Does the PI System Input Adapter Get Its Data from the PI Server or the Inteface Node?

Discussion created by richard Employee on Apr 23, 2010

In the first release due out by June 2010, the PI System Input Adapter gets its data streams from the PI Update Manager, which, in this release is still only available on a PI Server. Note: the PI System Input Adapter can run on any node that has access to the PI Server; it does not have to run on the PI Server.


There are two exciting developments coming in the PI System for optimized "Edge" data streams.


First, PINS will get a makeover in the next year which will include a version of PI Update Manager running on the PINS Node (Interface Node). Once this is in place, if the source of a data stream (tag) is local (collected on the PINS node where the query is running) the stream will be picked up locally without a trip to the server's PI Update Manager. This will not only lower the latency but will improve reliability, enabling queries where all the data is local to continue to run regardless of the state of the connection to the server.


Second and even more "interesting" is that our next generation "Interface" infrastructure, Codename "Galileo", will support streaming of "raw" data collected from our interfaces to be streamed directly to StreamInsight via our "Galileo" Adapter. This will allow queries to be written on the data our interfaces obtain before that data is transformed to PI data types (tags.)


So, we have a very solid, high performance solution rolling out in June which we expect will enable and benefit many solution scenarios. And the future is looking bright indeed.