| Alfonso Florio 2005-09-28, 5:51 pm |
| One of the factors of increasing latency is the cpu power and the ram of
the server.
Two servers with identical number of hops and identical ping roundtrip time
can have different latency times according to their cpu power and the fact
that the ram quickly saturates...
I usually use the .livedelay attribute. Is really useful for this kind of
tests.
Different client with different flash version can communicate without
problems, there is no change in protocol or something else.
Webcam netstreams still use the old spark codec also in flash 8.
Try recompiling. I've notice some improvements also without recompiling.
At 12.29 28/09/2005, you wrote:
>Thanks for your responses. I will try Flash Player 8 and let you know. I
>have not tried FP8 or Studio 8 yet. Do all clients participating in the
>experiment need to be on FP8? Do I need to recompile FLA using Studio 8?
> I was beginning to be convinced that it is the number of hops (there may
>not be a magic number 15, but latency would be proportional to number of
>hops). With South America we are getting 18+ hops, in the US we get 8 to 10
>hops. We are considering hosting with Akamai ... they claim to have servers
>closer to end points and claim to reduce number of hops. I was told they
>will have their FCS hosting available in a month.
> As far as firewall, proxies and other issues, I have checked with client
>machines and they have no firewall, proxy, etc. In most cases the US clients
>have firewalls, etc. And as I mentioned US to US latency is 0.5 secs.
> pramod
> Date: Tue, 27 Sep 2005 23:14:36 -0700
>From: "Edward Chan" <echan-14osZcCZf762oZ/6fjIToQ@public.gmane.org>
>Subject: RE: [FlashComm] High latency for web conference on broadband
>inSouthAmerica
>To: "FlashComm Mailing List" <flashcomm-1Ss2GqJETD3yZ38Mhd3e/9ZfFG6BLHNm@public.gmane.org>
>Message-ID:
>< F76BFE714C0B0F4787321930648EBD8501F44537
-uozdpHsGBUlVE+j733TidCaUA//5aTcQ@public.gmane.org>
>Content-Type: text/plain; charset="us-ascii"
>
>Dario is right. It could be any number of things. But FP8 fixed some
>latency issues due to a bug in the way we handle jitter. If you're not
>using FP8, it wouldn't hurt to try it.
>
>As for FCS/FMS having problems with more than 15 hops...I'm not sure
>where that number came from. But that's the first I heard of that it.
>
>Ed
>
>
>---------------------------------------------------------
>Supported by Fig Leaf Software - http://www.figleaf.com
>---------------------------------------------------------
>
>To change your subscription options or search the archive:
>http://chattyfig.figleaf.com/mailman/listinfo/flashcomm
=-----------------------------------------------------------
Supported by Fig Leaf Software - http://www.figleaf.com
=-----------------------------------------------------------
To change your subscription options or search the archive:
http://chattyfig.figleaf.com/mailman/listinfo/flashcomm
|