Any ip and local ip binding? - c++

I'm working on a network platform that creates a basic I/O interface for make something fast/easy. I need to understand this, what is the main differences between binding any ip or local ip to socket? Why I need to bind local ip, why I need to bind any ip? Simple but important question :)

Binding a server socket to a local ip instead of to 0.0.0.0 can act like a poor man's firewall if you want to have a server that is not open to the entire world, but only to a specific group.
For example:
- You can bind a socket to the local ip if you want it to be used a way for different processes in the same machine to communicate.
- You can bind a socket to an ip if you want all access to it to occur through some external load balancer.

Remember that a TCP connection is made up of four things: Source address, source port number, destination address and destination port number.
When e.g. a server program binds a socket to a local address, two of those things are provided, and allows clients to connect to a well know destination.
When you bind to any address, it doesn't actually bind to any address, it just creates the binding for all local addresses. (For hosts that have multiple interfaces, which all hosts have.)

Related

C++ / Qt: How can I detect when a hostname or IP address refers to the current system?

I'm writing a macOS C++ application using Qt that acts as both a UDP client and UDP server. The server functionality allows the user to configure which port the UDP packets will be received on, and the client functionality allows specifying both a target host and a port number. The host can be either a hostname or an IP address, including addresses or hostnames that resolve to a multicast or broadcast address.
In order to help prevent circularity, when the server functionality is enabled I need to be able to warn the user when the host and port they've entered for the client would send the packets directly to the app's server. Of course it's easy to check if the port numbers match, but this means I need to know whether the host they've entered refers to the current system.
Examples of hostnames or IP addresses that would typically be problematic:
127.0.0.1
localhost
192.168.1.255 (assuming the system is on a 192.168.1.0/24 subnet)
any of the IP addresses assigned to the current system's network interfaces
the system's local DNS name
any other loopback addresses that may be configured other than 127.0.0.1
How could I go about detecting this?
Since this app is being written using Qt, a solution that exclusively uses Qt's framework would be ideal. But if that's not possible (since Qt doesn't handle every possible use case) a macOS-specific solution will work too.
QNetworkInterface class should provide the most information you may need.
You can obtain a list of IP addresses using QNetworkInterface::allAddresses() static function.
You can also get a list of all interfaces using QNetworkInterface::allInterfaces().
Calling QNetworkInterface::addressEntries() for each QNetworkInterface returned by QNetworkInterface::allInterfaces() will give you more information about address entries for each interface.
auto ifs = QNetworkInterface::allInterfaces();
foreach (auto interface , ifs){
auto addresses = interface.addressEntries();
foreach ( auto addy , addresses){
///play with the addy here.
}
}
You can also test hostnames against those ip addresses which you are going to ban, using QDnsLookup class.

How to bind specific source IP address using Poco::Net::HTTPSClientSession

I have a https client that can run on one of two redundant hosts. Would like the host that is the 'Active' client host to use a specific IP address as the source address so that traffic appears to be coming from one source. Low level OS network setup to allow this is all in place. Our https client is built using the Poco::Net::HTTPSClientSession class.
From reading about this situation binding the shared source IP address to the client socket is the standard solution. Attempted the following code:
// set the shared IP address and use 0 (represents wildcard) for port
Poco::Net::SocketAddress sa = Poco::Net::SocketAddress(ipAddress, 0);
// bind the shared IP address to the socket
Poco::Net::SecureStreamSocket sss;
sss.impl()->bind(sa);
Poco::Net::HTTPSClientSession session(sss);
Unfortunately, this doesn't work as the bind() is throwing Poco::InvalidAccessException and exception message is "Cannot bind() a SecureStreamSocketImpl". Looked at the Poco source and bind() is just throwing this exception. Seems to not be allowed.
How to make this work with Poco::Net::HTTPSClientSession class?
Is it even possible to make it work with this class?
Not currently possible - there is a pending issue for it. It will be done eventually but currently there's no timeframe.

How to make a .cpp file to act as an accessible server

I have written a simple program with Linux (Cent OS 7.0) and C++. It is a very small server which sends back a string of characters to the client. But my problem is that I don't know how should I access that server using an IP address?
I have used Linux Socket Interface (Berkeley), and in the section which defines the address, my code does the following:
serverObject.
sin_family = AF_INET;
serverObject.sin_addr.
s_addr = htonl(INADDR_ANY);
serverObject.
sin_port = htonl(portNumber);
I use INADDR_ANY as my server's address which is defined in its definition as:
/* Address to accept any incoming messages. */
Now, how should I run the server, and then use my simple client program to send request to it. My simple client program accepts an IP address as it's destination address, this address should be the one destined toward to the server. How should I relate it then?
INADDR_ANY goes to specify that all active network interfaces in the system should be bound to. So if you're connected to more than one network, you'll be able to communicate with connections coming in from all of them. Most systems will usually have just one, though, and this still goes to say that if the IP bound to that interface happens to change, you'll still bind to that interface.
So, once you specify INADDR_ANY, you need to initiate connections according to the following rules:
If you're connecting from the same physical machine, the easiest thing would be to use the loopback interface (127.0.0.1). However, you can still do (2).
If you're connecting from another machine, you need to pick the accessible IP address of your server from that machine. As said above, if your server is only connected to one network, this will simply be the IP address of the server. Within an internal network this will often be something like 192.168.x.y, or 10.0.x.y—but it doesn't have to.
If you're connecting from a different network which uses a gateway to access your server, then you will need to set up port forwarding in the relevant routers so that when they receive connection to port X, they will know to internally transfer it to your server.
As a server programmer, you decide the port on which to listen, but not the address.
The internet address is provided by your internet provider, or 127.0.0.1 to test on your own machine.
There are plenty of web pages on internet that provide tools to tell you your current public address (search for What is my Ip).
Most of the "home" internet routers implement NAT: they have a single internet address and map them to many device, that carry the Port number to be changed (your port 80 become port (e.g.) 2345 for outside). To allows a client from outside your home to access your server, you are required to configure your router to map the server port, so for example your public port 80 map to your server port 80.
With that said, you should be able to connect your client to your server through an address and port.
If then you want to use a name (example.org) instead of an IP (93.184.216.34), a Domain Name Server is used. But that is another topic.

Is it possible to bind and listen to one IP address with TCP/IP sockets? (Linux/C)

I have always used:
serv_addr.sin_addr.s_addr = htonl(INADDR_ANY);
which means that I will accept connections from any interface. If I replace the INADDR_ANY with “192.168.0.1”. Does this mean I will only accept connections from the IP address of 192.168.0.1, or does it mean I will accept connections from the interface of 192.168.0.1 resides on?
I have a situation now where I have multiple clients (each with a unique IP address but same port number), trying to connect to one server. Can I have multiple listens functions (separate threads) listening to a unique IP address and port? Or do I have to accept any connection and get the network information after I've connected?
Edit To give more clarification.
If I say serv_addr.sin_addr.s_addr = inet_addr("192.168.0.1") and a client with IP address 192.168.0.2 tries to connect, will the listen command reject this?
The bind address is the local address to listen on; you can specify the address of a local interface.
If you want to control who can connect, you can either inspect the peer address inside the select/accept loop, or limit inbound connections using iptables.
Update
If I say serv_addr.sin_addr.s_addr = inet_addr("192.168.0.1") and a
client with IP address 192.168.0.2 tries to connect, will the listen
command reject this? I want to be able to have multiple thread, each
servicing a unique IP address.
No, the address is an address on a local machine. Given that you're going for a multi-threaded design, I'd recommend you run the listen/accept code in one thread, check the client address, decide what worker thread is appropriate, and then spawn it.
At the risk of showing my age, I still prefer using listen/accept/select for most socket code - it's a personal taste thing, and yes, does influence the design when it comes to blocking/non-blocking IO, buffering, etc.
If I replace the INADDR_ANY with “192.168.0.1”. Does this mean I will
only accept connections from the IP address of 192.168.0.1, or does it
mean I will accept connections from the interface of 192.168.0.1
resides on?
Yes
I have a situation now where I have multiple clients (each with a
unique IP address but same port number), trying to connect to one
server. Can I have multiple listens functions (separate threads)
listening to a unique IP address and port? Or do I have to accept any
connection and get the network information after I've connected?
Yes. You can "fork" (i.e. create a new thread) for each connection you wish to process, however you'd have to do the filtering yourself. There is no way (that I know off in the API) to offload that work to the stack.

Hold a network connection although IP address change

Is it possible to hold an open TCP connection with a client, while the IP address of the client is externally changed?
For example, the connection is establishes against address X, but somewhen while the connection is open, the client-side user asks for IP renew and gets another IP address. Can the connection remains alive in this case?
Thanks in advance.
No, it cannot.
Even if the local side could be massaged to understand that the connection is suddenly between different addresses, the remote side will not understand and will refuse to work with it.
You'd need to re-add the old IP address to continue using the connection.
To do so:
Linux: ip addr add 172.16.10.20/22 dev bond0
Windows: do some pointy-clicky or**fill in command here**
This is possibe with tcp v6 connections, if you're using e.g. tunnelbrocker.net. Every time your IPv4 address changes, the connection between your IP4 address and the tunnelbrocker's IP4 address is dropped then re-established, however your IP6 attress is the same and all the TCP/IPv6 conections from your IP6 address to your destination IP6 addresses are still here.
Or, if you're designinng both your client & server, your could design your protocol allowing the client to reconnect in a transparent way..
What I know is, when using UDP, the connection can be remained, one of my application is to change a remote network adapter address via UDP.
For TCP, I agree with Soonts, you should design your server and client application to allow the reconnect in a transparent way.