PI Interface Node Architecture - White Paper

Document created by srobertson on Mar 7, 2015Last modified by Michael2019 on Nov 4, 2019
Version 4Show Document
  • View in full screen mode

  This document has been replaced with an updated version found here.

The PI System Architecture is composed of multiple OSIsoft applications and solutions.  Among these components are interfaces to other systems such as Control Systems, OPC Servers, Relational Databases, Text Files, etc. OSIsoft has over 400 interfaces available for use in the PI Architecture depending on the customer’s needs. Typically, most of the interfaces are aggregated on an Interface Node (or Nodes if using Interface-level Failover) which is physically separate from the destination PI Server/PI HA (High Availability) Collective.


Although having separate PI Interface Nodes is a commonly accepted practice, the question remains where to position the PI Interface Node(s) in order to maximize the interface’s reliability, performance and security.  The purpose of this white paper is to define the optimal geographical and logical network position of the PI Interface Node(s) and to explain the reasons behind this recommendation.  Any special cases or exceptions to the general solution will be discussed in further sections of this document.


This paper will establish the current recommended location for positioning the PI Interface Node(s) in a geographical and logical network architecture.  This includes, but is not limited to:


  • Defining the best architecture for positioning the PI Interface Node(s) to maximize data availability and minimize data loss in an HA environment
  • Defining the best architecture for positioning the PI Interface Node(s) when interface-level failover is implemented
  • Identifying other considerations when positioning the PI Interface Node(s)