tag:blogger.com,1999:blog-995740623166720236.post4043022744001956907..comments2024-03-18T05:38:07.579+00:00Comments on use case driven: Low latency framework: the must read presentation of Richard CroucherThomas PIERRAINhttp://www.blogger.com/profile/10273021421292279417noreply@blogger.comBlogger1125tag:blogger.com,1999:blog-995740623166720236.post-32153477623507030672011-10-25T16:56:18.255+00:002011-10-25T16:56:18.255+00:00Very interesting, but some details bothers me :
S...Very interesting, but some details bothers me : <br />Slide 11 <br />"Stretched VLAN can be used to remove router insertion latency cost"<br /><br />This is not really true as when using extended VLANs, L2 packets are encapsulated in L3 ones.<br />So there will always be routers on the network path, it's transparent, but it's not direct switch to switch communication.<br /><br />Slide 21<br />I heard that noop scheduling isn't the best.<br />To improve latency, the deadline scheduler is recommended, as it's based on service time concept rather than FIFO queue.<br />Good article to read ; http://www.redhat.com/magazine/008jun05/features/schedulers/<br /><br />Anyway, it's really hard to identify I/O workload on systems, so there's no perfect scheduler because processes are not behaving the same way. Would be great to have OS I/O and Applicative I/O all seggregated (schedulers, interruptions, storage) :)Jonathan Skrzypeknoreply@blogger.com