Development and Comparison of a Dynamic Hash Method for a Real-time Application
In many web applications the performance depends very strongly on that the workload is evenly distributed upon the underlying server landscape. Several different load balancing strategies exist, e.g. Round Robin or URL-based methods, and work fine for some domains. But if clients must contact exactly those servers responsible for (and exclusive holder of) the considered piece of information then a correct and equable assignment to these servers is necessary. With the help of static hash functions responsible servers can be found only for a fixed server set. If servers should also be added or deleted during runtime then dynamic hash methods must be introduced. And if the application additional is realtime and requested data must be reliable and up-to-date then all these requirements can be met with a new developed dynamic hash method.