Using zsync for .deb downloads: initial benchmark results

[Martin Pitt]

Hello Lars,

first, thanks for doing those benchmarks, very insightful.

Lars Wirzenius [2009-07-14 19:19 +0300]:
> Before I continue working on this, I'd like to have some feedback on
> this: is a 25% reduction in download sizes worthwhile to pursue?

Just to be completely sure that I understood this correctly: if I have
package foo 5ubuntu1, and then a security update (which changes two
lines of code) 5ubuntu2 comes along, then I need to download 75%
of 5ubuntu2 with zsync'ing against -5ubuntu1?

To be honest, this looks like a very unexpected failure to me. I had
expected something like a 90% saving (i. e. only downloading the 10%
that actually changed), similarly to what we are used to with daily
karmic CDs (which change a lot, and yet an 1:10 download:size ratio is
realistic). 

For OO.o or kernel-sized updates which are in the 50 MB range, it
doesn't feel significantly different whether I have to download 50 or
40 MB, it's a pain either way.

Any idea why this is so bad? rsync has an intelligent server and thus
can do very fine grained blocking. Perhaps zsync, being limited by
HTTP, tries to weigh round trips against block size, or something like
that?

Thanks,

Martin

-- 
Martin Pitt                        | http://www.piware.de
Ubuntu Developer (www.ubuntu.com)  | Debian Developer  (www.debian.org)
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: Digital signature
Url : https://lists.ubuntu.com/archives/ubuntu-devel/attachments/20090714/fa231baa/attachment.pgp