+- JamKazam Forums (https://forum.jamkazam.com)
+-- Forum: Jamkazam Forums (https://forum.jamkazam.com/forumdisplay.php?fid=1)
+--- Forum: Miscellaneous (https://forum.jamkazam.com/forumdisplay.php?fid=6)
+--- Thread: Distributed JamKazam Servers (Host) to improve performance (/showthread.php?tid=1211)
Pages:
1
2
Distributed JamKazam Servers (Host) to improve performance - kaminskied@comcast.net - 06-21-2020
Because of the routing behavior of the internet latency can be highly propredictable between the same locations. In fact it is possible for a connection between Europe and the US to have fewer hops than a connection between two locations 100 miles apart. Sessions on JamKazam likely connect to a jamKazam hub and that point on the network is where each user is connecting.
Perhaps Jamkazam as a company would consider licensing the JamKazam server as a product, a product that could be installed at locations closer to the users in session. This distributed server licensed to a collection of users (for a fee) could improve performance because it would be dedicated to that user group with lower latency connections.
Has anyone proposed this concept... a concept that is typical in large global corporation networks.
Wondering
(06-21-2020, 01:43 AM)kaminskied@comcast.net Wrote: Because of the routing behavior of the internet latency can be highly unpropredictable between the same locations. latecy is influced by routing In fact it is possible for a connection between Europe and the US to have fewer hops than a connection between two locations 100 miles apart. Sessions on JamKazam likely connect to a jamKazam hub and that point on the network is where each user is connecting.
Perhaps Jamkazam as a company would consider licensing the JamKazam server as a product, a product that could be installed at locations closer to the users in session. This distributed server licensed to a collection of users (for a fee) could improve performance because it would be dedicated to that user group with lower latency connections.
Has anyone proposed this concept... a concept that is typical in large global corporation networks.
Wondering
RE: Distributed JamKazam Servers (Host) to improve performance - Hans Peter Augustesen - 06-21-2020
The audio and video signals is not going though JamKazam server(s) in Texas.
They are going directly from person to person.
JamKazam is so-called peer-to-peer based.
Then it would of course make no difference to have servers all over the place, so to speak
RE: Distributed JamKazam Servers (Host) to improve performance - kaminskied@comcast.net - 06-21-2020
(06-21-2020, 12:33 PM)Hans Peter Augustesen Wrote: The audio and video signals is not going though JamKazam server(s) in Texas.
They are going directly from person to person.
JamKazam is so-called peer-to-peer based.
Then it would of course make no difference to have servers all over the place, so to speak
RE: Distributed JamKazam Servers (Host) to improve performance - Hans Peter Augustesen - 06-21-2020
If JamKazam was not peer-to-peer - and the audio and video should pass through the server(s) in Texas, USA, the total latency to my JamKazam-friend 60 miles away (in Denmark, that is) would be about 200 milliseconds. Instead of the about 22 milliseconds it is now
RE: Distributed JamKazam Servers (Host) to improve performance - kaminskied@comcast.net - 06-21-2020
Hans there is no evidence that this is a point to point network. The behavior of
This site strongly indicates that a “session” is hosted on
A Jamkazam server managed by Jamkazam. You may want to look at Dr Walkers patent diagram for
The Jamkazam architecture online. The software on your laptop is client software not a server
That session is the connection point for users to the session
And Jamkazam controls the quality of service by virtually
Connection more servers as more sessions are initiated
The concept I am suggesting. Jamkazam would license bands, jam group to host their own server
And this would likely improve latency
One of the reasons for this approach is that there are internet connection points
provided by web services companies like Amazon that
provide better performance than everyday residential
Connection points. This is the best place to host a session
The second reason for hosting the session on a server
Architecture is scaling up addition servers virtually as more sessions
and users authenticate
If the architecture is as you suggest this in itself would account
For very poor performance relying on residential connections
To wash user and no ability to scale up computer processing power
If Jamkazam licensed Jamkazam servers. It would likely
Be a Linux box with the software installed and the
Box would be licensed to a limited number of users
It becomes a dedicated session processor. This would
Put the user authentication and access under the control
If the licensed jam group. It could be installed on a residential
Internet connection like xfinity I million bps service
The proximity of users would still be a performance factor
A trial if this concept could verify the possibility for improved performance
I would like to have a conference call with Dr Walker, myself and
An expert audio engineer. I will bring the engineer to this discussion
There my jam group is a 501 C3 nonprofit and I would be willing
To sel grant funding to test a project with a distributed server
(06-21-2020, 12:41 PM)Hans Peter Augustesen Wrote: If JamKazam was not peer-to-peer - and the audio and video should pass through the server(s) in Texas, USA, the total latency to my JamKazam-friend 60 miles away (in Denmark, that is) would be about 200 milliseconds. Instead of the about 22 milliseconds it is nowHans that is correct but with virtualized servers typically
there would be more than one server connected to the internet at key
locations. There is likely a server hosted by Amazon connected
at a point in Europe. Perhaps in Denmark
performance in the internet is impeded more by
number of hops and less about the length of the wire
between hops those routing conditions are not controlled
by Jamkazam or the user. It is possible to have fewer hops across the ocean than across a city
RE: Distributed JamKazam Servers (Host) to improve performance - Hans Peter Augustesen - 06-21-2020
Yes, a session is hosted on a Jamkazam server managed by Jamkazam.
But the audio and video is not going though the server - it is going directly from person to person.
Yes, I "may want to look at Dr Walkers patent diagram for The Jamkazam architecture online"
But that would not change anything. JamKazam is peer-to-peer. It was peer-to-peer yesterday - and it is peer-to-peer today - and it is peer-to-peer tomorrow.
RE: Distributed JamKazam Servers (Host) to improve performance - kaminskied@comcast.net - 06-21-2020
Please look at the patent. Let me know what you think. There are many global functions on Jamkazam
Musicians database, session database. Those functions would have to be updated point to point
With every user if this was point to point
This behavior strongly implies a central server.
RE: Distributed JamKazam Servers (Host) to improve performance - Hans Peter Augustesen - 06-21-2020
Can you by the way provide a link or something to "Dr Walkers patent diagram"?
And who is Dr Walker?
RE: Distributed JamKazam Servers (Host) to improve performance - kaminskied@comcast.net - 06-21-2020
The patent shows how music packet management is
Controlled. Music is very low data content
Dr Walker is the co inventor of Jamkazam. He holds the patent and he serves
As their chief technology officer
Send me a note and I will send you the patent
Kaminskied@comcast.net
RE: Distributed JamKazam Servers (Host) to improve performance - Hans Peter Augustesen - 06-21-2020
Ok - thank you. That is Peter Walker, then.
What about the link to the diagram? Can - and will - you provide it?
Here, of course - it should not be needed to send a note to whatever. Simply put it here ...
"First of all, the audio is routed peer-to-peer between musicians in a session. The audio does not run through JamKazam servers. So the location of our servers is not actually relevant to your latency at all"
[...]
"Best,
David"
David Wilson is CEO of JamKazam
from >http://forum.cakewalk.com/Anything-like-Jamkazam-out-there-m3289696.aspx
Search result for peer-to-peer:
https://forum.jamkazam.com/search.php?action=results&sid=fbc820ad11ea8153fe6280a9057eb9c4&sortby=lastpost&order=desc
Packet Rate Control And Related Systems For Interactive Music Systems
http://www.freepatentsonline.com/y2015/0256473.html
"Inventors:
Walker, Peter A. (Cedar Park, TX, US)
Wilson, David J. (Austin, TX, US)
Comer, Robert Scott (Austin, TX, US)
Call, Michael Seth (Austin, TX, US)"
What is "501 C3 nonprofit"?
I could figure it out for myself, maybe.
But, of course, I do not want to spend that time - it is your job to explain. For all people here ... even the zulues, the chinese, the danes and so on