learning - Not forwarding and adding to the MAC table Forwarding - Data passes through and is fully operational Disabled - Admini...
Read more

Unicasts, Multicasts, and Broadcasts

Image
Unicast One station sending information to another station Send information between two systems Web surfing, file transfers Does not scale optimally for streaming media Multicast Delivery of information to interested systems One to many Multimedia delivery, stock exchanges Very specialized Difficult to scale across large networks Broadcast Send information to everyone at once One packet, received by everyone Limited scope - the broadcast domain Routing updates, ARP requests Not used in IPv6 - focus on multicast
Read more

Protocol Data Units

PDU (Protocol Data Unit) A unit of transmission A different group of data at different OSI layers Ethernet operates on a frame of data It has no idea what's inside IP operates on a packet of data Inside is TCP or UDP segment, UPD datagram TCP or UDP PDU - TCP segment, UDP datagram Maximum Transmission Unit (MTU) Maximum IP packet to transmit - but not fragment Fragmentation show things down Losing a fragment loses an entire packet Requires overhead along the path Difficult to know the MTU all the way through the path Automated methods are often inaccurate, especially when ICMP is filtered Troubleshooting MTU MTU sizes are usually configured once Based on the network infrastructure and don't change often A significant concern for tunneled traffic The tunnel may be smaller than your local Ethernet segment What if you send packets with Don't Fragment (DF) set? Routers will respond back and tell you to fragment Hope you get the ICMP message! Troubleshoot using ping Ping with DF...
Read more