home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!olivea!sgigate!sgi!rhyolite!vjs
- From: vjs@rhyolite.wpd.sgi.com (Vernon Schryver)
- Newsgroups: comp.protocols.nfs
- Subject: Re: NFS packet size
- Message-ID: <u97b4fc@rhyolite.wpd.sgi.com>
- Date: 31 Dec 92 04:17:11 GMT
- References: <1hprl7INNmbj@rs1.rrz.Uni-Koeln.DE> <C01s5x.5Bn@news.iastate.edu> <1992Dec30.231852.22129@phx.mcd.mot.com>
- Organization: Silicon Graphics, Inc. Mountain View, CA
- Lines: 24
-
- In article <1992Dec30.231852.22129@phx.mcd.mot.com>, mat@phx.mcd.mot.com (Mat Cucuzella) writes:
- > In article <C01s5x.5Bn@news.iastate.edu> john@iastate.edu (John Hascall) writes:
- > >schittel@geo.Uni-Koeln.DE (Christoph Schittel) writes:
- > >}What is the maximum value for the "rsize" and "wsize" options
- > >}to make an NFS datagram fit completely into one ethernet packet
- > >}(max. 1500 byte)?
- > >
- > > 1024
- >
- > An Ethernet frame size is 1514. Subtracting Ethernet header (6+6+2), IP
- > header (20) and UDP header (20), that leaves 1460. I'm not sure if RPC
- > headers have to be included in this, but I believe 1460 or slightly
- > less will still keep everything in one Enet packet.
-
-
- After the UDP header and before the data, an NFS READ REPLY message
- tends to have about 108 bytes of RPC and NFS header.
-
- Powers of 2 tend to be nice choices for rsize and wsize because powers
- of 2 tend to be nice multiples or divisors of the "natural" buffer
- sizes disk cache in the client as well as the server.
-
-
- Vernon Schryver, vjs@sgi.com
-