I'm trying to build a p2p terminal sharing utility where PCs behind a NAT can share their terminals to other PCs behind a NAT through the public internet. (both PCs are on different private network basically). Another requirement is to provide a web based terminal as well ( and hence webrtc seems useful ) .Came across two promising solutions - webrtc and libp2p. But what is the fundamental difference between the two?
PS: ssh Tag might be misleading here. But on a broad level, all I'm trying to do is a make-believe SSH into a system that does not have a public IP
WebRTC is a low-level Browser API for P2P communication between browsers. See MDN documentation or Wikipedia article for high-level introduction and history.
libp2p is a library written in JS, NodeJS, Go, Rust (with more in the way) that implements Peer to Peer communication based on IPFS. From what I see it uses WebRTC and WebSockets in Browser.
Related
I've been asked to implement VPN capabilities in an existing software project on an embedded system, in order to make the device available via network to an external server while avoiding trouble with firewalls (no need for encryption, just to make it accessible).
Unfortunately, the embedded system is based on a Cortex-M4 MCU, therefore Linux, which would allow for VPN nearly out of the box, is not an option. All I've got is an RTOS and a working LwIP stack.
I've used VPNs in the past. However, my network knowledge is rather limited concerning implementing VPNs, so I'm rather stumped. As I think, I'd use the current LwIP instance for building up the tunnel connection, and the application would use a second instance for the actual network communication, while the network interface of the second instance is a virtual one (like a tap device on linux), encapsulating its low level data and tranceiving it via the tunnel connection of the first LwIP instance.
Maybe this way I'd be able to create a custom solution for the problem, but the solution should conform to any standards (as the server will be any kind of sophisticated system).
So I wonder if anyone has been confronted with a task like this, and would appreciate any hint what to do, at least a direction where to look at.
Thanks in advance!
I try to make a P2P Program and need help with getting through the NAT of the clients.
I have read many questions here on stackoverflow, but i never got what the drawbacks and benefits of all the Methods to get through a NAT are.
How many routers support which methods?
Which methods are commonly used by the big companies? (BitTorrent, TeamViewer,...)
What are the drawbacks/benefits of the three methods i listed and which else exists?
UPnP requires software support in your router. Even when routers have it, it's sometimes turned off due to security concerns.
STUN uses a simple public server to detect the presence of NAT and the real IP addresses. This is useful for other, higher-level protocols that do the useful work.
Hole punching is a simple technique to connect two computers. It requires a 3rd public computer to communicate between the 2 computers behind NAT. You could run your own server with a custom protocol, our you could use STUN and existing STUN servers. (Only a few packets are needed for setup, the rest are peer-to-peer).
Of course, some NAT firewalls are incompatible with the above, so you may need a custom protocol. Bittorrent does this: they use different techniques depending on which clients are behind firewalls and whatnot.
I wonder if there exist any other technologies used to establish internet connection between applications. Are there any other? I am searching and so far I haven't found anything else described.
There are many abstractions on top of sockets, if you don't want to deal directly with a socket API. UDP, TCP/IP, various RPC protocols, HTTP (which is on top of TCP/IP), etc. Many programming languages have easy methods of doing, say, an HTTP request and getting the resulting document. You can use that to allow applications to talk to each other over the internet without using a socket API.
What are you trying to accomplish?
If you want to skip sockets you basically have to implement your own means of talking to the network card hardware and telling it to communicate with other devices. A socket is just the abstraction chosen for *nix and Windows machines.
I am trying to implement a peer to peer communication ..
I use sockets for communication between them. I want to know if there are any ways i can use the same sort of communication when the two peers are behind two private Networks (They dont know each others public ip address.) I can think of a shared buffer in a well known location as a means of communication. But is there some other way to get this done?
You may also have to consider the case where two peers don't even have public IP addresses, in which case you could use STUN techniques to avoid the need for a central server. STUN has been used effectively in products such as Skype (I think), but it remains quite a black art at this stage.
For a central approach, consider XMPP, of which ejabberd is a very popular implementation. Various other messaging protocols exist as well, such as the emerging AMPQ standard.
These technologies are likely to be more robust and scalable than a home-grown socket-based protocol.
STUN and TURN are the common ways to achieve this as already posted and are standards based approaches. However, I have seen a really interesting approach to doing this using icmp messages recently. It's called pwnat, and available here
Since both are in different private networks, both peers cannot communicate with each other.
You need to have a public server in place, so that you can establish connection via that server.
Refer resiprocate , Resiprocate for details
I know that a protocol is a set of rules that governs communication between two computers on a network, but how are thoses rules implemented for the computer? Is a protocol basically a piece of code or, in other words, software?
Protocols are generally built upon each other. At the risk of sounding pedantic, here's an example of a protocol and where/how it's implemented:
Application Protocol - the way a particular application talks to another instance of itself or a corresponding server; this is implemented in the application code or a shared library
TCP (or UDP, or another layer) - the way that information is sent at the binary level and split up into usable chunks, then reassembled at the destination; this is usually implemented as part of the operating system, but it is still software code
IP - the way that information (having already been split or truncated by something like TCP or UDP) makes its way from one place to another by routing over one or more "hops"; this is always software code, but is sometimes implemented in the OS and sometimes implemented in the network device (your LAN card, for example)
base-T (ethernet), token ring, etc - Here we are physically getting into how the hardware talks to one another; ie, which wire corresponds to a particular type of signal; this is always implemented in hardware
electricity /photons - the laws that govern (or at least define) how electrons (or photons) flow over a conductive material or over the air; this is usually implemented in hardware ;)
In a sense, these are all "protocols" (a set of rules or expected behaviors that allow communication to take place), and they're built on one another.
Bear in mind that (aside from electricity) this is not an exhaustive list of the sort of protocols that exist at any of these layers!
Edit Thanks to dmckee for pointing out that electricity isn't the only physical process used in networking ;)
Networking protocols are not pieces of code or software, they are only a set of rules. When software uses a specific networking protocol, then the software is known as an implementation. There can be many different software implementations of the same protocol (i.e. Windows and UNIX have different TCP/IP implementations). It is possible to understand networking protocols without any knowledge of programming.
EDIT: How are they implemented? Here's a paper on taking an abstract specification of a protocol and implementing it into C. You'll see that less-strict protocols leave out certain details that programmers have to guess on, which makes some implementations incompatible with others.
A network protocol is basically like a spoken language. It is implemented by code that sends and receives specially prepared messages over the network/internet, much like the vocal chords you need to speak (the network and hardware) and a brain to actually understand what someone said (the protocol stack/software).
Sometimes protocols are implemented directly on the hardware [for speed reasons] (like the Ethernet protocol for LANs) - but it is always software/code required to do something useful with a protocol.
This might be interesting for you:
The OSI Model
Protocol (Computing)
Software implements the rules defined in the protocol, some protocols are formal defined and some informal.
a protocol is a set of rules governing the communication between two entities.
in the computer/programming context, a protocol is a set of rules governing the communication between two programs.
in the computer network context, a protocol is a set of rules governing the communication between two programs, well, over network.
in computers, in the end everything is embodied in code...
Protocols are basically set of rules. The way to implement them is to first of all make a state machine diagram as it completely tells that what is going to be the current state and how the state is going to change on the basis of input and what output actions are going to be performed.
Your answer is a very short one:
BY READING THE RFC.
The main networking problem is to share data between computers. All the networking protocols try to solve is a little part of that major problem. Some of them (the protocols) are implemented as software, some others as hardware. In short, protocols like algorithms, can be implemented it in many programming languages.
Back to the TCP, it is implemented by the operating system.