Richard Hartmann wrote: > Also, I am not sure if it would not be better to default to no > fragmentation and enable it optionally. I am aware that changing default > behaviour is always a bit of a problem but to the best of my knowledge > enabling fragmentation is a bug in any and all real-world applications. It worked well and was enabled by default on all the Bay Networks equipment I used ~15 years ago. And I know for certain that we tested with other gear (Ascend and Clam, probably) that did it right. If it works with the equipment you're using, it's a useful feature in that it can balance out the latencies among the links, resulting in much lower overall latency observed by higher layers -- especially so on lower-speed links where MP is more likely to be used. Without it, you're left either waiting for the one slow link choking on a big packet to catch up, or (worse) disabling the sequence headers altogether, resulting in reordering unless you're really "clever." It's a darned shame that lame implementations would force a change in the default ... -- James Carlson 42.703N 71.076W <carlsonj@xxxxxxxxxxxxxxx> -- To unsubscribe from this list: send the line "unsubscribe linux-ppp" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
