[vbox-dev] SPICE vs VRDP performance?
Klaus.Espenlaub at Sun.COM
Thu May 20 17:41:33 GMT 2010
Huihong Luo wrote:
> Can anyone shed some lights about the performance between kvm's SPICE
> and VRDP protocols? If SPICE is ported to vbox, would it have better
No one is able to tell what performance it could deliver. When I had a
look at the sources I immediately thought that it won't be easy to
interface it efficiently with the way VirtualBox works. This was quite
some time ago, and didn't feel the urge to check again since.
When I last had a peek at SPICE it was to a large proportion a list of
cool features which were not 'yet' implemented. This might have changed
by now, however I never understood why the world needs yet another
protocol. RDP could easily be extended to accommodate most SPICE
specific features only if someone would see the benefit - it is
happening actually. MS does keep the RDP specs up to date, and is adding
useful new features. The basic design (from the ITU T.120 set of
standards) makes it very flexible and extensible. Lots of clever people
contributed to this.
> According to this video, SPICE seems to consume much more network
> bandwidth than Microsoft RDP.
I've seen the benchmark videos, and I still don't know what value to
attribute to them. They measured something, but so far I couldn't make
up my mind if it is releavant or not. Maybe it's the presentation form
they're chosen which I don't like.
> The reason I asked this question, is to explore whether it's worth the
> efforts to make SPICE to work with vbox.
Cannot really comment whether it's worth doing.
I was immediately and completely put off by the licensing nightmare the
developers created there. Different licenses all over the place, lots of
files without even the faintest hint who wrote it, leave alone having a
license statement. This looks to me like a time bomb. In that form
there's no way to get this through the legal review process of a
More information about the vbox-dev