
MANAGEABILITY 191
Reader B could all happen quickly enough for Reader B to step in and read the next item
on the conveyor belt.
Even a distributed system has its weaknesses, though. A distributed system, like a central-
ized system, works best in a static environment where the same edge devices report virtu-
ally the same things day after day. Adding a new sensor may be fairly simple, requiring
only the physical installation and a call to IT operations to update configuration settings,
but adding a reader from a different vendor or even a new model from the same vendor
may be just as complex in a distributed system as in a centralized system, requiring days or
weeks of customization. Adding a completely new type of sensor, such as a Real Time
Location System (RTLS), could require a dedicated software development project.
Self-Organizing Deployment
The self-organizing approach is a radical departure from traditional deployments. In this
configuration, almost all of the computing power moves to the edge, with the data center
no longer central to the system but merely another participant in a mesh network. (For
more about mesh networks, see Chapter 11.) Figure 9-4 is an example of a self-organizing
configuration.
Unlike in centralized or distributed deployments, new devices added to a self-organizing
deployment are themselves responsible for announcing their capabilities and providing
software components ...