Home AMX User Forum NetLinx Studio

Open IP Sockets

Greetings,

In your experiences, how many IP sockets can be opened and held open simultaniously with no data transfer per controller? Would it matter with the vintage of the controller? Thanks.

Comments

  • ericmedleyericmedley Posts: 4,177
    This is a good question. I have systems out there that have 20-ish ports running. I would imagine it is probably a combination of port count and how chatty the devices are. Not to mention how many other network AMX devices are trying to talk to it...
  • Reason I ask is that I am experimenting with IP control of Sharp TVs. I have about 20 of them connected via WiFi to the network. I can't seem to get them all connected at the same time. There is no data going back and forth, just open connections.
  • viningvining Posts: 4,368
    Read this post by Prodigal.

    http://www.amxforums.com/showthread.php?7268-#5

    The 200 number is what I recall from training but I don't recall if that's the max number, max simultaneous, max definable?

    The other day some one posted they thought the max simultaneous ip connection were 20 but I think he intended to write 200.

    Can you watch the connections in WireShark and see who initiates the closing. I don't have much faith in Sharp products so maybe they timeout when idle. Have you tried to add connections incrementally to see when the problem begins?
  • viningvining Posts: 4,368
    Not exactly on point but this reinforces the 200 concurrent connections.
    http://amx.com/techsupport/techNote.asp?id=288
  • For the sake of experimentation, I am trying to open all of them at the same time. It seems to me that about eight of them will open while the others eventually timeout. However, the ones that timeout then come online after a retry at the onerror timeout. Once they all come online they seem to stay on.

    With no traffic, the TVs will disconnect after three minutes, as described in the manual.

    Just working with one TV, they work exactly as advertised in the manual.
  • By queing the ip_client_opens, I am able to get all my connections open simultaniously.

    It seems to me that the master can only attempt to open a small quantity of connections simultaniously.
  • jimmywjimmyw Posts: 112
    Ee sure to update to the latest firmware! At the cost of processor memory and a bit of processor time, there is a new tcp/ip stack well worth it IMO!
  • viningvining Posts: 4,368
    By queing the ip_client_opens, I am able to get all my connections open simultaniously.

    It seems to me that the master can only attempt to open a small quantity of connections simultaniously.

    I wanted to suggest opening one at a time initiating the next when the current ip_client_open triggers the online handler. IP connections can tend to bog down the processor while they negotiate their connection.
  • Well now that I got all of the sets online, your mistrust of Sharp seems to be justified.

    So here is the deal:

    Open IP connection.... Send message...... Close connection

    Everything works fine.

    However, if the open connection is broken by a network issue or a reboot of the AMX, etc, the sets will no longer respond until the IP Control function is reinitialized in the set's menu system.

    To Sharp's credit, they did excalate my tech support call to engineering. I have not heard back yet though.

    What's interesting is that a normally functioning TV will automatically close the IP connection after 3 minutes and I see the offline event from Netlinx. However, a locked up set will show an online event when the connection is made, but no three minute automatic offline event ever comes. I can close the connection though Netlinx and the offline event is triggered. Hmmm.
  • ericmedleyericmedley Posts: 4,177
    Rube Goldberg engineering...

    What happens if you pull the network plug from the TV?

    If this brings a TV's connection back alive after an AMX power cycle or network error (as you describe) then you could put the TVs network connection on a power controller and reset it from the control system.

    Kludgy, but it might work...
  • TurnipTruckTurnipTruck Posts: 1,485
    Sharp has acknowledged the problem and says that a firmware fix will be coming shortly.

    I must say that Sharp was VERY helpful. I was impressed with their responsiveness.

    Sent from my MB860 using Tapatalk
  • viningvining Posts: 4,368
    Hopefully you'll have better luck than I did when I last dealt with them. Once I was able to talk with an engineer she was very helpful, gave me a firmware fix but still no joy. Then they stopped talking to me after they realized that there was nothing else they could to make the TV work as advertised. To this day it still requires both IR and 232 for control. I haven't spec'd a Sharp since.
  • TurnipTruckTurnipTruck Posts: 1,485
    ...Giving them the benefit of the doubt for now.
  • TurnipTruckTurnipTruck Posts: 1,485
    Update

    Sharp sent a new firmware to try and the problem was resolved. They have since put that firmware on their auto-update server making it very easy for me to update all of the sets in this particular installation over their Internet connections. Sharp acknowledged the problem as being on ALL of their IP-controllable sets. New firmware has been released applying the fix to across the whole line.

    Especially given the pessimism expressed by many about about Sharp, I was very impressed by how they handled it. Within a few hours of calling their main number, I got a call back from someone who knew exactly what I was trying to describe and took the steps to get the problem fixed. Several weeks later the new firmware was made available.

    Way to go Sharp!
  • Joe HebertJoe Hebert Posts: 2,159
    Good to know. Thanks for the update.
  • mushmush Posts: 287
    How many IP sockets can be opened and held open simultaniously?

    From M2M - Design Considerations, Constraints, and Topologies

    The second constraint is the maximum number of 250 simultaneous TCP/IP connections supported by a single master. The maximum number of simultaneous TCP/IP ICSP (NetLinx device) connections supported by a single master is 200. The top ~25 of the remaining 50 are intended to be used for internal services i.e. FTP, Telnet, HTTP, etc…

    The next 25 are intended to be used for IP connections used in the NetLinx code via IP_CLIENT_OPEN, IP_SERVER_OPEN, and Duet modules.

    If there are more than 25 IP connections made from within the code they will utilize the required number of remaining 200 IP sockets which reduces the number of available socket connections and subsequently the number of available NetLinx device connections which will reduce the number of available entries within the URL List.
  • champchamp Posts: 261
    Has anybody attempted to test the maximum IP limits of a Netlinx master?

    For instance if I had to quote on a job with hundreds of IP devices then I need to do a stress test to figure out how many simultaneous devices I can connect to each master and how much data can be parsed before it all goes brown.

    What I was thinking was creating a server program on one master that can take up to 200 simultaneous IP connections and echo whatever is sent by the clients, then make a client program on another master where I can dynamically increase IP connections, string size and repeat rate.

    Next I play around until I figure out the point where I blame AMX for providing a faulty processor because it couldn't possibly be my code causing a it to crash :)

    Has anyone ever tried this, what were the results and were they consistent enough to use when designing systems?
    Does anyone have a better way of figuring out how to determine the maximum size of a system per processor?
  • TurnipTruckTurnipTruck Posts: 1,485
    The maximum number you put on a master has more to do with how chatty the devices are more so than the total sockets of the master. Your maximum number of devices will be something less than the total master sockets.

    I recently did a system with lots of IP-controlled TVs. There is very little communication to them. 50 per master worked fine.

    I would stagger the opening of the sockets. You won't be able to open them all at once. In the case of my TVs, I built a delay into my comm module that held off the openings based on the virtual device number as brought into the module. Each subsequent instance of the module held off its socket opening for two seconds.

    Sent from my MB865 using Tapatalk 2
  • a_riot42a_riot42 Posts: 1,624
    champ wrote: »
    Has anybody attempted to test the maximum IP limits of a Netlinx master?

    Does anyone have a better way of figuring out how to determine the maximum size of a system per processor?

    You can figure it out if you really wanted to break it all down by connections, throughput, events, etc. However, I don't think you want to be running systems so close to their tolerances. Typically the AMX controller is one of the cheapest pieces of equipment in a project. Hell, most TVs cost more. Trying to figure out the maximum size of a system per processor is like trying to figure out the maximum capacity in the elevator you are in. Theoretically interesting, but you don't want to be anywhere near it in practice.
    Paul
  • TurnipTruckTurnipTruck Posts: 1,485
    a_riot42 wrote: »
    Trying to figure out the maximum size of a system per processor is like trying to figure out the maximum capacity in the elevator you are in. Theoretically interesting, but you don't want to be anywhere near it in practice.
    Well said!!!
  • mpullinmpullin Posts: 949
    a_riot42 wrote: »
    Trying to figure out the maximum size of a system per processor is like trying to figure out the maximum capacity in the elevator you are in. Theoretically interesting, but you don't want to be anywhere near it in practice.
    Of course you can always find that information written in the elevator somewhere. Makes you wonder how they come up with that information though. My vote is for "put a bunch of 500-lb people in there and have them get out one by one until the elevator is able to go up."
  • PhreaKPhreaK Posts: 966
    If you do want to start breaking elevators [thread=8770]this may be of use[/thread].
  • TurnipTruckTurnipTruck Posts: 1,485
    vining wrote: »
    Hopefully you'll have better luck than I did when I last dealt with them. Once I was able to talk with an engineer she was very helpful, gave me a firmware fix but still no joy. Then they stopped talking to me after they realized that there was nothing else they could to make the TV work as advertised. To this day it still requires both IR and 232 for control. I haven't spec'd a Sharp since.

    The latest models have all the IR remote commands available as serial and IP. Maybe too late to make a believer out of you, but I thought you should know.
  • viningvining Posts: 4,368
    I'll keep this mind next time i have to spec a TV. Thanks for the update.
Sign In or Register to comment.