monkey@unixprt.UUCP (Monkey Face@unixprt) (01/01/87)
Now that System V, Release 3 has been available for 6 months, and many vendors are somewhere along the road to porting it to their systems, how about some discussion on RFS. Does it work? How does it compare to NFS? What transport providers are people using? Is anyone going to produce a transport provider for re-sale to other computer vendors? Will RFS or NFS become the de-facto standard? (NFS and 4.[23]BSD fanciers welcome in the discussion.) Monkey Face - Uni-xperts, Milpitas, CA
monte@oblio.UUCP (Monte Pickard) (01/04/87)
In article <261@unixprt.UUCP>, monkey@unixprt.UUCP (Monkey Face@unixprt) writes: > Now that System V, Release 3 has been available for 6 months,.... > ... how about some discussion on RFS. > > Does it work? How does it compare to NFS? What transport providers > are people using? Is anyone going to produce a transport provider > for re-sale to other computer vendors? Will RFS or NFS become the > de-facto standard? It appears to me that all the vendors that purchase a license to SVR3 get RFS for free with it, and it is ported to SVR3. At a cost of $25,000, plus more for later updates, license fees, and the cost of porting it to SVR3, NFS gets prohibitive fast. These things may help define the standard. Also, NFS does not implement the full UNIX file system symantics, so applications do no port as easily (or transparently). Also, follow on charges for NFS versions that will (speculative) get it up to the full file system symantics will probably cost even more. Monte Pickard - Counterpoint Computers
mo@seismo.CSS.GOV (Mike O'Dell) (01/04/87)
Sorry guys, RFS is dead meat. Until you have managed a large collection of machines, you do not understand the importance of stateless servers, filesystem semantics be damned (in point of fact, they aren't damaged very much at all, much hoopla to the contrary). Many small Unibox vendors will support RFS since they get it for free with SVR3, but there is also a good NFS port to SVR3 available, too (see Lachman associates). The simple fact is that over 150 different Unix system vendors and builders have *already* gone with NFS, and MANY of those folks ship SysV boxes. Even more importantly, with servers for other operating systems (DEC and IBM) either available or coming, the arguments are even more compelling. Which do you think will sell more computers? Being able to share files between 3B2's, or between VAXes (VMS&UNIX),IBM 3090's, PC's, Macintoshes, and a host of Uniboxes?? Rumor has it that AT&T privately admits they will probably be forced to support NFS because of its commanding lead. Believe what you wish. Faith is a terrible thing to waste. -Mike O'Dell
guy%gorodish@Sun.COM (Guy Harris) (01/04/87)
(Why is this in comp.bugs.sys5? I've redirected followups away from there.) > It appears to me that all the vendors that purchase a license to SVR3 get > RFS for free with it, and it is ported to SVR3. > > At a cost of $25,000, plus more for later updates, license fees, and the > cost of porting it to SVR3, NFS gets prohibitive fast. 1) It already *is* ported to S5R3; ask Lachman Associates about it. 2) Other people seem to have decided that it's not too prohibitive, since they've signed up for it. > Also, NFS does not implement the full UNIX file system symantics, so > applications do no port as easily (or transparently). Also, follow > on charges for NFS versions that will (speculative) get it up to the > full file system symantics will probably cost even more. Also, the only transport layer that's supplied with S5R3 is Starlan; no TCP/IP or ISO protocol implementation comes with S5R3. (Yes, NFS has been demonstrated on top of ISO protocols.) There is an MS-DOS client implementation of NFS; there is none for RFS. NFS is supported by non-S5R3 UNIX systems; RFS isn't. No, you can't access devices over NFS. (If a diskless machine's root file system is being accessed over NFS, you don't *want* to access remote devices over NFS; were NFS ever changed to permit remote device access, there would need to be an option to indicate whether special files refer to local or remote devices. If RFS were ever changed to permit diskless machines, the same would be true.) S5-style locking does work with NFS (technically, it's not part of NFS; locks are managed by a different server). Forced-append mode isn't guaranteed to work, because a forced-append write isn't idempotent. It does make a "best effort", which amounts to an "lseek" to the current EOF followed by a write (done within the kernel, so no other process on the same machine can get in between them). Some other operations also aren't idempotent, although a server-side reply cache seems to prevent this from being a problem (at least in my experience). On the other hand, I haven't seen how well stateful servers work in an environment with more servers than a few carefully-administered central machines. When I had a disk, I exported the file system containing my home directory to other machines; I didn't have to worry about re-mounting my home directory on those machines if I rebooted my machine. Another machine I manage is used for development, so it is sometimes rebooted frequently; it exports an infrequently-used file system to several clients, and it would be inconvenient if they were forced to remount that file system ever time we needed to bring up a new kernel on the server. (It's also convenient not to have my sessions smashed if a server crashes.) I would say off the top of my head that a stateful server *could* reestablish a connection when the server comes back up, but as far as I know RFS doesn't do so. In short, you pays your money and you takes your choice. I see no evidence that RFS is a clear winner, and I will avoid speculation as I don't have enough data.
mash@winchester.UUCP (John Mashey) (01/05/87)
In article <371@oblio.UUCP> monte@oblio.UUCP (Monte Pickard) writes: >In article <261@unixprt.UUCP>, monkey@unixprt.UUCP (Monkey Face@unixprt) writes: >> Now that System V, Release 3 has been available for 6 months,.... >> ... how about some discussion on RFS. >> >> Does it work? How does it compare to NFS? What transport providers >> are people using? Is anyone going to produce a transport provider >> for re-sale to other computer vendors? Will RFS or NFS become the >> de-facto standard? > >It appears to me that all the vendors that purchase a license to SVR3 get >RFS for free with it, and it is ported to SVR3. > >At a cost of $25,000, plus more for later updates, license fees, and the cost >of porting it to SVR3, NFS gets prohibitive fast. > >These things may help define the standard. > >Also, NFS does not implement the full UNIX file system symantics, so >applications do no port as easily (or transparently). Also, follow >on charges for NFS versions that will (speculative) get it up to the >full file system symantics will probably cost even more. > >Monte Pickard - Counterpoint Computers I don't think the dust has settled on this one yet. Our boxes run either 4.3BSD or SVR3, and we do a lot of OEM selling, so we hear a lot of different viewpoints. So far, it seems that: a) Most people believe that RFS has a number of technical pluses, which Monte mentions. b) Very few, if any, large computer companies have signed up yet for SVR3, given the confusion/controversy over the SVR3 licensing terms. [We have a number of very large customers who want SVR3, but whose lawyers won't let them sign the contract. Hopefully, the newest revisions will fix this.] So far, the people who've publicly supported RFS are those that have or have had special relationships with AT&T [Convergent, Monte's Counterpoint, Intel/Moto], or have other reasons not to want NFS [such as Apollo]. c) It's still not clear how well RFS will do an a heterogeneous environment, although it would certainly appear more attractive in a homogeneous environment. [This is not to say it won't be OK, it's just that NFS is clearly already OK in a heterogeneous environment.] d) At the last NFS Connectathon in December, about half of the 20-or-so vendors were there to try NFS in a SV environment. One must also give credit to SUN: they are doing a fine job of encouraging ports with good support: the Connectathon was well-run and well-supported. Anyway, there seems to be merit in both approaches, and it is just not clear yet, if only because RFS doesn't really seem to be in widespread use outside AT&T [readers: Correct me if I'm wrong.] Also, one may well draw a parallel with the XNS versus TCP/IP wars a while back: XNS had a lot of merit, but TCP/IP got there first, and in networking, it often seem to be a question of who gets there first. In any case, it appears that many people are hedging their bets by structuring their code to allow you to go either or both ways at some point. -- -john mashey DISCLAIMER: <generic disclaimer, I speak for me only, etc> UUCP: {decvax,ucbvax,ihnp4}!decwrl!mips!mash, DDD: 408-720-1700, x253 USPS: MIPS Computer Systems, 930 E. Arques, Sunnyvale, CA 94086
henry@utzoo.UUCP (Henry Spencer) (01/05/87)
> It appears to me that all the vendors that purchase a license to SVR3 get > RFS for free with it, and it is ported to SVR3. > > At a cost of $25,000, plus more for later updates, license fees, and the cost > of porting it to SVR3, NFS gets prohibitive fast. Unfortunately, licensing SVR3 is not something that can be taken for granted, since the license contains some troublesome clauses about mandatory SVID compliance that have made a lot of vendors think twice about it. -- Henry Spencer @ U of Toronto Zoology {allegra,ihnp4,decvax,pyramid}!utzoo!henry
monkey@unixprt.UUCP (Monkey Face@unixprt) (01/07/87)
In article <7478@utzoo.UUCP>, henry@utzoo.UUCP (Henry Spencer) writes: > Unfortunately, licensing SVR3 is not something that can be taken for granted, > since the license contains some troublesome clauses about mandatory SVID > compliance that have made a lot of vendors think twice about it. It seems appropriate that if you have a UNIX port based on SVR3, that at least one of the environments available to the user is SVID compliant. I do not believe that this this disallows having other 'known' environments available. Application writers need to have a finite set of environments to address their work to, else so many things will only run on very specific environments (like now). Several vendors are providing this multi-environment solution already. Monkey Face - Uni-xperts, Milpitas CA (408) 946-6572
monkey@unixprt.UUCP (Monkey Face@unixprt) (01/07/87)
> ...Also, one may well draw a parallel > with the XNS versus TCP/IP wars a while back: XNS had a lot of merit, but > TCP/IP got there first, and in networking, it often seem to be a question of > who gets there first. .... I believe, in the past, there was more wide-spread use of the Berkeley rendition of UNIX (4.xBSD) outside of AT&T also. Lately it seems that most vendors feel obligated to provide SV. Many do as MIPS and provide both. Few still only provide the Berkeley based product. Monkey Face - Uni-xperts, Milpitas CA (408) 946-6572
mishkin@apollo.uucp (Nathaniel Mishkin) (01/10/87)
In article <43052@beno.seismo.CSS.GOV> mo@beno.CSS.GOV.UUCP (Mike O'Dell) writes: > >Sorry guys, RFS is dead meat. Until you have managed a large collection >of machines, you do not understand the importance of stateless servers, >filesystem semantics be damned (in point of fact, they aren't damaged >very much at all, much hoopla to the contrary). This is an interesting (and probably correct) argument. But it's not clear that people will understand or act upon it. For example, I could state: Until you have managed a network of 100+ machines, you do not understand the importance of not having to explicitly mount a file system before you can access it. This argument has not always had the effect I might have wished :-) -- Nat Mishkin Apollo Computer Inc. apollo!mishkin
HN1@PSUVM.BITNET (01/19/87)
qquit qquit QQUIT