Clients of ejabberd XMPP server are receiving messages too slowly - performance

I have a ejabberd server and a few clients connected to it. They normally exchange messages and etc.
Clients are not the real people but the software.
It's all great when they just start working, but then they need to exchange with huge amounts of text data (for example 4K characters) and this is where problem is starting to appear. Each new message is sent to the server in ~10ms but receiving client gets it after ~10-40 seconds!
I tried to increase serversided limits (to 1 million) and it didn't help.
What can I do and what's the problem could be?

Related

WebSocket client disconnect due to network loss doesn't get intercepted by Spring server

I have an application in which clients use websockets to connect to a server which is running Spring Boot Tomcat.
My question is if there is a way for the server to detect a client disconnect due to a network loss.
Thanks.
if you are using stomp , check SessionDisconnectEvent.
For raw Websocket connections, you can use :
WebSocketHandler-->afterConnectionClosed
I have searched before for this and the solution I was able to find was to implement a ping-pong mechanism between the server and the clients.
For example, each few seconds send a dummy message to the client on a specific topic and receive back another dummy reply, if you didn't get a reply for a configured period you can consider the client disconnected.
As mentioned here,
STOMP and Spring also allow us to set up topics, where every
subscriber will receive the same message. This is going to be very
useful for tracking active users. In the UI, each user subscribes to a
topic that reports back which users are active, and in our example
that topic will produce a message every 2 seconds. The client will
reply to every message containing a list of users with its own
heartbeat, which then updates the message being sent to other clients.
If a client hasn't checked in for more than 5 seconds (i.e. missed two
heartbeats), we consider them offline. This gives us near real time
resolution of users being available to chat. Users will appear in a
box on the left hand side of the screen, clicking on a name will pull
up a chat window for them, and names with an envelope next to them
have new messages.

How to drop inactive/disconnected peers in ZMQ

I have a client/server setup in which clients send a single request message to the server and gets a bunch of data messages back.
The server is implemented using a ROUTER socket and the clients using a DEALER. The communication is asynchronous.
The clients are typically iPads/iPhones and they connect over wifi so the connection is not 100% reliable.
The issue I’m concern about is if the client connects to the server and sends a request for data but before the response messages are delivered back the communication goes down (e.g. out of wifi coverage).
In this case the messages will be queued up on the server side waiting for the client to reconnect. That is fine for a short time but eventually I would like to drop the messages and the connection to release resources.
By checking activity/timeouts it would be possible in the server and the client applications to identify that the connection is gone. The client can shutdown the socket and in this way free resources but how can it be done in the server?
Per the ZMQ FAQ:
How can I flush all messages that are in the ZeroMQ socket queue?
There is no explicit command for flushing a specific message or all messages from the message queue. You may set ZMQ_LINGER to 0 and close the socket to discard any unsent messages.
Per this mailing list discussion from 2013:
There is no option to drop old messages [from an outgoing message queue].
Your best bet is to implement heartbeating and, when one client stops responding without explicitly disconnecting, restart your ROUTER socket. Messy, I know, this is really something that should have a companion option to HWM. Pieter Hintjens is clearly on board (he created ZMQ) - but that was from 2011, so it looks like nothing ever came of it.
This is a bit late but setting tcp keepalive to a reasonable value will cause dead sockets to close after the timeouts have expired.
Heartbeating is necessary for either side to determine the other side is still responding.
The only thing I'm not sure about is how to go about heartbeating many thousands of clients without spending all available cpu just on dealing with the heartbeats.

Winsock Array with data arriving simultaneously stops sending data to clients

I've been searching online for days now and I can't find anyone who has this same problem with the VB6 Winsock. So here's my problem, I have a server with two winsocks in an array. I then have two clients each with one winsock control. Now the I have it set up is that the server first sends data to client A. Then client A receives that data and sends its own data back to the server. Then the server sends data to client B, once client B receives the data, it sends its own data back to the server. So data is being sent from a client to the server, then from the server to another client, then back to the server, and then on to another client. I did it this way because when I tried doing it the way I really want it to work, I was having problems. So I had to set it up this way in order to ensure that only one client is sending data to the server at any one time. This process works perfectly, however I want it to work in a different way, since as you can image, the more clients there are, the longer the delay in data transfer between each client and the server.
So what I really want to do, which I can't get to work, is have the clients send and receive data to the server whenever they want. That is, client A sends data to the server, and then the server sends data back, all while client B is doing the same thing. When I do this, even though I have a winsock array on the server, I run into a problem. When client A first connects, it begins sending and receiving data to the server. But once client B connects to the server, all communication between client A and the server stops, and only client B sends and receives data to the server. Now I've done some tests and client A remains connected to the server the whole time. But for some reason, it seems that if two clients send data to the server simultaneously, only one data arrival event fires, even though each winsock is on a different port. I have not installed VB6 SP6 yet, as I'm not sure this will fix the problem.
So I really do hope someone will read this and explain to me what it is that I'm not understanding or what it is that I'm doing wrong.
If you search for Microsoft KB articles on the Winsock control you will find a long history of flaws and bug fixes. There is absolutely no reason not to install SP6 before even attempting to use VB6, since a vast number of issues were resolved over time.
Once you've done that (and only then) is it really worth talking about problems of the sort you describe. At least it eliminates a significant number of known problems, and then it might be worth discussing your code.
Are you using none blocking sockets? I guess you should.
You should probably create a thread for each incoming connection.
So the main loop should act none blocking and create a thread for each incoming connection, that receives the data and sends the answer.

Does websocket only broadcasts the data to all clients connected instead of sending to a particular client?

I am new to Websockets. While reading about websockets, I am not been able to find answers to some of my doubts. I would like if someone clarifies it.
Does websocket only broadcasts the data to all clients connected instead of sending to a particular client? Whatever example (mainly chat apps) I tried they sends data to all the clients. Is it possible to alter this?
How it works on clients located on NAT (behind router).
Since client server connection will always remain open, how will it affect server performance for large number of connections?
Since I want all my clients to get real time updates, it is required to connect all my clients to server, so how should I handele the client connection limit?
NOTE:- My client is not a Web browser but a desktop application.
No, websocket is not only for broadcasting. You send messages to specific clients, when you broadcast you just send the same message to all connected clients, but you can send different messages to different clients, for example a game session.
The clients connect to the server and initialise the connections, so NAT is not a problem.
It's good to use a scalable server, e.g. an event driven server (e.g. Node.js) that doesn't use a seperate thread for each connection, or an erlang server with lightweight processes (a good choice for a game server).
This should not be a problem if you use a good server OS (e.g. Linux), but may be a limitation if your server uses a desktop version of Windows (e.g. may be limited to 200 connections).

Ajax vs Comet (not a chat application)

I've developed a web-based application in which a signed in user should send a message to the server telling he is still online every 3 seconds. The message is then processed by the server and a stored procedure is called in Mysql to set the user's status to online.
I've looked in to similar issues in which Comet and Ajax are compared (here or here) but considering that 3 second delay is acceptable and maximum users of 1000 are online in the system, is using Ajax a wise choice or Comet should be used?
For this kind of feature comet is more appropriate:
Your clients send messages (i'm online)
Your server broadcast the processed message (user X is still online)
In an ajax way you are only serving messages to server.
In order to get the "broadcast effect" in an ajax way. You will end up doing something similar to comet but with less efficient bandwidth.
Ajax:
Client send server - i'm in
Server process
Server send back to client list of user in.
In this case every client ask every 3 second the database for the COMPLETE "in" list.
In comet:
Client X send server - i'm in
Server process
Server send back to client S that user X is still online
In this case every client tell the server every 3 second that he is in.
The server send back to every connected client ONLY that x is still in
Comet is just the technique to broadcast back and push messages to client
Ajax is the technique to push client information to the server without having to refresh all the page.
Quoting wikipedia:
http://en.wikipedia.org/wiki/Comet_%28programming%29
Comet is known by several other names, including Ajax Push, Reverse Ajax , Two-way-web, HTTP Streaming,and HTTP server push among others.
So go comet :)
If you do not broadcast anything, then simple Ajax is the best option
In this particular case, since you do not need to send any information from the server to the client(s), I believe Ajax is the more appropriate solution. Every three seconds, the client tells the server it is connected, the database is updated, and you're done.
It could certainly be done using Comet, in which case you would basically ping each registered client to see if it is still connected. But, you would still need to run a query on the database for each client that responds, plus you would still need the client to notify the server on its initial connection. So, it seems to me that Comet would be more trouble than it's worth. The only thing that might make sense is if you could ping each registered client and store the responses in memory, then once all clients have been pinged you can run one single query to update all of their statuses. This would give you the added bonus of knowing as soon as a client disconnects as opposed to waiting for a timeout. Unfortunately, that is beyond the scope of my expertise with Comet so, at this point, I can't help to actually implement it.

Resources