I'm writing a server in C ++ for both Windows and Unix systems.
A key feature of this server is that it must be able to receive and send network packets at any time.
Specifically, the server must be able to send data to the client not only in response to their messages, but also be able to send packets to them asynch in push.
I'm having difficulty in implementing a solution that uses the select() function in the scenario described above.
The solution I have currently implemented does not convince me at all and I think it can be implemented with better patterns/solutions.
I currently have a dedicated thread (selector) that performs the select by listening on events in the reading for the server socket (to accept new connections) and for the sockets connected to the server.
This is the main select() loop:
if((sel_res_ = select(nfds_+1, &read_FDs_, NULL, &excep_FDs_, &sel_timeout)) > 0){
if(FD_ISSET(serv_socket, &read_FDs_)){
//we have to handle a newly connection.
...
if(sel_res_ > 1){
//in addition to the newly connection, there is also some other message incoming on client sockets.
...
}
}else{
//we have to handle incoming messages on client sockets
...
}
}
This solution works well for receiving the data and to respond to client requests in synchronous form.
However, the server must also be able to send asynchronous data, and send when necessary, packets in push.
To do this I currently use separate threads that perform directly the send() on the client sockets.
This solution does not convince me, and I would like to centralize the packets receiving and sending on the selector thread.
The main difficulty is that the select() by its nature is blocking and I have no control until a client does not send any packet or the timeout is triggered.
The solution to set a timeout very low does not convince me; I see it as an easy solution that is actually doing active wait, and not only, however, the worst case I would pay the price of the timeout before sending the push packet.
I thought a more 'elegant' solution; I think, will work well, but only for a Unix/Linux platform.
I thought to use an anonymous pipe and insert into the select() read_FDs_ the anonymous pipe read descriptor.
In this way, when a thread wants to send a data in push, it writes something on this pipe, interrupting the select() and returning control to the selector that can then predispose to send the data to the client, without significant loss of time.
I think that this solution, unfortunately, cannot be implemented on Windows because the select() function on that system works only with fds that are actually sockets.
So the question is: Is there some well known solution that can be used to address this kind of scenario (both Linux and Windows)?
You can create a self connected UDP socket, this works equally well on Windows and Linux.
Basically, you create a UDP socket, bind() it to INADDR_LOOPBACK and port 0, and connect() it to itself (with the address taken from getsockname()).
At this point, you can send yourself a single byte message (or something more specific) to wake yourself up.
Related
I'm currently implementing a daemon server that acts as 2 servers. One of the servers is recieving logs via UDP from a collection of producers. The second server is broadcasting every log that was received from a producer to a consumer who is currently connected via TCP.
These are 2 separte sockets. My current(pretty basic) implementation is to use select() on these 2 sockets, and handle every read signal accordingly, so my code is basicly(NOTE this is pseudo code)
for(;;) {
FDSET(consumers_server)
FDSET(producers_server)
select()
if consumers_server is set:
add new client to the consumers array
if producers server is set:
broadcast the log to every consumer in the array
}
This works just fine, the problem ocurres when this code is put in to stress. When multiple produers are sending logs(UDP) the real bottleneck here is the consumers which are TCP. Sending a log to the consumers can result in blocking, which i can't afford.
I've tried using non-blocking sockets and select()ing the consumers write fds, the problem is this would result in saving the non-sent logs in a buffer, until they can be sent. This results in a very unelegant massive code, and the system is also low on resources(mainly RAM)
I'm running on a linux distro.
An alternative approach to synchronize between these UDP and TCP connections would be welcomed.
This is doomed to failure. Sooner or later you will be unable to send to the TCP consumer. Whether that manifests itself as blocking or EAGAIN/EWOULDBLOCK isn't really relevant to the underlying problem, which is that the producer is overrunning the consumer. You have to decide what to do about that. You can have a certain amount of internal buffering but at some point you will have to stop reading from the UDP producers. At that point, UDP datagrams will be dropped and your system will lose data, and of course it is liable to lose data anyway by virtue of using UDP.
Don't do this. Use TCP for the producers: or else just accept the data loss and use blocking mode. Non-blocking mode only moves the problem slightly and complicates your code.
I'm trying to write a server/client pair to run over udp, and the only way i've been able to get it going is having the server aware of the client's ip and port before the connection starts. My new design involves waiting for a packet to come in, recording the sender address, forking to a child process (the parent loops around and continues listening), which then connect's to the client the transmitted the packet. The child should then only receive packets from the associated client, but the documentation is unclear is the parent socket will continue receive traffic from that client. I'm working on a program to try it, but i figured i could ask the question at the same time.
EDIT: It seems that when the child's socket is connected'd it will connect the parent's socket too.
UDP protocol does not operate connections, it's a connection-less protocol. It is enough for one side to listen and other side to just send datagrams for data channel to work.
On the question (sorry, didn't got the point before): forking is not the way out when working with UDP. Connection-based protocols are widely used with that technique. That is possible because:
you can fork right after listen()
the first process accepts connection works with it (and only that process posesses the newly created connected socket.
When you work with UDP you don't have such gap (as before accept() with TCP) to know when exactly to fork (especially when you have intensive datagram flow).
So, when you design UDP service, you need either
use non-blocking I/O with event loop or
design threaded solution.
I'm breaking my mind trying to understand how make a client/server write by myself accept multiple socket connections.
The connection is a datagram (UDP), for now was implemented based on getaddrinfo(3) man page works nice, but each client needs to wait the process of early connections be processed.
I've heard about select, but in its man page says:
select() can be used to solve many problems in a
portable and efficient way that naive programmers try to solve in a
more complicated manner using threads, forking, IPCs, signals, memory
sharing, and so on.
and more:
The Linux-specific epoll(7) API provides an interface that is
more
efficient than select(2) and poll(2) when monitoring large numbers of
file descriptors.
So, it is? epoll is simply better than select? Or it depends? If it depends, on what?
epoll man pages has a partial sample, so I'm trying to understand it.
At now, (on server) I think, I need a thread to listen in a thread and write in another. But how to control the completion of a partial message? If two clients send a partial message interleaved, how to identify? By the sockaddr? if it's the only need, I can manage it without a pool, so why use a epoll?
Can anyone try to explain to me, how to make, or where to learn about a multi-connection client-server UDP app?
I think there is a misunderstanding here about UDP. UDP is not a connection oriented protocol which means there is no permanent connection like TCP. UDP just bind to an address/port and waits for packets from everyone. At the server there is only one socket listening per address/port number. When a packet is received you may find out who is the sender by the packet's source IP, you can reply to the sender thru this IP.
As I see it, there is no need for poll() o select() you bind to an address/port and asynchronously receive packets. That is, when a packet is received you get a signal/message alerting your asynchronous function. This function should be reentrant, it means that in the middle of a reception another signal could be received and care must be taken when accessing/modifying global stuff (variables/objects). When dealing with an incoming packet it should be processed as soon as possible or, in case the process takes too long, you better keep the packet in a packet spool and process them in another [less priority] thread.
For UDP packet size read this question.
For UDP fragmentation read this
For UDP packet header read this
I want two functionalities to be implemented on my udp server application.
Creating thread that continuously receives data coming from any client.
Creating a thread that continuously sends data on server socket after specific time period and waits for reply from client. (I implemented this to make aure that whenever any client goes down, the data is not received back from client and server comes to know that client is down.)
Now, the problem I am facing is that since two threads are sharing same connected socket, whenever both threads try to access this socket simultaneously, a deadlock is established.
One of the solution I found was to create two sockets. One that continuously receives data, and the other socket that is meant for sending data from server time to time to clients and wait for their response, but since Server can must be bind()ed and I have bind()ed my socket to INADDR_ANY once, how would I create a separate socket for sending data from server and waiting for replies from client.
Please help me with this complication.
Also do let me know if there is some other better way of its implementation.
Thanks in advance :)
You will have to use non-blocking net functions and use a mutex to ensure no two threads access the socket at once.
A single thread may, however, be enough, if you use non-blocking functions. Using many threads will probably not improve performance, but may make the code more readable.
I'm implementing a TCP socket server in C and Linux. It will be a chat server handling many requests simultaneously. Here is the pseudo code for how it is setup
create a socket
bind to a well-known port
use listen to place in passive mode
while (1)
{
accept a client connection
fork
if (child)
{
communicate with new socket
close new socket
exit
}
else
{close new socket}
}
From what I've seen, the "communicate with new socket" part consists of reading the buffer and then possibly writing to the buffer, then the socket closes. I thought sockets were suppose to be a persistent connection? If the client has to reconnect every time it wants to send more data to the server, isn't this inefficient and defeating the purpose of sockets?
If the client has to reconnect every time it wants to send more data to the server, isn't this inefficient and defeating the purpose of sockets?
It's impossible to say without knowing more about the nature of your application:
In some cases, it makes sense to maintain a persistent connection for the duration of the client's lifetime.
In other cases, connecting on every message would also work well.
You are quite right on the efficiency front: there is quite a lot of overhead involved in establishing a TCP connection. Therefore, if the messages are frequent it may turn out to be very costly to set up a new connection for each message.
Don't close the socket. Put the code which receives and responds to the messages in a loop and using blocking I/O ( the default ) to block on reading the socket until a request arrives. This way you'll have a forked process dealing with each client and using minimal resources. You only close the socket and exit the child process when the client explicitly closes the connection because it's done.
Use UDP instead of TCP. This way there is no "connection" between the server and the client. You multiplex multiple clients to a single socket without forking. If you care about reliability you'd need to add a sequence number to each message to keep them straight and allow retransmission on failures.
Keep a single process using TCP, but store a socket descriptor set and use select to check for readable sockets. You can then retransmit that data on each or some of the other client sockets.
I'd point out that forking a child for each client in a chat server seems like a poor implementation. How do two people who want to chat communicate? Each one will be served by a separate process - how do the processes communicate with each other so that chat data can pass between clients?