Tuesday, January 3, 2012

Implementing an async servlet in Scala with less than 10 lines of code

What is async servlet and why use it? An async servlet can tell the container that it will handle a request later, so that the current thread is not tied up and can be used for another request. Why is it useful? For example, if the request is going to do something relatively time consuming (e.g., generating a report from a DB), instead of tying up the thread doing nothing waiting for the data, you can free up the thread for other requests and wait for the data in a background thread. When the data is received, you can complete the response.


It turns out to be very easy to implement an async servlet (new in servlet 3.0). In Scala, it took less than 10 lines of code (not counting brace brackets and imports):



package com.ttdev.webqos
import javax.servlet.http._
import java.util.concurrent._

class MyServlet extends HttpServlet {
private val pool = Executors.newFixedThreadPool(1) // create a thread pool

override def doGet(req: HttpServletRequest, res: HttpServletResponse) {
req.startAsync //tel the container the response will be generated later
pool.execute(new Runnable() { // add the runnable to the queue
def run { // when a thread in pool picks up the runntable...
res.getWriter.append("hello world!") // write the response
req.getAsyncContext.complete // tell the container that the response is done
}
})
// return without a response yet
}

}