Abstract: A kernel provides a mechanism allowing various applications (e.g. CRM, logistics, procurement, etc.) hosted on a plurality of different application servers, to share access to a common underlying system (e.g. database). An automated process for implementing a kernel change (e.g. for upgrade or replacement) may employ a “Stop-the-World” approach involving suspension of application server instances, coordinated by the start service of the last application server whose kernel is to be changed. As used herein, suspending refers to halting any processing prior to calling a central service (CS) as long as a CS instance is unavailable, and then to resuming processing once the CS instance becomes available. This suspension of relevant clients avoids errors from arising during the downtime of the CS instance. Once the application server instances are in a stopped state, the CS instance is restarted. Then, the application server instances are resumed, and their kernels are changed.