Well hello again! Today, we're gonna do some more routing, I think this will be the second to last one. I anticipate doing about 8 pages today. I might also start doing entries from home. Regardless, let's get right in!
So last time we talked a bit about RIP and BGP. Well basically, the limitation of RIP motivated people to create a fast protocol, with low bandwidth requirement. This brought forth the "Link State" dynamic routing protocols. There are only two protocols for Link State, they are: OSPF and IS-IS (Not Islamic State.)
Firstly, Open Shortest Path First (OSPF) is the most common one of the two. Even most ISPs use OSPF, chances are, you do too! It converges at a much faster rate than RIP, and works within a single AS. However, if you plan on using it, it's a pretty complicated protocol. There will be no OSPF on low-end routers cause it takes a lot of computational power. The main reason this is important to cover, is cause it will be on the Network+ exam! The difference with OSPF and RIP, is that OSPF is designed to work with the Internet. Let's use the example from last time of RIP, and this time use OSPF. By the way, if you don't have this textbook, you'll probably be out of luck (I'm using "All-in-One CompTIA Network+ Fifth Edition.) If you do, I refer you to Pg. 255, and go back a couple pages for the example on RIP (I forgot to put that it was based on a textbook figure last entry.)
This time we're giving Router B an upstream connection to the ISP. Starting an OSPF capable router for the first time will cause it to send out LSAs (Link State Advertisements), a.k.a "hello packets", which are used to look for other OSPF capable routers. The new router will send many LSAs upon starting it, and this is known as "flooding", for obvious reasons. Another difference with OSPF and RIP are the hop costs. In RIP, the cost of a hop was always 1, unless changed manually. For OSPF however, it's dependent on the speed of the link. The formula is 100,000,000/bandwidth in bps. The 10BaseT link's OSPF was 10 based on "100,000,000/10,000,000 = 10". This shows, the faster the bandwidth, the lower the costs. You may override it manually as you could with RIP, but I don't know why you would. When an OSPF router sends its hellos, it will exchange info and update their link state databases. The hello message is project throughout the network to any other routers. Of course you wouldn't want the flood to leak out of the network, so you'd assign an Area ID to the router. Usually it's assigned one by default, and it will accept the area ID like the well behaved router it is! All routers are in the example, given let's say, 0.0.0.0 which is known as "Area 0". I wonder what Area 51 would be, haha. So Area 0 is actually important in the OSPF world, as you could make more areas if the network is to get more complex, but Area 0 would remain the default backbone, and the most important part. The reason why areas are even important are to minimize traffic between routers. Every area has one router which is the "Designated Router" (DR), I'd say it's like the President of the United Network. Then there's a backup designated router (BDR), which is like the Vice President. When the routers communicate, there's an election for the DR and BDR! Could you imagine, these routers have minds of their own! It's like if the Terminator was electing a leader (it would probably be him.) In this case, the DR is Router B, and Router A is the BDR. The election will take place during the hello packets. Most of the tie you'd just let the router decide (unless you're a dictator!), but you can manually change it, but you would be violating the humanity of the routers, I say! It's very rare that anyone sets the DR, since the routers are good at picking.
After the elections, the routes are distributed across the area. Routers A and B will send separate LSAs telling each other that they are connect to NIX and NIY. These messages don't contain the entire routing table from the respective routers however. Basically, as you can tell this process is going a lot quicker than with RIP. These routers will send each other LSAs and hello messages in intervals of 30 minutes. They also keep alternate routes, unlike RIP, which would simply discard them. Now, what if there were to be a sudden disconnection between Routers A and B? Well they'd detect the break instantly, and would try and reconnect. If that failed, then the routers would send out an LSA announcing the broken connection. It's no surprise that OSPF became so popular. In addition to all these great features, it also supports authentication and the prevention of loops. Is there any reason to why people would want an alternative? Well for a while it didn't have support for IPv6, but it does now. So there really is no reason not to use it!
Then there are the other protocols which aren't as widely used. First, IS-IS which is really the only other Link State option. It's basically a lot like OSPF, except it had an advantage to the IPv6 addressing. However, that's not much of an issue anymore. To put it bluntly, this isn't even a close second to OSPF. Then there's EIGRP which isn't either a distance vector or link state protocol. This protocol belongs to the mighty Cisco. They released it when people were demanding a better version of RIP, so Cisco released this. It's fading away as well. without dynamic routing, the Internet would cease to exist, so appreciate this, cause it make it all possible. I'll refer to the chart on Pg. 259 to give you a description of all the protocols and a couple notes. On that bombshell, thanks for reading, and until next time!
Completion Status: 38%
Pages Left:
- Book: 420 pages
- Chapter: 11 pages
No comments:
Post a Comment