In a previous post, I discussed traffic coloring. Modifications to a data packet such that the packet itself informs your network equipment that it requires handling different from normal packets such as best effort traffic. In other words, if you looked at your Ethernet frame or at the IPv4 header, you'd notice a few bits different on these packets than on normal packets.
What does this look like on Ethernet? Here's a capture of two almost identical packets. Take a look at the bytes of the raw data in the first image. The first 6 bytes (in green) are for the MAC address of the receiving NIC. The second 6 bytes (in red) are for the MAC address of the source NIC. The next 2 bytes (in black) are for the Ethernet type. This tells the receiving host that the data after the Ethernet header is IPv4 traffic.
So how do you map these 802.1p values? Which priority value should you use when? That's also defined (or buried depending on your perspective) in another IEEE standard. If you scanned the end of the 1998 IEEE 802.1D standard, in Annex H, you'd find a discussion of user priorities and traffic classes. For the sake of simplicy, this is the important table from this annex:
Table H-15 — Traffic type acronyms user_priority Acronym Traffic type 1 BK Background2 — Spare0 (Default) BE Best Effort3 EE Excellent Effort4 CL Controlled Load5 VI “Video,” < 100 ms latency and jitter6 VO “Voice,” < 10 ms latency and jitter7 NC Network Control
Now we know how to color our Ethernet frames and which values to select for our packets. If only it was this easy…
- Mathias