* Simplifies the implementation of telemetry buffering in Open MCT.
* Switches from per-parameter buffering to a shared queue. Per-parameter buffering is too easily overwhelmed by bursts of telemetry from high-frequency parameters. A single shared buffer has more overhead to grow when a burst arrives without overflowing, because the buffer is shared across all parameters.
* Removes the need for plugins to pass serialized code to a worker.
* Switched to a "one-size-fits-all" batching strategy removing the need for plugins to define a batching strategy at all.
* Captures buffering statistics for display in the PerformanceIndicator.
