 |
In the early days of embedded computing, data management was simply treated as part of the application; developers just used the data structure capabilities built into their programming language to manage their data. Embedded devices have grown geometrically in complexity, and this approach now falls woefully short of the requirement. Modern embedded devices, such as those found throughout the communications infrastructure or in factory automation systems, often need to process and store many megabytes of data. Examples include routing tables, encryption keys, configuration information, and physical data of many different kinds. What’s more, due to the distributed nature of many systems, critical data now often reside on multiple nodes throughout a complete system, and the system’s reliability depends on the integrity of this distributed data. Clearly, the time has arrived for a much more advanced approach to managing data within modern embedded devices.
The right approach for managing data is to use a toolkit that provides fast, reliable, scalable, and mission-critical level support, including full support for transactions and replication. For maximum performance, an in-process execution model, supported by a simple function-call API, is appropriate. Administration overhead should be kept to an absolute minimum.
For the ultimate in reliable file systems, we recommend Reliance from Datalight, and for managing flash memory, we recommend their FlashFX product.
|
