It's been years since I've dealt with traffic control in Linux. It's so awesome there's a mailing list now! I was hoping to get some guidance from people who have been doing this lately, it's a lot to swap back into the brain. Too bad the human brain doesn't run as well as Linux, it's more like Windows 95 or something. The object is to make sure no single connection gets starved under load, and also to give some services pseudo-reserved (or maybe just prioritized) bandwidth. For example, if the total available network bandwdith is 9Gb, I'd like virtual machines to be able to consume the entire 9Gb if nothing else is going on. But as soon as some filesystem replication needs to happen, I'd like to dedicate a good 4.5Gb or so to the replication (if it needs it), and the ot her services and VMs can divide up the rest. Most of these setups have a single 10Gb ethernet interconnect to a single 10Gb ethernet switch. Some use 1Gb interconnects, and future ones will have mul tiple NICs. I realize that since the upstream bandwidth is shared and the system can't control the queue, this has to be on a best-effort basis. But I think doing so mething is much better than doing nothing. Ideas/recommendations? -- To unsubscribe from this list: send the line "unsubscribe lartc" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
