[9fans] the backporting continues - /. stories maybe of interest
matt
Network stack cloning / virtualization extensions to the FreeBSD kernel
http://www.tel.fer.hr/zec/vimage/
/. story http://bsd.slashdot.org/article.pl?sid=03/05/26/2117254
This is an old one, I didn't think it warranted it's own post
Remote Direct Memory Access Over IP
Posted by timothy on 27/04/03 18:22
doormat writes
"Accessing another computer's memory over the internet? It might not be
that far off http://www.commsdesign.com/story/OEG20030425S0043
Sounds like a great tool for clustering, especially considering that the
new motherboards
have gigabit ethernet and a link directly to the northbridge/MCH.
ron minnich
> Remote Direct Memory Access Over IP
such an old idea. Let's see, how old? Well if you don't flame me for being
involved in such terrible ideas, consider this:
- node exports NFS file system
- exporter and client mmap same file
- client stores to its mmap'ed file, does an msync, memory in
process on server changes
Voila: remote memory access over IP
OK, enhance it:
- make NFS coherent with a few mods
- put a 'fast bypass' in the ethernet interrupt handler so that, in
special cases, you bypass the nfs stack and put network data into the
process memory directly.
QED: MNFS, do a google and you'll find it.
Ok, you don't want the interrupt handler in your way, right? You just want
to dump your data into that procs memory, but still over IP over
something.
OK, google MINI for an interface we built long ago that does this too
-- over IP over ATM, and ATM was the big mistake, but we believed DARPA
:-(
It's still pretty doable. People are acting like this is the Second
Something or something. Yikes.
Plan 9 guys never made the mistake of doing this stuff. It's basically
dumb to worry about RDMA over lines that are 10s of ms in length.
Oh, yes, for real fun, google the Japanese researchers who succeeded in
putting a real NIC into the DIMM slots. Very cool.
You want to make your way in the CS field? Simple. Calculate rough time of
amnesia (hell, 10 years is plenty, probably 10 months is plenty), go to
the dusty archives, dig out something fun, and go for it.
It's worked for many people, and it can work for you.
ron
northern snowfall
>
>It's worked for many people, and it can work for you.
>
>
boyd, rounin
weird stuff with the 8000 series VAX CI and that was
some ~12 years ago. i know we [PRL] routed IP across it.
Charles Forsyth
in Plan 9 (and from the same stock, early Inferno),
the single stack was at an intermediate stage described by
some data structures rooted in a single instance of the Fs structure.
having more than one stack required allocating an Fs per stack, and
changing functions to accept Fs* where needed. a bit tedious but
otherwise fairly obvious and straightforward. user-level programs
weren't much affected because you can bind what you like on /net and
if it's somewhere else the dial string names it. you couldn't
honestly get a paper out of it.
what they did in the approach in the /. article is an interesting
application of brute force: if you replicate
the kernel environment (including user processes) you might well
replicate the IP stack but it's almost a side effect.
(i'm sure there's at least one paper in it. there's certainly at least one lesson.)
i wonder if i can link my virtual card punch to a virtual card reader
to do e-mail...
Aharon Robbins
ron minnich <9f...@cse.psu.edu> wrote:
>You want to make your way in the CS field? Simple. Calculate rough time of
>amnesia (hell, 10 years is plenty, probably 10 months is plenty), go to
>the dusty archives, dig out something fun, and go for it.
>
>It's worked for many people, and it can work for you.
>
>ron
Alas. C.A.R. Hoare made this point in his 1980(!) ACM Turing Award Lecture, which
I happened to read recently. It is chilling to see how little things have changed
in this industry in 40 years!
--
Aharon (Arnold) Robbins --- Pioneer Consulting Ltd. arn...@skeeve.com
P.O. Box 354 Home Phone: +972 8 979-0381 Fax: +1 928 569 9018
Nof Ayalon Cell Phone: +972 51 297-545
D.N. Shimshon 99785 ISRAEL
Geoff Collyer
works. (At one time, my employer sent mail over that transport, with
a sendmail at each end, and an RJE link thrown in too. Matching
technologies, I suppose.)
What's more amusing is connecting a virtual line printer (132 columns
wide) to a virtual card reader (80 columns wide). As I recall, you
get a combination of a virtual card jam and virtual printer jam. I
forget what it takes to fix it; perhaps destruction of all relevant
virtual machines, or at least virtual devices?
ron minnich
> i wonder if i can link my virtual card punch to a virtual card reader
> to do e-mail...
>
speaking of which, does anybody know if there is a program out there to
take lines of text and display them as punch cards? I don't know why I
want this, but I do.
Manila is the preferred color.
ron
ron minnich
> I've seen virtual card punches connected to virtual card readers; that
> works. (At one time, my employer sent mail over that transport, with
> a sendmail at each end, and an RJE link thrown in too. Matching
> technologies, I suppose.)
oh, all right, the ASP story. ASP was an early IBM "multi-processor" ca.
1975. ASP = Attached Support Processor.
Here is my memory of it.
They connected (at my site) 2 370/158s via a channel (the magtape channel
interface was the final low-level data path, of course). The attached
processor saw a "virtual card reader" and the main processor submitted
jobs via that channel to the ASP. At that point, of course, the ASP had
its own tape drives and printers. This was allegedly easier to admin that
two separate machines.
Virtual card readers ... via magtape channels ...
Now *that's* a kludge.
Although V5 if memory serves did something similar, I guess, I never saw
this but I did find a manual set once with a jcl(1) command.
Actually, there has been progress.
ron
Jack Johnson
> wide) to a virtual card reader (80 columns wide). As I recall, you
> get a combination of a virtual card jam and virtual printer jam. I
> forget what it takes to fix it; perhaps destruction of all relevant
> virtual machines, or at least virtual devices?
That's virtually funny....
boyd, rounin
http://www.insultant.net/repo/a2b.html
http://www.insultant.net/repo/b2a.html
iirc, i think rob said i got 'em 'round the wrong way, at some point.
small amount of sam'll fix that :)
John Stalker
>speaking of which, does anybody know if there is a program out there to
>take lines of text and display them as punch cards? I don't know why I
>want this, but I do.
>Manila is the preferred color.
>ron
How about this?
$bcd
Hello world!
________________________________________________
/HELLO WORLD! |
|]] ] |
| ]]] ]]] |
| ] ] |
|111111111111111111111111111111111111111111111111|
|222222222222222222222222222222222222222222222222|
|33]]33333]33333333333333333333333333333333333333|
|4444444444]4444444444444444444444444444444444444|
|5]5555555555555555555555555555555555555555555555|
|6666]6]]6666666666666666666666666666666666666666|
|77777777777]777777777777777777777777777777777777|
|]8888888888]888888888888888888888888888888888888|
|99999999]999999999999999999999999999999999999999|
|________________________________________________|
$ man bcd
BCD(6) FreeBSD Games Manual BCD(6)
NAME
bcd, ppt - reformat input as punch cards or paper tape
SYNOPSIS
bcd [string ...]
ppt [string ...]
DESCRIPTION
The commands bcd and ppt read the given input and reformat it in the form
of punched cards or paper tape. Acceptable input are command line argu-
ments or the standard input.
--
John Stalker
Department of Mathematics
Princeton University
(609)258-6469
ron minnich
well with seeing how easy plan9 graphics is compared to X11, I might try a
port of bcd to plan9 graphics.
After all, if the card isn't salmon, it's just not the same.
thanks for the links :-)
ron
James Sainsbury
> amnesia (hell, 10 years is plenty, probably 10 months is plenty), go to
> the dusty archives, dig out something fun, and go for it.
> It's worked for many people, and it can work for you.
Not only CS I suspect, but very true.
RGDS