The standard implementation of this mechanism has no means of setting the bounds for the duration of that pause. One can imagine the potential consequences if such a pause happened just as an automated trading system was exploiting a fleeting arbitrage opportunity or a defense system was countering incoming missiles. Only through a fundamental innovation within the Java virtual machine itself—as deep in the software stack as you can get before entering the operating system—can this lack of predictability be addressed; rendering Java viable for applications with real-time needs such as in these examples.
The point is not simply that innovation happens and makes applications stronger and faster. Rather, it’s a reminder that many of the most critical innovations are happening way down the stack, in areas that may be too easily neglected in a myopic focus on the top layers as the most fertile ground for innovation. By all means, leverage the benefits of flexible loose-coupling and efficient component reuse that SOA and BPM provide. But don’t forget that all of that coupling and reuse will be happening to assets that need to run somewhere. In that somewhere, the supply of innovation and differentiation is far from depleted. The application server has a long life ahead of it.
Mike Piech is senior director at Oracle with responsibility for Oracle Fusion Middleware products: Oracle WebLogic, Oracle Coherence, Oracle JRockit, and Oracle Tuxedo. Mike joined Oracle as part of the BEA acquisition, prior to which he spent seven years running product marketing at Dorado Corporation, which builds a WebLogic-based cloud solution for mortgage banking.