Using Keep-Alive to retain connection for duration

[robertybob]

Hi.

I am building an app that acquires data from a web service via HttpURLConnecti on. The app requests data every minute continuously whilst it is open and activated.

At the moment I am opening and closing the connection each time it is required but is it better practise to have a single connection that persists?

The current code is

URL url = new URL(tradeURL);
connection = (HttpURLConnection)url.openConnection(Proxy.NO_PROXY);
connection.setRequestProperty("Content-Type", "application/json");
connection.setRequestProperty("Content-Length", Integer.toString(urlParameters.getBytes().length));
connection.setUseCaches(false);
connection.setDoOutput(true);
DataOutputStream wr = new DataOutputStream(connection.getOutputStream());
wr.writeBytes(urlParameters);
wr.close();
InputStream is = connection.getInputStream();
BufferedReader rd = new BufferedReader(new InputStreamReader(is));
StringBuilder response = new StringBuilder();
String line;
while((line = rd.readLine()) != null) {
    response.append(line);
    response.append('\r');
}
rd.close();
connection.disconnect();

If it is better to re-use the connection from the initial request, how would I go about doing that?

Also, is the keep-alive related to the domain level only - ie api.abcd.com - or is the entire url path involved?

Many thanks!

[chaarmann]

Use a connection pool!

Your problem is not only that a connection is kept alive even if it is not used anymore, but also that the maximum number of connections is limited.
This is all managed by a connection pool. (I have written my own connection pool many years ago, but now there are a lot of ready-made connection pools out for free usage. Just google for that.)
Instead of opening and closing connections every time which is very costly, a connection pool will re-use existing connections.

[robertybob]

Many thanks Chaarmann - I'm investigating this and will let you know.

[robertybob]

Sorry - had to work on something else for a while and forgot all about this!

For my needs I managed to get good enough stability using the generic keep-alive functionality and tweaks to the requests but Chaarmann's method seems to be decent solution for more complex requirements.

Many thanks.