I have two servers, written in Java, that communicate through sockets and TCP. The servers are both on the same Linux machine. If the servers were on different machines then data would have to go through network adapters and network cables, but since the servers are on the same machine how does the OS actually move data efficiently form one server to the other.
Generally, the messages will be sent over the loopback interface:
In TCP/IP a loopback device is a virtual network interface implemented in software only and not connected to any hardware, but which is fully integrated into the computer system's internal network infrastructure. Any traffic that a computer program sends to the loopback interface is immediately received on the same interface.
However, you can manage to configure your situation so that, for example, the server is referred to via an external IP address so that messages actually go out over the network before being routed back to the same machine.
If two programs are using TCP/IP to communicate on the same machine, they are probably connected through the loopback interface
Related
I have a local Linux machine (L), with a network interface (eth0) connected to (currently) only one device (D).
L/eth0 only has an ipv6 link local address.
D also only has an ipv6 link local address.
Client software on L discovers D's link local address by sending a custom multicast packet over eth0. The response contain the device link local address as well as some configuration information. From there it communicates with the device using various UDP or TCP port.
My goal is to be able to use the device from a remote machine (R). The remote machine being most likely a Linux docker container running on some host - although it could be a native macOS or other.
That means running both the discovery protocol and communicate with the device. I definitely cannot modify software on the device, and cannot modify most of the client software running on the local Linux or remote machine (Limited modification could possibly be requested)
My idea was to somehow create a virtual interface on the remote machine, that would remotely be bridged to the the local machine, so that the multicast discovery works from the remote, and so that I can just connect to D link local address from the remote.
I'm not sure which tool(s) I should be using.
I'm googling various things about VPN and tunnel, TUN/TAP interfaces, bridges, VETH, VLAN, etc... but I'm having trouble connecting the dots here... I have no budget for this, so I'm looking for open source tools, or just something I can cobble up together with common tools.
Thanks
I'm working on a simple Swift app where one user can find other machines on a local network which are running an instance of my app and then send data to that machine using TCP sockets.
My question is how to find IP addresses of devices on same network which are running an instance of my app (cross-platform)?
I was thinking about listing all devices on local network and then
checking whether they have opened specific port (the port my app is
using)?
I also found that Apple provides service called Bonjour which could make my process discoverable. I'm not sure if this solution is good for cross-platform communication.
Apart of Mac-related stuff and high-level solutions (I believe keywords "network service discovery" will bring you to them), there are a couple of things that will work for a local network:
I Have a server that clients should report to. Some short hello-like UDP message and a timeout mechanism will be sufficient to keep a list of available clients in the network.
II Use IP or UDP multicast groups to notify others that a client has just connected to the network. Send a message to a multicast group and listen to this group to build a list of clients.
However, broadcasts and multicasts won't be transmitted through a router. So if your network is large enough only neighbouring clients will hear your notification. In order to overcome it,
III DHCP servers can be configured to provide custom data to clients via unassigned DHCP options. Large networks have usually such server. You probably can use it to send out a list of clients, but I'm not sure about this.
I am intending to write some software modules that will all run on a single machine, and I would like for them to communicate with each other via TCP.
It is likely that the machine I am using will not have a physical network adapter. Will TCP connections to localhost be able to work correctly if there is no physical network adapter?
The operating system that I am running on is Linux for Tegra.
Thanks.
I have a question about network connections among computers.
I've made some applications where messages pass through the Internet (via sockets) to make a connection between two devices. However, a strong condition is that two devices must be connected to the same network.
Can anyone give me a trick how to create a communication using sockets between two computers even if they are connected to different netwkorks?
Thank you in advance.
Here is a great tutorial on how to use sockets and general networking
(in java) http://www.thenewboston.org/watch.php?cat=25&number=38
In order to communicate between two diffrent networks over the internet, you will need to do something called port forwarding. What that does is that when your public IP of your network receives a packet with a spesific port number. The router knows where to send that packet to which local IP.
If you dont port forward and receive some data. The router doesent know where to send the packet. Therefore it discards it, which means others wont be able to connect to you.
You will only need to port forward the network with your server (using the example i linked). How you do that is by logging in to your router, and say that a port which the server uses gets forwarded to the IP of the PC hosting the server.
On the other network (client) you will need to change the IP address of which the client shall connect to. That IP address needs to be your public IP of your server's network. You can find that by connecting to the server's network and go to: http://www.whatsmyip.org/ . Keep in mind that public IP addresses may change over time.
Hope this helped!
-Kad
According to http://www.cisco.com/en/US/solutions/collateral/ns340/ns517/ns224/ns892/ns894/white_paper_c11-525307.html
Each virtual machine is given a dedicated network interface card. My question is, how do a server containing about 10 virtual machines, ever support 10 NIC's ?
Those NICs are probably virtual. Packets from them are routed to the physical NIC(s) and the other way around. It's pretty much the same thing as you get in modern WiFi routers: at home you only have one Ethernet port from your Internet Service Provider, it's in the modem. You connect your router to it, but your router may have 2+ Ethernet ports to which you can connect multiple PCs.
They can be physical too and either be directly accessible to VMs or indirectly.