Re: Do we need a reliable multicast in kernel?

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Tue, 21 Oct 2003 13:29:43 +0800 (CST)
"Zhu, Yi" <yi.zhu@intel.com> wrote:

> On Tue, 21 Oct 2003, David S. Miller wrote:
> 
> > People on clusters use their own special clustering hardware and
> > protocol stacks
> 
> And their own operating systems...

I've stated publicly on many occaisions that if OS's other than
Linux solve someone's problem better than Linux does they should
go use it.

There is nothing wrong with this, and it should not in any way decide
how or what we implement in the Linux kernel.  In fact I invite other
systems to facilitate small groups of users in ways that would be bad
for Linux in the long run.

This is what Sun has always done, and they enjoy an ever shrinking
market share.  This is therefore not what Linux will do.

> Reliable mulitcast is a very basic requirement from all the cluster
> softwares and IP is so widely used. I don't think every cluster
> implementation needs to write its own protocol stacks from scratch
> except those for rigorous network performance.
> 
> If Linux can provide one, I bet a big number of people (at least for
> Linux cluster people) will use and take advantage of it, if it is
> written good enough. :)

The things cluster people want is totally against what a general
purpose IPV4 implementation should do.  Linux needs to provide a
general purpose IPV4 stack that works well for everybody, not just
cluster people.

I'd rather have millions of servers using my IPV4 stack than a handful
of N-thousand system clusters.

If your company had the choice between making their CPUs useful to a
handful of technical people or the millions upon millions of Microsoft
users writing Word documents, which way would you prefer them to choose?

Next, how is that choice any different from the choice I'm telling you
we're making for the Linux networking here?

Sure, many people would like to simulate the earth and nuclear weapons
using Linux, but I'm sure as hell not going to put features into the
kernel to help them if such features hurt the majority of Linux users.
-
: send the line "unsubscribe linux-net" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[Index of Archives]     [Netdev]     [Ethernet Bridging]     [Linux 802.1Q VLAN]     [Linux Wireless]     [Kernel Newbies]     [Security]     [Linux for Hams]     [Netfilter]     [Git]     [Bugtraq]     [Yosemite News and Information]     [MIPS Linux]     [ARM Linux]     [Linux RAID]     [Linux PCI]     [Linux Admin]     [Samba]

  Powered by Linux