I'm working on an idea of mine which is comparable to a home automation system.
The layout of network devices would be like this:
What I want is for my Mobile Phone App to be able to communicate with the home server at all times, but also for the server to be able to reach my phone (Push-notifications).
For this I thought it would be a good idea to implement TCP-Holepunching using my server in the datacenter as the center for traffic.
The problem is though, Since my Java skills aren't good enough yet, I'm forced to write my server in the PHP scripting language, but PHP isn't capable of keeping sockets active.
Is there TCP-holepunching server software (executable in Debian) available which would be able to keep the sockets alive and push any commands from my server towards the correct destination (ie: phone) and vice versa?
There are ways to get sockets working in PHP. One library that does this can be found here:
http://socketo.me/
Also an easy way of setting a server like this up can be done with NodeJS. You code your server in JavaScript (run on Google Chrome's V8 Engine) and then you can handle persistent connections based on events.
It should be possible (with NodeJS) to code a relay server (relaying from your phone to your home PC and vice versa) with only a few lines of code.
Related
I recently built a web application for my company, but the client needed the application to work both offline and online, depending on the availability of internet connection.
Anyways, I'm looking for a way to allow my application to run offline and online. The following technologies are used: PHP, MySQL, jQuery, HTML, and CSS.
Clients could be using that app without internet connection; if there is internet connection the app will automatically move to using the hosting server.
This is either trivial or insanely difficult, depending on a few things:
Does your client have an internal network it can use?
Can that internal network be accessed from the outside?
How many users will your application have?
Are they all on the same network?
On the trivial end of the scale, you can host the entire application locally on your client's network, that can also be accessed from outside, if there is an internet connection. But if there isn't as long as you're on the internal network, the service continues.
On the insanely difficult end, you can build your service so that each browser stores enough information on the machine to allow the user to continue to use the service even if internet connection drops. And then when the connection resumes, the service syncs any updates to the browser.
But this all depends on the answers to those questions.
I want to send a file from my browser to a server which is not the server on which the site is hosted.The problem is the another server is not a web server.
Is it programmatically possible to send a file to different server without the involvement of host server? I do not want copy of the file I am uploading on my hosting server.
Sorry for not being clear on the first go. I hope this makes sense.
Note: I have the IP address of server. Implementation has to be done in a secure way.
Create a signed application to handle the upload.
One of the fundamental rules of browser sandboxing is that you don't let the browser communicate to a server other than the one which provided the website. This has been a rule for quite some time, and is done for security reasons.
The only way I know of to bypass this restriction is to use a signed application. This is possible in several languages including at least Flash and Java. Once you've signed your application with an authoritative source, then you can specify that your application is known to communicate with a third-party server and should be trusted to do so. The process of doing this will be different depending on the technology you choose.
The server to which you are sending the file needs to allow some sort of protocol in order to accept the file. You do not specify what that is, but these languages should be more than sufficient to implement whatever you need.
given that you have the necessary authentication of the secondary server you can
i'll explain using ftp, but there can be other socket related ways
1) upload the file to your server as usual, using $_FILE
2) use the ftp class in php to transfer the file to the remote host
Thanks all for valuable suggestions. I am opting for bit weird way but it seems the only feasible solution considering the current requirements of the project.
I am planning to install lite version of web server and php on the external server which can accept uploaded files through browser.
Our team has worked out this is the best solution we can consider as of now.
Thanks again to all for their suggestions.
This is done all the time with Firefox plugins, (youtube etc.). Does it have to be driven from your page source ? How stealthily do you want to do this ?
The other server should use a common protocol, SSH or ftp or NFS or CIFS.
In any case the user is going to have to agree with the plugin or the applet or a third party service (think punkbuster) running in the background.
BF4 runs in a browser window via a plugin and routinely sends data to servers other than the game server.
I'm working on a project that is located on 2 domains within same server:
1. DataSource system, which provides data for main app
2. Main app, providing the data for front-end app.
App 1 needs to work on seperate domain, as it's data source for more applications. I'm trying to find some way to boost communication performance. Simple call from app 2 to app 1 takes approximately 0.3-0.4s.
Is there anyway to force server to bypass TCP/IP communication and call service directly from localhost?
Both applications are written in PHP with Zend Framework. The server is IIS. Both applications are based on SOAP solutions.
Would appreciate any tips. Will provide additional information if needed.
Thank you in advance for any help.
You have a misunderstanding here. If you call services from localhost (i.e. via Zend_HTTP_Client), this means you are using the tcp/ip and http layers here. Everything works via sockets, no matter if localhost or external ip address.
If the other application needs to be accessible "from the outside" (no integration possible) you can imho only speed up by using a faster webserver (e.g. nginx), turning off modules in your webserver that you don't need or writing your own socket server, dismissing a whole lot of the processing apache and nginx do. http://devzone.zend.com/209/writing-socket-servers-in-php may help you with your first steps.
I will be developing an application for a club where they will have visitors use biometric systems(finger print) or magnetic cards to mark their attendance.
This application is planned as a web app, made using PHP/MySQL/Javascript. It does a lot of other things as well.
What I wanted to find out is how is the interface between Biometric/Megnetic Card systems to a web app done?
I've never worked on this and am hoping if someone who has experience with this can throw some light on this as to how this could be accomplished. Any pointers will be appreciated.
Also, we will be hosting this application remotely. So we won't have physical access to the web-server.
(I'm afraid there's got to be some form of application installed on a pc that would interface with the hardware and probably makes calls to the web app. But, if there was a way to connect it to the web app directly, then the app would be easy to deploy to any location with minimal installation.)
Thanks! :)
How is the hardware connected? Directly to the app server, or to a standalone box of some sort? If it's a standalone box, then yes you'll have to have some kind of program on the box collect the data and send it to the web app. If it's connected directly to the app server, then you need to write something that either polls the hardware or receives messages/events from the driver and DTRT wrt the web app.
There three type of communication channels are supported by the biometric machines. They are serial, TCP/IP and HTTP. In your case, you either need to implement TCP/IP or HTTP.
If your application is running the intranet, then you can implement the TCP/IP server application and host at a intranet system which is expected to write in a common database between the biometric application and your web-app.
If your application is hosted at the server, then you can implement the http server and associated the URL with the machine. For this you need buy the http api supported machines. Here is the google search link for finding such machines.
Normally, I would use a PHP webservice to do this, but since the front-end is hosted on a linux box, I need another way to do this (so I don't have to go through the trouble of installing FreeTDS, etc. I will if I have to).
Is there a better way to do this? I'm not a web guy, but I'm trying my best.
If the web host is a linux box then you will have to write a web service as a go between for Silverlight and SQL Server. If you have an ASP.NET capable web host available you can host the Silverlight in a RIA Services project, which gives you more direct access (still not actually direct, but you treat it very similarly).
Not really - especialyl given that for security reasons silverlight apps should only connect back to the host they were loaded from (otherwise a silverlight app could be abused as network scanner etc.).