Why don't they update RIP to use a better metric?

feng.lianfeng.lian Member Posts: 47 ■■□□□□□□□□
I mean hop counts seems to be a horrible metric and using something else that is simple enough probably won't make a difference in bandwidth and cpu use. I mean they could use something like a very simple metric that uses 8 bits of information instead of 4 bit (I'm guessing 4 since 16 is infinity). It would go something like 10Gb or higher links cost 1, 1 Gb or higher links cost 2, 100 Mb links cost 4, 10 Mb links cost 8, T1 or higher links cost 12, lower than T1 links cost 16. This simplistic scheme would still have 16 hops as infinity. A much more experienced network person could probably come up with much better costs with the same amount of bits and if more bits are allowed, there could be a better scheme.

Who is responsible for maintaining those public standards? IEEE? Why aren't they doing something, especially when creating RIPng?

Comments

  • jovan88jovan88 Member Posts: 393
    Its easy to pick on RIP but thats just the limitation it has being a billion years old. If you need something more scalable you'd want to go for one of the other protocols out there. Also now days your average joe-schmoe home router can run it, being as simple as it is.
  • networker050184networker050184 Mod Posts: 11,962 Mod
    Because it doesn't need a more complex metric for its intended implementations. If you have all those different types of links and a large network you will pick a better suited protocol. RIP is used in smaller network, usually LANs, where the links have eqaual bandwidth and a small diameter network.
    An expert is a man who has made all the mistakes which can be made.
  • wbosherwbosher Member Posts: 422
    Other than a CCNA lab environment, does anyone still use RIP?
  • jovan88jovan88 Member Posts: 393
    I've seen it around but from my own experiences if a network is small enough for RIP then you can probably get by with static routing
  • JSimOJSimO Member Posts: 8 ■□□□□□□□□□
    I have never seen RIP implemented on a network, EVER.
  • networker050184networker050184 Mod Posts: 11,962 Mod
    We have RIP running on some of our older devices that can't run anything else. They are being phased out though.
    An expert is a man who has made all the mistakes which can be made.
  • QHaloQHalo Member Posts: 1,488
    I worked for a Fortune 500 company that was running RIP for quite a while. I think they only within the last few years moved to EIGRP. It's still out there, don't be fooled.
  • APAAPA Member Posts: 959
    JSimO wrote: »
    I have never seen RIP implemented on a network, EVER.

    Don't be fooled into thinking it isn't out there........

    We purchased a company who had a wholesale L3 subscriber agreement with another well-known ISP....... the agreement was signed off with a design that included reachability for the L3 subscribers routes to be delivered via RIP.....

    Yuck...... but it does work......without issues.... so don't be thinking it doesn't have its uses...

    :D

    CCNA | CCNA:Security | CCNP | CCIP
    JNCIA:JUNOS | JNCIA:EX | JNCIS:ENT | JNCIS:SEC
    JNCIS:SP | JNCIP:SP
  • Paul BozPaul Boz Member Posts: 2,620 ■■■■■■■■□□
    I have audited and pentested some of the largest financial networks in the country and have seen RIP deployed more frequently than any other protocol combined, not counting static routes (which aren’t a protocol per se). In the banking world you almost don’t see any routing protocols in institutions worth less than $500m. The $500m to $2b range is usually where some basic routing comes in, but even then I’ve only seen two or three running something sophisticated like EIGRP or OSPF. I just thought about it pretty hard and recall seeing OSPF deployed once and EIGRP maybe 3-4 times. Everyone else used RIP (including the largest bank in Florida).

    RIP is a perfectly fine protocol for what it does. They’ve made enhancements along the way which really make it a better protocol. If you need something more robust or with more complex metrics just scale to a link state protocol or EIGRP.
    CCNP | CCIP | CCDP | CCNA, CCDA
    CCNA Security | GSEC |GCFW | GCIH | GCIA
    [email protected]
    http://twitter.com/paul_bosworth
    Blog: http://www.infosiege.net/
  • hypnotoadhypnotoad Banned Posts: 915
    For what it's worth, I've seen Windows Servers participating in RIP; so they receive backup routes on a regular basis for fail-over.
  • APAAPA Member Posts: 959
    Paul Boz wrote: »
    I have audited and pentested some of the largest financial networks in the country and have seen RIP deployed more frequently than any other protocol combined, not counting static routes (which aren’t a protocol per se). In the banking world you almost don’t see any routing protocols in institutions worth less than $500m. The $500m to $2b range is usually where some basic routing comes in, but even then I’ve only seen two or three running something sophisticated like EIGRP or OSPF. I just thought about it pretty hard and recall seeing OSPF deployed once and EIGRP maybe 3-4 times. Everyone else used RIP (including the largest bank in Florida).

    RIP is a perfectly fine protocol for what it does. They’ve made enhancements along the way which really make it a better protocol. If you need something more robust or with more complex metrics just scale to a link state protocol or EIGRP.

    Well said.. :)

    CCNA | CCNA:Security | CCNP | CCIP
    JNCIA:JUNOS | JNCIA:EX | JNCIS:ENT | JNCIS:SEC
    JNCIS:SP | JNCIP:SP
Sign In or Register to comment.