Category: Physical layer

If a computer on local LAN network is turned off and administrator needs to do some regular maintenance on it, he will need to use Wake-On-LAN (WoL) to power the system up remotely.

Of course, network devices need to be configured to enable that kind of “magic” packet forwarding.

NIC cards on machines need to support WoL for this to work, but we don’t bother with this here..

WoL is sending “magic packets” to computer NIC card in order to start the system up. NIC which supports WoL is still receiving power when PC is turned off. NIC then keeps listening on the network for the magic packet and if received it will initialise the system boot process and power up the PC.

Magic packet is specially crafted network directed broadcast packet typically sent with connectionless UDP, port 7.

You would usually have a WoL server somewhere on you network which will be used to source magic packets. If you send magic packets across network segments (between VLANs or from some remote subnet), last router in the path, one having client subnet locally connected, needs to be configured with directed broadcast. The first router on the path, router with server subnet locally connected, should have ip helper configured pointing to directed broadcast IP address (in our case 172.19.1.255).

In our example below, both ip helper and directed broadcast are configured on the same L3 device since this is the only router connecting two subnets.

Directed broadcast on Cisco devices is off by default since IOS 12.0 and needs to be configured on specific subnets where WoL will be needed.

You need directed broadcast because PC which needs to be woken up is asleep and while asleep it will not have an IP nor it will respond to ARP. Only way to get some packets to that PC without an ARP resolution is by using local subnet L2 broadcast.

Furthermore, we can surely assume that your PCs are connected to L2 Access Switch. That switch will not know to which port is the PC connected while that PC is asleep. Only a Layer 2 broadcast (and unknown unicast) will be sent out all ports on a switch.

As this is networking blog I will focus mostly on the usage of CAM and TCAM memory in routers and switches. I will explain TCAM role in router prefix lookup process and switch mac address table lookup.

However, when we talk about this specific topic, most of you will ask: how is this memory made from architectural aspect?

How is it made in order to have the capability of making lookups faster than any other hardware or software solution? That is the reason for the second part of the article where I will try to explain in short how are the most usual TCAM memory build to have the capabilities they have.

CAM and TCAM memory

When using TCAM – Ternary Content Addressable Memory inside routers it’s used for faster address lookup that enables fast routing.

In switches CAM – Content Addressable Memory is used for building and lookup of mac address table that enables L2 forwarding decisions. By implementing router prefix lookup in TCAM, we are moving process of Forwarding Information Base lookup from software to hardware.

When we implement TCAM we enable the address search process not to depend on the number of prefix entries because TCAM main characteristic is that it is able to search all its entries in parallel. It means that no matter how many address prefixes are stored in TCAM, router will find the longest prefix match in one iteration. It’s magic, right?

Image 1 shows how FIB lookup functions and points to an entry in the adjacency table. Search process goes through all entries in TCAM table in one iteration.

Router

In routers, like High-End Cisco ones, TCAM is used to enable CEF – Cisco Express Forwarding in hardware. CEF is building FIB table from RIB table (Routing table) and Adjacency table from ARP table for building pre-prepared L2 headers for every next-hop neighbour.