The key to performance is understanding...
Many of our customers have benefitted from CustardWare when we have helped to diagnose problems with systems and networks.
Over the years we have worked with many customers to help them isolate and rectify faults in a wide variety of systems and applications. These range from simple networks, through high performance networked systems and into digital broadcast networks.
Throughout this period we realised that there were a specific group of tools we would turn to or diagnostic information. Some of these were open source, others we developed ourselves. The common features were:
- Regular collection of status from a variety of sources and devices
- Frequently a need to regularly execute some automated test
- The development of custom code to analyse some event
- The storage of a large volume of statistical information
- Highly resilient operation that can be left unattended for months or years
- Ability to display information in a variety of simple ways
CustardWare has effectively evolved out of four generations of development. This started with tools configured or written for one of use and to a specific need or goal. From there we developed a generic solution (Reithmon) that could manage a wide range of tools making deployment easier for a given customer. This then evolved into a more generic solution (Normon) that provided most of the features needed to handle a growing range of customers with a great diversity of monitoring or diagnostic challenges. Finally a redesign of the architecture allowed us to produce a scalable architecture able to manage many thousands of devices within a high performance, high availability package.
This gives us the ability to use datamining concepts to gather diagnostic information across a network or collection of systems to rapidly assimilate issues that may lead to sub optimal performance or failure.