Linked by Thom Holwerda on Sat 17th Dec 2005 23:44 UTC, submitted by Lazarus
BSD and Darwin derivatives Fans of DragonFly BSD will be getting their Christmas present late this year, and plans for 1.5 have been announced. MP safe networking code, the long awaited cache coherency management system, and a port of Sun's ZFS. Read here for more. Update: Refresh, empty cache, whatever, and check the shiny new beastie icon! And there was much rejoicing. Can we now please discuss DragonFly BSD?
Thread beginning with comment 75315
To view parent comment, click here.
To read all comments associated with this story, please click here.
RE: Work Continues
by on Mon 19th Dec 2005 06:17 UTC in reply to "Work Continues"

Member since:

In what ways would it be "technically superior" to other operating systems?

Clustering, clustered filesystems, etc have been around
for quite a long time and operating systems like OpenVMS
and Linux are quite good at them.

SSI shared memory clusters are possible today with Linux
using a page based coherency protocol. I don't think
they're particularly practical because non-transparent
cluster software is really mature now and one actually
needs to program for a cluster in a non-transparent
way (far more important than even a slow NUMA) to get
decent performance.

However I'm not sure about what sort of SMP scalability
DFBSD would have if all code was fully MP safe using
LWKT as you say, nor advantages of XIO or ZFS.

Reply Parent Score: 1

RE[2]: Work Continues
by Ronald Vos on Mon 19th Dec 2005 14:10 in reply to "RE: Work Continues"
Ronald Vos Member since:
2005-07-06

Now the iconbusiness is out of the way, let's discuss.

ZFS no advantages? ZFS has numerous advantages over older filesystems, all to do with data-coherency, capacity and speed. Being one of the first BSDs to implement it would be a good thing.

And if the LWKT model works out, it would provide an excellent basis for SSI, wether or not Linux does it as well.

Reply Parent Score: 2

RE[3]: Work Continues
by on Tue 20th Dec 2005 09:14 in reply to "RE[2]: Work Continues"
Member since:

ZFS is not an *advantage* as such because it is available
on another system.

Sure it may be an advantage for DFBSD + ZFS versus Linux + something worse, but if you're talking about innovation
and advantages of DFBSD project then ZFS is not so interesting.

Now if you can do something with DFBSD + ZFS that you
cannot do with Solaris, that too would be interesting.

I fail to really see how LWKT (lightweight kernel
threads?) is a basis for SSI...

Reply Parent Score: 0

RE[2]: Work Continues
by on Mon 19th Dec 2005 14:25 in reply to "RE: Work Continues"
Member since:

Well that is the the reason (cache locality) that Light Weight Kernel Threads were developed. The processes themselves are per CPU and are migrated (and communicate via a light weight messaging protocol.) This suits itself to clustering because you have an increased L2 utilization and a means of migrating processes to systems with heavy loads to other non-taxed systems in a cluster. Cache coherency systems are not new, however they are new to this system and because of the way the system is developed it has a potential to perform and scale really well in a clustered environment, talk on ZFS which Matt Dillon thinks will work very well in such an environment and you may end with something that works VERY well.

Reply Parent Score: 1

RE[3]: Work Continues
by on Tue 20th Dec 2005 09:22 in reply to "RE[2]: Work Continues"
Member since:

Per-CPU scheduler and CPU affine processes which migrate
between CPUs are nothing new. Nor would it particularly
be of help for a cluster because migrating processes
between CPUs with a per-cpu scheduler is by no means
equivalent to migrating processes between nodes in a
cluster.

Now if the cluster is SSI on a page level (like Linux's
virtual iron capability AFAIK), then you may not have to
explicitly migrate processes but they could migrate
naturally according to the scheduler. This also seems
like a pretty bad idea unless the scheduler is very
specialised, because you could have a process with local
memory scattered all over your cluster.

"and because of the way the system is developed it has a potential to perform and scale really well in a clustered environment"

Well yes, we are talking about software cache coherency
schemes for SSI clusters. So what exactly is it about
the way this system is developed that gives it this
potential?

"talk on ZFS which Matt Dillon thinks will work very well in such an environment and you may end with something that works VERY well."

I recall him saying ZFS is a clustered filesystem or can be used over a cluster. This is not the case.

Reply Parent Score: 0