Skip to main content
Known Participant
November 1, 2010
Question

Edge Origin Advantages on live apps.

  • November 1, 2010
  • 1 reply
  • 799 views

Hi .. Can someone advice how would edge origin be advantageous in terms of network latency for a live (publisher / consumer) application ( I guess live apps wont need caching of data in edges)

user1 ---------------   ASIA````````````````````|

                                                  USA

                                                       |

user2 --------------   EMEA ```````````````````

                            (Edges)               (Origin)

How the above is advantageous from the below one.

user1 --------------------------````````````````````|

                                                  USA

                                                       |

user2 --------------------------````````````````````

In both the cases, requests are reaching the same server  (USA) where the application is running.

( I know if I have multiple edges behind a load balancer it ll help in failover. But how about network latency?)

2) Can I make it so that if the ASIA server dies, User1 would automatically be connected to EMEA server? ( I wasnt able to understand the client autodiscovery)

3) I wanted to have a fail over of origin server.   I went through   http://www.adobe.com/devnet/flashmediaserver/articles/origin_redundancy.html   But this is too complicated. Is there any other simple way for Origin's Failover ?

4) Is it possible to have two origin server in active-active mode ?

5) If for live applications, Edges are not going to process / store much things, can I use a low end machine for edges and a very high end machine for Origin?

Please help .. Thanks in advance...

    This topic has been closed for replies.

    1 reply

    Participant
    November 1, 2010

    Hi,

    1) Whether or not you use Edge/Origin won't really affect your latency in a noticable way.

    2) You can program your flash client to listen for a disconnect event (NetConnection.Connect.Closed or NetConnection.Connect.Failed or something). Once your client recieves a disconnect event, you can have it automatically attempt to connect to another server.

    3) A simple ways to have your servers load balanced and redundant could include:

    - using a load balancer

    - if you don't have many servers, have each client round-robin connect to your FMS cluster. If a connection fails for any reason, simply reconnect until you hit an active server

    - you can use "DNS load balancing"

    4) i'm not sure what you mean by active-active mode

    5) Yes, if your edges won't be under much strain, you could use a server that isn't as beefy as your Origin server.

    Hope this helps,

    Jason

    Known Participant
    November 6, 2010

    Hi .. Sorry for my delayed response ...

    1. Does that mean that edge/origin shows no improvement in performance. I understand that i can provide failovers of edges.

    or

    if they have the same performance, what is the use of,

    publisher       -     edge1

                                            origin ( fms server)

    consumer      -     edge2

    over

    publisher

                                       fms server

    consumer

    3) Sorry I couldnt understand fully.

    How do i setup

                                  ,------ Origin 2 ( will take over if Origin1 fails )

                   Origin1----`                (How to set up if Origin1 dies Origin2 will takeover)

         edge1           edge2

    4) By active-active I meant, if its possible to have two Origin servers up  and running and both of them serving. If any one of the Origins goes down the other Origin takes over. Is it possible ?

    One more question,

    Are there any tools to measure the performances of Live AV Chat or any load test tools for FMS applications like whiteboard ?

    Thanks

    Karthik