| |
|
 |
|
 |
|
|
|
| |
|
|
xen-users
Re: [Xen-users] A few newbie questions about Xen devices - virtual netwo
practically, the earlier answer about using a bridge would suffice. But
if you are looking at MPI kind of loads or aiming for better
performance, i suggest you take a look at the Xen-Socket proposal
(possibly a code exists out there too), that was presented in the Xen
Summit April 2007 by a researcher from IBM .
--It had excellent results compared to a standard tcp/ip communication
mechanism between two DomUs in the same physical host.
Jeremie Le Hen wrote:
Hi there,
I'm new to Xen. I've read thoroughly Xen design papers [1] [2] [3], but
now I'm experimenting it, I have a couple of questions in order to match
my practical experience with the idea of Xen design I've made up from
the papers.
So here is what I've understood. Please correct me if I'm wrong, it
may be useful for future newcomers. There are a few questions along the
text flow.
Hints for other newbies: PV = para-virtualized; FV = full-virtualized.
According to slide 49 in [4], Isolated Driver Domain (IDD) have only
been implemented experimentaly but is not used at all the current
official releases of Xen 3. Are there any plans about this?
Currently, Virtual Block Devices (VBD) and Virtual (Network) Interfaces
(VIF) are the most common way to provide storage and network devices
within Xen PV guests. It is yet possible to assign exclusively a PCI
device to any one DomU, in which case the DomU's driver talks to the
hardware through the "Safe Hardware Interface" (see figure 1 in [2]),
enforced by I/O Spaces, as described in [2] section 4.
How are other devices (simple serial console emulator, VGA emulator,
...) implemented in PV guests? I mean where does the console output
make its way into my pty? Is it possible to emulate a full VGA console
somehow? How can a PV guest display graphics, when running X for
instance? Any documentation pointer would be welcome.
What's the goal of the "ioemu" keyword used in FV guest configuration?
I couldn't find any documentation about all this. My VBD seems to work
both with or without it. Does this mean that Xen should use IOMMU
(Intel's VT-d) for this device, if present? Or does it mean it should
use QEMU as backend? Again, any documentation pointer will be welcome.
What's the benefit of using QEMU backend for a disk, instead of using
a VBD? BTW, I know that QEMU can emulate multiple ISA or PCI network
adapters, is it possible to use them as for FV guests? What would be
the benefit over a VIF?
BTW, are there any chipset that currently provide VT-d? If yes,
according to [2] section 8, this shouldn't be a big deal to implement it
in Xen, so is it supported in Xen 3?
Thank you.
Best regards,
[1] http://www.cl.cam.ac.uk/netos/papers/2003-xensosp.pdf
[2] http://www.cl.cam.ac.uk/netos/papers/2004-oasis-ngio.pdf
[3]
http://www.intel.com/technology/itj/2006/v10i3/3-xen/4-extending-with-intel-vt.htm
[4] http://www.cl.cam.ac.uk/netos/papers/2005-xen-ols.ppt
_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users
|
|
|
| |
|
Home