From mboxrd@z Thu Jan 1 00:00:00 1970 Message-ID: <600308d605091014365118a95@mail.gmail.com> Date: Sat, 10 Sep 2005 23:36:05 +0200 From: Francisco Ballesteros To: Fans of the OS Plan 9 from Bell Labs <9fans@cse.psu.edu> Subject: Re: [9fans] reliability and failing over. In-Reply-To: Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable Content-Disposition: inline References: <600308d605090916023b856b47@mail.gmail.com> Topicbox-Message-UUID: 8782fe32-ead0-11e9-9d60-3106f5b1d025 Funny, I admit. We saw the problem when combining multiple user-level file servers (not fossil, but device drivers for X10, guis, and the like) within the same dir, specially on wireless at home. Issuing requests in parallel as Russ suggested will help, for sure, but if things dont get too bad, the simplest user-level file server will be a good option= . We just happen to have the kernel changed, because I wanted to be able to s= ay "no overhead, but for the unions", which required a kernel to measure. That=B4s why I was a considering the kernel changes besides the user level alternative. = Guess I have an opportunity to experiment with a few things now that it=B4s going= to be a user program :-) On 9/10/05, Sape Mullender wrote: > > Beware of the latency. > > > > Plan B usage shows that latency is the biggest problem. >=20 > That's funny. We've been working on a low-latency wireless comms protoco= l > here at the Labs that we've called Plan B (our wireless emulator runs on = Plan 9, > of course), because we'd come to the conclusion that latency is the bigge= st problem > in current cellular systems. >=20 > Sape >=20 >