WANs IV: Route Globally, Act Locally

GLOBAL Route-Sharing

As we talked about in our last discussion, it’s really amazing that the Internet functions as well as it does.  Anyone with access on one side of the planet can instantly send information to someone on the other side, including email, voice, video, and many other types of data.  Ever wonder how it all manages to work so well most of the time?  The answer, which is not usually covered in CCENT/CCNA discussions, is the Border Gateway Protocol, currently at version 4.

When I first started reading up on networking protocols, I was impressed at how OSPF could communicate vast amounts of accurate reachability information to devices in its domain.  The problem, though, is that comparing even a large, 2000-node network, for example, is far different from the Internet.  Communicate the intricate details about every conceivable route (including host routes) would make Internet routers capable of cooking eggs from the sheer heat, because of the massive routing tables and computations involved.

This is where we can talk about an important but sometimes overused networking term—scalable.  Scalability is the ability for something to grow in a controlled, measured fashion, rather than haphazardly or too rapidly.  To allow the Internet to be scalable, routing has to be simplified in some form or fashion, and this is the unique ability of BGP as an exterior gateway protocol (between autonomous systems, namely devices under a common administration).

Routing across the Internet is a lot like the old line you hear from military personnel in most movies—“That information is on a need-to-know basis” (and usually implies that the person being told that does not need to know).  For example, the Internet Service Provider that I first worked at had a block or range of Class C addresses from 216.145.0.0 to 216.145.31.255, and were allocated to various customers that they serviced.  Advertising out all 31 separate routes, or even worse, even smaller subnets, would have created a minimum of 32 entries if not more.  Instead, because of the beauty of BGP, they advertised a single entry, 216.145.0.0/19.  This greatly reduced the size of the potential routing table, and if you multiply that out across the world, you can see why this works so well.

Another important concept is peering, where service providers interconnect and exchange their routes.  Originally this took place at public exchange points, but companies like AT&T and Verizon connect directly to one another to accomplish that, called private peering.  Most of these entities also have rules about the sizes of network advertisements that they will accept as well.

Next time we will start exploring the newest version of the Internet Protocol, IP Version 6!

– Joe

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: