sharp@cpsc.ucalgary.ca (Maurice Sharp) (12/12/89)
In article <15103@joshua.athertn.Atherton.COM> joshua@Atherton.COM (Flame Bait) writes: >> = Maurice Sharp from sharp@ksi.cpsc.ucalgary.ca >>> = my eariler posting >I look at RPC as a sort of temporary kludge, which should go away in three >to ten years. Right now programmers, compilers, designers, and everything >else is set up for sequential execution on one machine. RPC alows us to >should replace RPC. Unfortunately, the temporary kludge will cost in total re-design when the other distributed paradigms come into their own. Besides, the Mach system from CMU (a new UNIX, used on the Next and a few others) is a message passing operating system. >> Actually, Apollo NCS IS faster than the Suns. I have the source, >>compiled NCS and NIDL on the Suns, and it WORKED first time. NO bugs >>(unlike NFS for machine X). Not only that, it took 3 Sun 4 file >>servers to equal the speed of 6 DN3000's for the Mandel demo. >Apollo used Sun's NFS. Your last sentence seems to say that Sun's are >twice as fast as Apollos, so I'm sure you did not mean it :-) NO !! Actually, it says that it took 3 Sun 4 SPARC file servers to meet the speed available from 6 DN3000's (68000's (not 020) 4 megs memory). The DN3000 is slower than a Sun 3/50. The Suns are SLOWER because they use TCP communication on a 10 Mb/s ethernet. DDS protocol on a Ring (or an ether for that matter) is much faster than TCP. >I assume that "faster" refers to installation time, not run time speed. >The current Apollo RPC system is 4 TIMES slower than Sun RPC system >for 8K packets, which happens to be the size my application uses. >That is using Sun/TCP (for reliablity as good or better than Apollo). >If I use Sun/UDP, it is 8 times faster than Apollo, but not as >reliable. See above. I do not know how you built your stuff, but the Sun RPC system is much slower than NCS. >I do not want to talk about interface langauge syntax (everyone has their >own favorite), but I will say one sentence: I thought that both NIDL's >PASCAL syntax and its C syntax looked like PASCAL, and were almost identical. >This is personal taste, however, just look at example code from all three >to see which one you like better, RPCGEN, NIDL(C), or NIDL(PASCAL). Actually, the C syntax is very close to ANSII C. This is a good thing. [bunch of stuff about NCS using only one transport protocol deleted]. NCS so far used 2 protocols concurrently. DDS on the Apollos and TCP/UDP on anything else. I have had systems running over the Suns and Apollos using DDS on the Apollos and the IP protocol on the Suns. They are also working on other protocols (PUP, ...). I guess I am biased. The Sun NFS system has been nothing but a buggy pain in the ^&*. I have constant problems with NFS fileserver not responding. And when that happens watch loads go from 2 -> 75 in a few minutes. The Apollos, no problem. A node goes down, no probs. If my files were not on it, I am OK. A server goes down on the Sun network and goodby network. Like a machine I did not know about mucks me up. Apollo NCS (and dds) is far superior technology to anything Sun has. BUT Sun is out there, they will push out half ready code (NFS) to get a user base, then fix it later. NCS will be ready eventually, but reasonably right. I am not saying Apollo is perfect, they have had their own cockups. But not in the name of grabbing the market by fooling the user community. I guess I should end with a black prediction. Right now I think Sun will win the market wars. They have inferior enginerring technology on their workstations, they have an inferior RISC (SPARC), they have a bastardized UNIX (sys5 with 4.3 extensions), but they have the glittzy catalogues, the technical reports, the cheap desktop power, the applications, the customer base, the SPARC hardware license, the road shows, and the good anti-Apollo propaganda. Apollo has superior engineering technology for workstations, superior OS, MUCH better file sharing, much better compute serving, superior networking. In other words, much better engineering, design, and thought (The network is really the computer, unlike Sun). What they do not have is marketing, tech support, technical reports, glittzy catalogues, road shows, cheap prices, decent sales support, decent backing from HP, etc... What I see is networks of Suns with some Apollos using NFS. Apollo better get their act together soon, or Sun will be the standard. maurice Maurice Sharp MSc. Student University of Calgary Computer Science Department 2500 University Drive N.W. sharp@ksi.cpsc.UCalgary.CA Calgary, Alberta, T2N 1N4 ...!alberta!calgary!sharp