EIGRP vs OSPF vs RIP
so i am currently learning about different routing protocols and am getting somewhat confused between the three i have learned about so far.
so this is currently what i know. EIGRP is proprietary to Cisco and uses cisco proprietary elements to manage routing (wont go into all the details ive learned). Next is OSPF, its an interior gateway,link state,dynamic routing protocol. Also OSPF routers and links are grouped into logical areas. Finally is RIP, RIP is also a gateway,link state,dynamic routing protocol. RIP also uses authentication to verify source.
so my question is which one is actually being used in enterprise networks? All three claim to be popular in today's networks. Also what is the difference between RIP and OSPF, they almost seem exactly the same?
so this is currently what i know. EIGRP is proprietary to Cisco and uses cisco proprietary elements to manage routing (wont go into all the details ive learned). Next is OSPF, its an interior gateway,link state,dynamic routing protocol. Also OSPF routers and links are grouped into logical areas. Finally is RIP, RIP is also a gateway,link state,dynamic routing protocol. RIP also uses authentication to verify source.
so my question is which one is actually being used in enterprise networks? All three claim to be popular in today's networks. Also what is the difference between RIP and OSPF, they almost seem exactly the same?
Comments
OSPF basically builds a database of routes to its neighbors and using an algorithm, calculates the best possible (i.e shortest) path. Its a link-state protocol because OSPF really cares and needs to be aware of the up or down state of its neighbors in order to calculate the best (shortest) path.
RIP uses "next-hop" as its metric. It calculates the best route based on the number of "hops" it takes to reach the specified subnet.
EIGRP is indeed cisco proprietary but is considered to be somewhat of a hybrid between link-state and distance vector; its often referred to as an advanced distance vector protocol. To put it simply, EIGRP basically does what OSPF does and what RIP does and combines the two into a set of algorithms to learn and place the best route for each subnet.
Hope that helps!
Certifications: CCENT; CCNA: R&S; Security+
Next up: RHCSA
so how would an enterprise choose which protocol to use? can different routers use different protocols? sorry if this is a dumb question, i am just starting to learn networking
There's a lot of variables/reasons to use one protocol over another in an enterprise. It really depends on the needs of that network. Smaller to medium sized businesses might be ok with using RIP-2; larger enterprises might get more efficiency using EIGRP or OSPF; or an ISP or telco might use IS-IS to manage the vast number of devices and routers. Each protocol is going to bring its own set of benefits to the table, a network engineer's job is to figure out which is best suited for their network.
Routers come in all sorts of flavors. Some routers are perfect for your home office/small office, whereas you'll need better routers for larger networks (to handle more traffic, etc). Your "average" enterprise router can probably handle your standard fair of IGP routing protocols like OSPF, RIP and EIGRP. Now keep in mind that EIGRP is Cisco-proprietary, so you'll only find that particular protocol on Cisco equipment.
Certifications: CCENT; CCNA: R&S; Security+
Next up: RHCSA
Packet Pushers did that show recently, and diverged into IGP's in general, including EIGRP. It's about an hour and a half long, but worth it if you want to hear the opinions on interior routing protocols by some industry veterans
This is normally not a good idea, however, as redistribution comes with a loss of information. Redistributing RIP into OSPF, for example, does not now include shortest path information, it makes OSPF behave in a distance vector manner, as the only shortest path involved is in the path to get to the router that's redistributing the routes. Then there's the fact that if you're redistributing at multiple points, you have to do some careful planning to ensure you're not causing routing loops.
And sometimes, you need to run multiple routing protocols to support new technologies. For example, our backbone runs IS-IS and OSPF. IS-IS was deployed to support ipv6 because at the time we needed it, OSPFv3 didn't exist in our vendors code, leaving IS-IS as the only option.
But the holy grail is to have your IGP running one protocol. It simplifies the network and makes troubleshooting alot easier, not to mention cutting out those nasty routing loop possibilities. Personally, my preferred design is to have every site run OSPF internally, with all cross-site links running eBGP. I do not appreciate fiber cuts in BFE, Kentucky causing convergence events in Georgia, and the area 0 requirement makes scaling OSPF while containing SPF runs after a link fails a bit onerous.
I agree that open standards is defiantly better unless you know for sure your going to be a Cisco shop for the next two decades.
On the other hand, I know better than that, and todays guarantee is tomorrows 'oh, by the way.... we're buying Juniper gear'.
I'd agree that OSPF is the better choice to implement, but it has nothing to do with it being a better protocol, that's subjective. I'd employ OSPF for three reasons:
#1 - Open standard. No need to worry about vendor lockin
#2 - Mature and active code development (comparitively, anyway)
#3 - The crapload of people out there that know OSPF.
Unfortunately, technical matters aren't the only consideration when you're designing a network. Integration, vendor support, and staffing are all very relevant factors.
Umm, thanks for the info. I agree! I always found OSPF to be the more complex protocol but complex in a good way (fun!).