RDBMS Fragility

An RDBMS is quite sensitive to overloading. One of a Database Administrator’s main tasks is to ensure that the server does not stray outside its linear zone, because once that happens, the RDBMS becomes quite unstable.

The graph below shows the results of a load test performed on a popular RDBMS running a well-known Customer Relationship Management application. It graphs query response time against the number of users. The application was originally running a 400-person help desk. We captured the workload and then fed it back into the system with various levels of speed-up, to simulate smaller and larger help desks. You can see that the RDBMS in this configuration became quite unstable under load with more than about 600 users.

RDBMS Instability Under Load

System managers faced with a sharp increase in user workload must plan around this fragility, purchasing additional hardware resources and additional database server licenses. These so-called “fork-lift upgrades” that involve major infrastructure upgrades, are very expensive, time consuming and intrusive into normal operations.

WireCache provides an alternative to this expensive and risk-fraught upgrade process. A WireCache Transparent Database Accelerator can reduce the RDBMS server workload by more than half. This means that you can double the user population without upgrading the RDBMS or its hardware. The WireCache Accelerator will cost you well under half the cost of the equivalent system upgrade. And unlike a system upgrade, WireCache results are guaranteed ahead of time.