Anyone done Video Conferencing with JMF?

[John]

I cant find any full applications designed for vid conf with JMF - the
closest is the sample app that comes from sun, but is no good.

Just want a Netmeeting replacement that has clean port usage - I could cope
with just vid/audio

Dont really want to write it from scratch - dont really have the time but
could help if someone's interested?

Cheers

 

[Sudsy]

I cant find any full applications designed for vid conf with JMF - the
closest is the sample app that comes from sun, but is no good.

Just want a Netmeeting replacement that has clean port usage - I could cope
with just vid/audio

Dont really want to write it from scratch - dont really have the time but
could help if someone's interested?

Cheers

<http://www.openh323.org>

 

[Mark Thornton]

<http://www.openh323.org>

H323 is not my idea of "clean" port usage. This is the protocol which
NetMeeting implements and is why it is such a pain with firewalls. Sure
you can get one machine to work inside many firewalls, but add a second
one and the troubles really start to mount.

If the two ends are inside UPnP firewalls (or NAT devices) and you don't
mind Microsoft and "Passport" then the current version MSN Messenger
works quite well. This uses a central (Microsoft) server to negotiate
the connection (transfer information on the ports opened at each end).

Mark Thornton

 

[Sudsy]

H323 is not my idea of "clean" port usage. This is the protocol which
NetMeeting implements...<snip>

I know. I was adressing the request for a NetMeeting replacement.
OP doesn't want to create it from scratch. H.323 might not be the
perfect player but it's an international standard and supported
widely. Heck, you could probably interoperate with NetMeeting if
you had to. YMMV

 

[Mark Thornton]

I know. I was adressing the request for a NetMeeting replacement.
OP doesn't want to create it from scratch. H.323 might not be the
perfect player but it's an international standard and supported
widely. Heck, you could probably interoperate with NetMeeting if
you had to. YMMV

It was designed before the widespread use of NAT. Eventually IPv6 may
remove the need for NAT, but in many current environments H323 is a
complete pain. To be sure there is gateway software which allows the use
of multiple H323 connections across a NAT firewall, but it isn't common
(and probably expensive).
Hmm, instead of replacing NetMeeting, it might be easier to write (in
Java) an H323 gateway which resolved the port usage issues and
transported the data across the firewall(s). Java could easily cope with
the data rates involved. Of course I am only guessing as to why the OP
wanted a replacement for NetMeeting.

Mark Thornton

 

[Sudsy]

Mark Thornton wrote:

It was designed before the widespread use of NAT. Eventually IPv6 may
remove the need for NAT, but in many current environments H323 is a
complete pain. To be sure there is gateway software which allows the use
of multiple H323 connections across a NAT firewall, but it isn't common
(and probably expensive).
Hmm, instead of replacing NetMeeting, it might be easier to write (in
Java) an H323 gateway which resolved the port usage issues and
transported the data across the firewall(s). Java could easily cope with
the data rates involved. Of course I am only guessing as to why the OP
wanted a replacement for NetMeeting.

I was thinking in terms of the effort involved in trying to develop a
brand new protocol from scratch. While H.323 is far from perfect, at
least they've addressed the major details. That's why I think it would
be more time/cost effective to look at something open-source which
could interoperate rather than trying to reinvent the wheel.
We all know that an elephant is a horse designed by committee but in
this case I've seen nothing else practical. Maybe if you painted some
racing stripes on it then it wouldn't FEEL like a lumbering elephant?
;-)

 

[Mark Thornton]

Mark Thornton wrote:



I was thinking in terms of the effort involved in trying to develop a
brand new protocol from scratch. While H.323 is far from perfect, at
least they've addressed the major details.

Except for usability across NAT which for many is a major stumbling
block. I can't use H323 to connect to my brother and cousin because they
have separate machines hidden behind a single IP address. Like most
consumer level firewalls it can forward the H323 stuff to only one of
the two machines.

Mark Thornton

 

[Sudsy]

Except for usability across NAT which for many is a major stumbling
block. I can't use H323 to connect to my brother and cousin because they
have separate machines hidden behind a single IP address. Like most
consumer level firewalls it can forward the H323 stuff to only one of
the two machines.

Mark Thornton

I did a quick search and found support from Cisco for their IOS routers
and netfilter.org for Linux/ipchains. There's even support from M$ in
their SecureNAT (ROTFLMAO!) software.
IOW, this is not as big of a stumbling block as it used to be just a
couple of years ago. As I said before, it's not a perfect protocol (and
the embedding of IP addresses in some protocol elements was a HUGE
mistake) but it's arguably the best we've got right now.
FYI: Both Linksys and Netgear claim to support H.323 (multiuser) via
NAT.

 

[Mark Thornton]

I did a quick search and found support from Cisco for their IOS routers
and netfilter.org for Linux/ipchains. There's even support from M$ in
their SecureNAT (ROTFLMAO!) software.
IOW, this is not as big of a stumbling block as it used to be just a
couple of years ago. As I said before, it's not a perfect protocol (and
the embedding of IP addresses in some protocol elements was a HUGE
mistake) but it's arguably the best we've got right now.
FYI: Both Linksys and Netgear claim to support H.323 (multiuser) via
NAT.

Alternatively this might be a solution: http://www.gnugk.org
There is even a Java GUI for it!
You are right, things have improved. I can't find any references on
NetGear's site for H323 for more than one client.

Mark Thornton