1

I have an interesting question regarding Java and networks.

I am creating a basic chatting program, which is only for learning purposes. Now, I have properly connected a client to a server before and sent text back and forth. That's perfectly fine. However I am running into a problem.

Say I have a database on a remote server somewhere. That database is going to contain login information, contact list information, etc for the instant messaging program. This means that I am going to need to perform queries on the database when the client does something. Obviously I know that the client should not contain the connection string or have any access what-so-ever to the database.

My question is: what is the best way to have a client-server architecture for chatting purposes with multiple potential clients, but also have a method of sending data over the network so that queries can be made on the server side?

It may be a simple answer, but for some reason I am having trouble wrapping my head around multiple sockets/serversockets sending data at the same time.

Thanks.

Improve this question

3 Answers 3

Reset to default
1

one solution is to set up tomcat and use java servlets, although that is technically http requests.

Sounds like you need to write a server. I think that all your chatting can go through on server, so its not a p2p architecture, but rather a hub-style architecture. That way your server can access the database on behalf of the client.

I hope this answers your question, there are a million different approaches when it comes to network programming :)

Example XML message as per comments:

<XML>
<msg_header type="chat"/>
<msg_body>
   hello world
</msg_body>
</XML>

<XML>
<msg_header type="query"/>
<msg_body>
   SELECT * FROM myTable
</msg_body>
</XML>

You don't even have to go that complex if you don't want. Or you can adopt a widely used standard like HTTP (Which uses TCP/IP which uses sockets) - that way you can use one of Java's built in http processors to get the job done. Its up to you :)

Improve this answer
Sign up to request clarification or add additional context in comments.

2 Comments

As I mentioned above, I do plan on using a server to receive the requests. I'm mostly just confused about having two different things going on between a client and the server (the client is sending chat text to the server, but can also send data for sql queries. and i would also like the client to be able to constantly receive query results).
That usually comes down to establishing a protocol. I usually use XML and serialize it to binary before i send it over the socket. Then, the reader can check for an XML attribute like "message type" and decide what to do with it. Check out http/1.1 headers and look at their protocol, maybe you can invent a light weight version of that?
0

I was using TCP connections (mostly in games and chat applications) like yourself some time ago, but in my opinion it doesn't suits what your asking for very well, except for games. I'm very fond of webservices atm. These communicate over HTTP instead of TCP and uses GET, POST and PUT methods to acknowledge what your trying to do. It's pretty simple to put up a running java webservice yourself and try it out. Try using XML or Json with it for your application.

Great tutorial on a jersey webservice: http://www.vogella.com/articles/REST/article.html#

Later on when you got a grasp of it, you can easily tie a database connection with your webservice the same with as you do normally.

Good luck!

Improve this answer

Comments

0

There are multiple approaches you can take here:

  1. If you're using a dedicated server-side thread per client, this thread performs the DB queries needed. The client passes in the initial login info etc; after validating these, the handler thread can send back the contact list and other info needed by the client.

  2. Another approach is to not have a dedicated thread. Instead, when the client authenticates, create some kind of token object and pass it onto the client. Everytime the client sends a message, it sends the token along. Whichever thread on the server receives the message+token can validate the client by just inspecting the token.

There's no one right answer here; I'd suggest you take a look at the architecture of some popular chat implementations such as IRC and jabber if you want to explore further.

Improve this answer

5 Comments

I am interested in the first option- it seems to be what I was going for. So basically, each connection the server makes with a client should spawn two threads (1 for chat, 1 for queries). The chat thread would be used for the client to send/receive data. the query thread would be used for sending queries / receiving the query results?
No, you'd just need one thread on the server for a client. Your queries are things like retrieving the contact list. Your client program can periodically send a message asking for the contact list in between the user sending his/her chat messages. Define different message formats for this.
This makes sense, thank you very much. Every time a client connects to the server, a thread is spawned and the messages are dealt with within that thread. I'm still confused about how the client and server will distinguish between sql queries and normal text. If the client sends chat text to the server, it will be sent over through the buffer and received by the server. Should the server have code that checks what type of text it is (if it's in a certain format or something) to see what type of text it is? I think this is the last think I'm confused about here.
I'm still not sure why you want the client would send sql queries to the server. The client should ask for data such as contact list etc. - how that is stored, queried are all things you want your server threads to worry about. As for the latter, define different message types, one for user chat messages, another for contact list requests, another for contact addition requests and so on. Your server thread, of course, has to inspect the message format and act on it.
Oh, ok. For logging in, the client needs to send the server the username and password so that the server thread can query the database and make sure it is all correct. I guess technically the client isn't sending sql queries, just parameters that will be put into sql queries on the server end. So technically, the client does have to send some data over that isn't chat text. The client will have to send over text with a different message type that contains the username and password so that the server can perform the query on the database, right?

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.