Re: SCTP. V4 v. V6 addresses... / comparing identies
Re: SCTP. V4 v. V6 addresses... / comparing identies
- Subject: Re: SCTP. V4 v. V6 addresses... / comparing identies
- From: J P May <email@hidden>
- Date: Mon, 2 Aug 2010 21:34:39 +0200
It's not safe to assume that UDP data from peer A to peer B will
always go via the same route, and hence appear to come from the same
address.
Duh! of course... thanks for the tip
The best way to avoid this whole problem is to include a sessionID
field in the UDP packets
right..
(or use TCP :-)
indeed, it's wildly easier, there's no doubt. For tcp I have been
using the AsyncSockets library, which is trivial to use and still
very simple and lightweight and does not have any guff you don't need
The pure performance of Udp is appealing, but I did lots of tests and
really could tell no difference between udp (folliwing udpecho) v.
tcp (AsyncSockets lib). (maybe if you had 10s of devices connected).
I feel the performance of either of these two approaches (UDP from
scratch or TCP from scratch) is considerably better than using
gamekit.
putting aside the connection performance, the time-to-connect aspect
(which is really important to users) is wildly better (I have found
anyway) using one of these two approaches (ie, write a couple lines
of bonjour and then connect using AsyncSockets, or code like in
UDPEcho), compared to using gamekit overall.
If you go Bonjour+Udp, the time to connect is mind-bogglingly fast.
I mean, I suppose at some level that is how mouses and keyboards
work. It is so fast, it's great. in contrast the gamekit approach
is slow 50% of attempts, and extremely slow 45% and just doesn't work
5%.
gamekit is terrific technology but the difference is stark in
specific situations. obviously gamekit gives you bluetooth and other
glamour for free, which is great
thanks
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Macnetworkprog mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden