<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	xmlns:atom="http://www.w3.org/2005/Atom"
	xmlns:sy="http://purl.org/rss/1.0/modules/syndication/"
		>
<channel>
	<title>Comments on: A Quick Note About Comcast Cable Modems</title>
	<atom:link href="http://andrew.pilsch.com/blog/2008/08/04/a-quick-note-about-comcast-cable-modems/feed/" rel="self" type="application/rss+xml" />
	<link>http://andrew.pilsch.com/blog/2008/08/04/a-quick-note-about-comcast-cable-modems/</link>
	<description>science fiction, new media, technical communications, transhumanism</description>
	<lastBuildDate>Fri, 11 Mar 2011 22:24:45 +0000</lastBuildDate>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
	<generator>http://wordpress.org/?v=3.2.1</generator>
	<item>
		<title>By: Gerrit DeWitt</title>
		<link>http://andrew.pilsch.com/blog/2008/08/04/a-quick-note-about-comcast-cable-modems/comment-page-1/#comment-1817</link>
		<dc:creator>Gerrit DeWitt</dc:creator>
		<pubDate>Wed, 08 Oct 2008 03:29:21 +0000</pubDate>
		<guid isPermaLink="false">http://andrew.pilsch.com/blog/?p=114#comment-1817</guid>
		<description>Hey Andrew.  Gerrit here.  :)

Yeah, that’s normal.

The way I see it, most of these “modems” fall into one of two categories:

A. those that can perform layer 3 functions such as routing and NAT
B. those that act as layer 2 bridges

Sounds like you have number B.  :)

DSL devices are usually type A – NAT routers – because the communication to the service provider is usually done via PPPoE.  It’s usually easier for customers to go to http://yeahigotdsl.com and enter a name and password than to configure PPPoE on their router or computer.  These NAT routers can also be told to go into a bridge mode for another router to perform the PPPoE “dial in.”

Cable “modems” are usually type B.  The big clue is if you plug your computer into the modem, you’ll get a public IP address via DHCP.  So, no NAT there.  The DHCP assignment is usually not coming from the “modem” – rather, it just serves as a bridge to the ISP’s network, which is where the DCHP server resides.

In either case, when operating in bridge mode, the only identifiers that the “modem” and device behind it have are the MAC addresses – the Ethernet frames are addressed this way.  That much makes sense.

Now our friend OSI layer 2 has two parts:  the top-part is the logical link control sublayer (LLC), and the bottom part controls MAC addressing.  All of a switch’s or bridge’s real magic lies in the LLC.

A regular Ethernet switch will let you unplug and connect new devices without rebooting it.  And a switch is better than a hub, because a hub is a layer 1 device that sends responses to everyone – e.g. no one-on-one communication between sender and recipient (no unicast).  So “regular” switches are good for unicast, multicast, and broadcast communication at layer 2.  This behavior is an unacknowledged, connectionless mode, which sends data efficiently to one or more recipients, but doesn’t verify that the recipients would get the data.  Verification could be left up to an upper-layer service instead.  This behavior is defined in IEEE 802.2, which is LLC.

Enough about a regular switch.

IEEE 802.2 also defines two other LLC control methods, one of which is an acknowledged, connectionless mode which only supports one-on-one communication.  For the service to be acknowledged, each of the two devices in the “one-on-one chain” need to know about each other.  Since we’re still in layer 2, the only way to know one another is via MAC address.

So that’s why the router “caches” the MAC.

For the curious, LLC can also operate in a connection-oriented mode, where packet order is maintained.</description>
		<content:encoded><![CDATA[<p>Hey Andrew.  Gerrit here.  :)</p>
<p>Yeah, that’s normal.</p>
<p>The way I see it, most of these “modems” fall into one of two categories:</p>
<p>A. those that can perform layer 3 functions such as routing and <span class="caps">NAT</span><br />
B. those that act as layer 2 bridges</p>
<p>Sounds like you have number B.  :)</p>
<p><span class="caps">DSL</span> devices are usually type A – <span class="caps">NAT</span> routers – because the communication to the service provider is usually done via PPPoE.  It’s usually easier for customers to go to <a href="http://yeahigotdsl.com" rel="nofollow">http://yeahigotdsl.com</a> and enter a name and password than to configure PPPoE on their router or computer.  These <span class="caps">NAT</span> routers can also be told to go into a bridge mode for another router to perform the PPPoE “dial in.”</p>
<p>Cable “modems” are usually type B.  The big clue is if you plug your computer into the modem, you’ll get a public IP address via <span class="caps">DHCP</span>.  So, no <span class="caps">NAT</span> there.  The <span class="caps">DHCP</span> assignment is usually not coming from the “modem” – rather, it just serves as a bridge to the ISP’s network, which is where the <span class="caps">DCHP</span> server resides.</p>
<p>In either case, when operating in bridge mode, the only identifiers that the “modem” and device behind it have are the <span class="caps">MAC</span> addresses – the Ethernet frames are addressed this way.  That much makes sense.</p>
<p>Now our friend <span class="caps">OSI</span> layer 2 has two parts:  the top-part is the logical link control sublayer (<span class="caps">LLC</span>), and the bottom part controls <span class="caps">MAC</span> addressing.  All of a switch’s or bridge’s real magic lies in the <span class="caps">LLC</span>.</p>
<p>A regular Ethernet switch will let you unplug and connect new devices without rebooting it.  And a switch is better than a hub, because a hub is a layer 1 device that sends responses to everyone – e.g. no one-on-one communication between sender and recipient (no unicast).  So “regular” switches are good for unicast, multicast, and broadcast communication at layer 2.  This behavior is an unacknowledged, connectionless mode, which sends data efficiently to one or more recipients, but doesn’t verify that the recipients would get the data.  Verification could be left up to an upper-layer service instead.  This behavior is defined in <span class="caps">IEEE</span> 802.2, which is <span class="caps">LLC</span>.</p>
<p>Enough about a regular switch.</p>
<p><span class="caps">IEEE</span> 802.2 also defines two other <span class="caps">LLC</span> control methods, one of which is an acknowledged, connectionless mode which only supports one-on-one communication.  For the service to be acknowledged, each of the two devices in the “one-on-one chain” need to know about each other.  Since we’re still in layer 2, the only way to know one another is via <span class="caps">MAC</span> address.</p>
<p>So that’s why the router “caches” the <span class="caps">MAC</span>.</p>
<p>For the curious, <span class="caps">LLC</span> can also operate in a connection-oriented mode, where packet order is maintained.</p>
]]></content:encoded>
	</item>
</channel>
</rss>

