I'm trying to send data from one software to another running on the same pc. The coding in the former s/w is done in python. I'm doing this:
sock = socket.socket(socket.AF_INET, socket.SOCK_DGRAM)
sock.sendto("msg,Hi!", ("127.0.0.1", 12345))
This works fine when I have internet access (AF_INET). However, this is just a sample code and I will need to include this piece in some other application where I will not have internet access and so AF_INET wouldn't work. Is there a possible replacement for this?
I've already tried AF_UNIX and AF_LOCAL but neither of them works. I need to use UDP.
AF_INET has nothing to do with the INTERNET. It just means use IP based communication. You can do it on the same computer or across computers. 127.0.0.1 refers to local-host IP i.e. my own IP. Each unix/linux machine has the same local IP.
So yes this code will work without internet access.
Of course, You do not need any Internet concepts while using Sockets. You can use an ad-hoc network to create a relaxed network and run a simple TCP or UDP socket script to transfer things you like.
You can infact use Multicasting or Broadcasting schemes to share data on different devices within the ad-hoc network.
Related
I need my own IP in a small script and in order not to hardcode it, I`ve found a piece of code from here(stackoverflow) that works.
This--
with socket.socket(socket.AF_INET, socket.SOCK_DGRAM) as s:
s.connect(("8.8.8.8", 80))
print(s.getsockname()[0])
--
What is not clear for me is why it only works on UDP and not TCP? It has something to do with the google dns server? Thanks in advance.
This has nothing to do with Google and nothing to do with DNS.
All what this code does is to "connect" a UDP socket to an external IP, so that the OS kernel figures out which local IP address it needs to use in order to reach this external system. This is no real connection, i.e. there is no traffic involved but the OS kernel is only checking routing tables and local interfaces in order to decide which IP address to use as source in case one would actually use the socket to send data.
One could do the same with TCP. But in this case a real TCP connection would be established which means that actual traffic would be exchanged and that the connect would fail if the external system would not be reachable on this port (i.e. no listener, firewall in between etc).
With UDP instead connect will not produce any traffic and would fail only if no route to the destination IP address could be determined. This also means that an arbitrary external IP address and port could be used, i.e. ('1.1.1.1',11) would work the same as ('8.8.8.8',80).
My goal is to have remote control of a device on a WLAN. This device has software that enables me to configure this wireless network (IP, mask, gateway, dns). I can successfully connect this device, and my computer to a common network. Since both machines share the same network, I made the assumption that I would be able to open up a socket between them. Knowing the IP and port of the device that I am attempting to control remotely I used the following code, only to receive a timeout:
import socket
s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
s.connect(('192.168.xxx.xxx', XXXX))
(I am using python 2.7 on mac OS 10.11.6)
The network that I am connected to is on a different subnet that the IP that I assigned to my device. I also tried this having an IP on the same subnet as my network. There could be a number of things keeping me from opening a socket. That's not really what I'm after. The heart of my question is whether or not I can use python's 'socket' module to connect to a device wirelessly.
Yes you can.
So you get a timeout when you try to connect to a wireless device. There are several steps you can take in order to troubleshoot this.
Make sure your device has a program running that is listening to the port you want to connect to. Identify if the device can answer ICMP packets in general and can be pinged in particular. Try to ping the device. If ping succeeds, it means that basic connectivity is established and the problem is somewhere higher in the OSI stack.
- I can ping the device - great, it means that the problem is somewhere in TCP or Application Layer of the TCP/IP stack. Make sure the computer, the device, and intermediate networking equipment allow for TCP connections to the particular host and port. Then proceed to your application and the device software. Add some code to the question, post the stack trace you get or ask another one on SO.
- I can't ping the device - great. There's no connectivity between the devices and you're to identify the reason.
I) Draw a network diagram. How many intermediate network devices are placed in between the computer and the device? What are they, routers, switches? (Just in case, home grade wifi modem is a router.) Get an idea of how IP datagrams should travel across the net.
II) You said that the device can be used to configure an IP network. At least for troubleshooting purposes I would ignore this option and rely on a static IP or your router's DHCP server. Using an existing DHCP will ensure there's no IP misconfigurations.
III) Review routing tables of all the devices you have. Do they have an appropriate default gateway? Does a router knows how to pass the packets to the device. You're probably in trouble if the computer and the device are in the same subnet but attached to different network interfaces. Split the network in two subnets if needed and set up static routes between them on the router.
You can also use wireshark to see if data you send leaves the computer or is dropped right there by some nasty firewall.
There's a lot of caveats in getting a LAN working. You may want to ask questions on networking.stackexchange if these simple steps doesn't help you or if you have major troubles following them. Or just leave a comment here, I'd be happy to help.
I need to transfer data via pyzmq through two computers connected by an ethernet cable. I have already set up a script that runs on the same computer correctly, but I need to find the tcp address of the other computer in order to communicate. They both run Ubuntu 14.04. One of them should be a server processing requests while the other sends requests. How do I transfer data over tcp through ethernet? I simply need a way to find the address.
EDIT: (Clarification) I am running a behavioural study. I have a program called OpenSesame which runs in python and takes python scripts. I need a participant to be able to sit at a computer and be able to ask another person questions (specifically for help in a task). I need a server (using pyzmq preferably) to be connected by ethernet and communicate with that computer. It wrote a script. It works on the same computer, but not over ethernet. I need to find the address
Tcp is a protocol that uses an internet connection to Transfer data, through an IP address and specific port, you have to ensure that those ip directions are in the same range and uses the same port. For example: one pc 192.168.1.50 and another pc 192.168.1.60 in 502 port. The easiest way is using a Modem and set an statical address for each pc.
In the following link, you can find easily in which ip direction is connected your Computer. http://www.howtogeek.com/howto/17012/how-to-find-your-ip-address-in-ubuntu/
Maybe you could periodically send datagram message containing peer's ip address (or some other useful information) to broadcast address, to allow other peers to discover it. And after peer's address is dicovered you can estabish connection via ZeroMQ or other kind... connection. :)
I made some python script and ran it both on my computer and some distant shell (some website that provides shell access).
I'm using threads, pipes and UDP sockets to transfer data in a P2P fashion, so each script can both receive and send through the same socket. To test if this works, I open one terminal on my computer and one other terminal with ssh, connected to my shell. I make sure the script is the same on both machines, and feed it with an ip address.
Here is the script: http://codepad.org/V9Q1KcDT
(I don't know if I should paste it directly here or not)
My problem is this: strings I send seems to land something 20% of the time, sometimes often, sometimes not, and it seems to be random...
What am I doing wrong ?
Are UDP so unreliable ?
Are python thread+pipe+socket too slow ?
Could it be some kind of network problem with my shell provider ?
Is my program flawed ?
Are pipes a good solution to communicate with threads ?
I have no problem not using a shell and I have not tried, but it's useful for testing purposes.
BTW if i'm behing a router, how does the router knows where to send the packet if I'm not the only computer connected ? (I tried when I was the only one, it behaved identically).
User Datagram Protocol(UDP) could easily stand for Unreliable Datagram Protocol. UDP provides no guarantees regarding delivery so if you need reliability you have to implement it yourself by resending messages.
That said, threading and multiprocessing and remote machines and networks are a lot of variables to diagnose at once. I'd advise you to step back and try to boil the problem down to something simpler.
For instance:
first try TCP instead of UDP
instead of threads, run 2 processes (one doing nit_send, one doing nit_recv)
run them both on your local machine (use localhost: 127.0.0.1)
Once you have this basic test working then add features back in. For instance, switch to UDP. Once you've got your UDP issues sorted out locally try introducing remote machines or threads, etc.
Regarding the router question all machines connected to your router are most likely NAT'd so they all appear to have the same IP address (the IP address of your router) to machines on the internet. Look into NAT and port-forwarding if you want to route traffic to a specific machine on your local subnet.
Basically what I'm trying to achieve is a program which allow users to connect to a each other over a network in, essentially, a chat room. What I'm currently struggling with is writing the code so that the users can connect to each other without knowing the IP-address of the computer that the other users are using or knowing the IP-address of a server.
Does anyone know of a way in which I could simply have all of the users scan the IP range of my network in order to find any active 'room' and then give the user a chance to connect to it?
Also, the hope is that there will be no need for a central server to run this from, rather every user will simply be connected to all other user, essentially being the server and client at the same time.
I can give you two suggestions. First of all, UDP packets to the broadcast address of your network will be received by everybody. Secondly, there is a protocol for programs offering certain services to find each other on a local network. That protocol is called mDNS, ZeroConf, or Bonjour.
Using broadcast UDP is likely going to be the faster route. But if I were you, I'd learn how to use ZeroConf instead. It's supported well under IPv6 and already used by several interesting programs such as SubEthaEdit and Gobby.
Here is a link to a nice tutorial for implementing something that speaks ZeroConf in Python.
Another recommendation... If you want to hand roll your own broadcast/multicast UDP code and you can be sure that all of the systems you're on are running a Linux that's newer than 2003 or so, and all the Windows systems are XP or better, you can probably get away with using IPv6. The IPv6 link-local (think same LAN) all hosts multicast address is ff02::1. That's really simple and easy, and it will reach all the other systems on the same LAN. It's much better than having to figure out what your network's broadcast address is with IPv4.