1. 4 Application Layer Protocols
• Learn the advantages of using OPC
• Understand the OPC Client/ Server architecture
• Review the milestones in the history of OPC
• Understand how to create a connection between an
OPC Client and OPC Server
• Recognize what factors can affect OPC communication
• Review the most widely implemented OPC
specification: OPC DA
4 Application Layer Protocols
4.2 Open Process Control (OPC)

2. OPC: Open Process Control
Source:

3. OPC: Open Process Control
Manufacturer-independent application programming interface (API) for automation
To implement clients which can access plant data coming from remote devices (PLCs, field bus devices, real-time databases) easily
Set of commands collected in a software library (DLL) to access OPC servers
OPC clients
read and write process variables, read alarms and events and acknowledge alarms, and retrieve historical data from data bases according to several criteria
Implemented on automation platforms (e.g. ABB Ax800), which may act themselves as an OPC server to publish their data, events and historical data.
OPC server
Supplied by manufacturer of automation devices supplies
Communicates with its devices through a proprietary protocol
Manages several devices of the same type, several servers can run in parallel and each server can be accessed by several clients in the same network

4. Usage and specifications Clients and Servers: configuration
OPC Overview
4.2.1 OPC Overview
Usage and specifications
Clients and Servers: configuration
4.2.2 OPC Data Access
Objects, Types and properties
Communication model
4.2.3 OPC Alarms and Events Specification
Events
Alarm Conditions
4.2.4 OPC Historical Data Specification
Overview

5. API covered by the OPC standard
4.2.1 What is OPC ?
OPC (formerly: "OLE1 for Process Control", now: "Open Process Control") is
an industry standard set up by the OPC Foundation ( specifying the software interface (objects, methods) to a server that collects data produced by field devices and programmable logic controllers.
API covered by the OPC standard
application
(OPC client)
servers
OPC server X
OPC server
(simulator)
OPC server Y
Ethernet :
Field bus
Manufacturer-independent application programming interface (API
- implement clients accessing plant data coming from remote devices (PLCs, field bus devices, real-time databases) easily
- set of commands collected in a software library (DLL) to access OPC servers
OPC clients
read and write process variables, read A&E, and retrieve historical data
Implemented on automation platforms (e.g. ABB Ax800
OPC server
Supplied by manufacturer of automation devices supplies
Communicates with its devices through a proprietary protocol
Manages several devices of the same type
Several servers can run in parallel and each server can be accessed by several clients in the same network
PLCs Brand X
PLCs Brand Y
Sensors/Actors
1) OLE (Object Linking and Embedding) is a Microsoft technology for connecting software components.
It has since been extended by the COM / DCOM technology. It corresponds to Java Beans.

6. Before OPC visualization history data base ABB PLCs Télémécanique PLCs
MasterBus
MMS driver
Profinet
driver
XWAY
driver
ABB PLCs
Télémécanique PLCs
Siemens PLCs

7. (Information Manager)
With OPC
Operator
Historian
(Information Manager)
application software is written independently from the type of controller
the drivers still exist,
but the clients do not
see them anymore
ABB OPC server
Schneider OPC server
Siemens OPC server
MMS
XWAY
ProfiNet
ABB AC800M
Télémécanique TSX
Siemens S7

8. Greatest improvement in automation since IEC 61131.
Importance of OPC
Greatest improvement in automation since IEC
More than 150 vendors offer OPC servers to connect their PLCs, field bus devices, displays and visualization systems.
Used for data exchange between applications and for accessing databases
DLL for Automation Interface (VB,..) and Custom applications (C++,..)
Three major components: 1) OPC - DA = Data Access (widespread, mature) 2) OPC - AE = Alarms and Events 3) OPC - HDA = Historical Data Access
… and some profiles* (batch,…) * A “profile” is a subset or a specialization of a standard to form a stricter standard, adapted to an application. E.g. 100 Mbit/s, full-duplex, fibre-optical Ethernet is a profile of IEEE

9. Specification 1: OPC DA for Data Access
Process variables describe the plant's state, they are generated by the sensors or calculated in the programmable logic controllers (PLCs).
Process variables can be sent upon a change, on demand or when a given time elapsed.
The OPC DA (Data Access) specification addresses collecting Process Variables. The main clients of OPC DA are visualization and (soft-) control.

10. Specification 2: OPC AE for Alarms and Events
Events are changes in the process that need to be logged, such as "production start"
Alarms are abnormal states in the process that require attention, such as "low oil pressure"
OPC AE (Alarms and Events) specifies how alarms and events are subscribed, under which conditions they are filtered and sent with their associated messages. The main clients of OPC AE are the Alarms and Event loggers.
determine the exact time of change (time stamping)
categorize by priorities
log for further use
acknowledge alarms (events are not acknowledged)
link to clear text explanation

11. Specification 3: HDA for Historical Data Access
Historical Data are process states and events such as: process variables, operator actions, recorded alarms,... that are stored as logs in a long-term storage for later analysis.
OPC HDA (Historical Data Access) specifies how historical data are retrieved from the logs in the long-term storage, filtered and aggregated (e.g. compute averages, peaks).
The main client of OPC HDA are Trend Displays and Historians.

12. Beyond Microsoft: OPC UA
In a move to get more independence from Microsoft and use web technology,
a new specification called " Unified Architecture" (formerly. OPC XML) that uses web services for all kinds of transactions: query, read, write, subscribe,...
The classical OPC DA, AE and HDA are implemented with XML / SOAP / WSDL
this allows encryption and authentication of process data.
OPC UA does not only standardize the interfaces, but also the transmitted data.

13. Usage and specifications
Client and Servers
4.2.1 OPC Overview
Usage and specifications
Clients and Servers: configuration and communication
4.2.2 OPC Data Access
Objects, Types and properties
Communication model
4.2.3 OPC Alarms and Events Specification
Events
Alarm Conditions
Automation Interface
4.2.4 OPC Historical Data Specification
Overview

14. Server(s) and Client(s) in the same node
client application
(OPC client)
client application
(OPC client)
OPC server
OPC server
OPC server
devices
devices
devices
devices
Clients and servers run as parallel processes
The OPC specification defines the interface between client and server in the form
of objects and methods.

15. OPC for internal communication: ABB’s SCADA (800xA) as example
ABB's Operator Workplace (800xA) is at the same time OPC server and OPC client. Software components (agents) within AIP expose their properties as OPC objects.
Internal (within the PC) and external communication (between PCs) takes place over OPC.
800xA
aspects
aspects
Enterprise
Historian
Asset
Optimizer
functions
OPC server
Windows PC
process
data base
OPC
connections
OPC client
ABB OPC server
Schneider OPC server
Siemens OPC server

16. Direct and Fieldbus access
direct connection
fieldbus connection
client application
(OPC client)
client application
(OPC client)
(local)
OPC server
(local)
OPC server
FB Manager
fieldbus
can also be a point-to-point link
proprietary protocol
I/O devices
fieldbus
fieldbus
The OPC server is running
all the time, even if no
client is present
FB agent
FB agent
PLC
PLC

17. Accessing a server in another node
client application
(OPC client)
stub
DCOM
TCP/IP
DCOM
TCP/IP
TCP/IP
DCOM
DCOM
OPC
server
OPC server
Limitation:
does not work over firewalls.
Solution:
Tunneller
OPC UA
FB Manager
fieldbus

18. What is the objective of OPC ? On which technology does OPC rely ?
Assessment Overview
What is the objective of OPC ?
On which technology does OPC rely ?
What is an OPC Server ?
What do the main OPC specifications describe ?
How can an OPC Server access data on another machine? And how does it work for OPC clients?

19. Usage and specifications Clients and Servers: configuration
OPC Data Access
4.2.1 OPC Overview
Usage and specifications
Clients and Servers: configuration
4.2.2 OPC Data Access
Objects, Types and properties
Communication model
4.2.3 OPC Alarms and Events Specification
Events
Alarm Conditions
Automation Interface
4.2.4 OPC Historical Data Specification
Overview

20. OPC DA: Item properties
The process data are represented by three dynamic properties of an item:
value: numerical or text
time-stamp: time at which this data was transmitted from the PLC to the server UTC, not local time.
quality: validity of the reading (not readable, dubious data, o.k.)
optional static properties
description: a text string describing the use and of the variable (optional)
engineering unit: the unit in which the variable is expressed (optional) ID V Q T D U

21. OPC DA: Objects as viewed by the OPC server
An OPC server is structured as a directory with root, branches and leaves (items)
Process Line 1
Tag Name
Controller 1
Controller 2
TAG
Level_1
An item is identified by its "fully qualified ItemID", e.g.
"Process_Line_1.Controller_2.Level_2"
Controller_3.Prog_1
TAG
Level_2
Controller_3.Prog_2
TAG
Ramp4
Cell 1
Machine 2
Branches may contain other branches and items
The structure may also be flat instead of hierarchical
This structure is defined during engineering of the attached devices and sensor/actors.
(Intelligent servers can configure themselves by reading the attached devices)

22. OPC DA: Objects as viewed by the OPC client
A client builds its own hierarchy, using the server’s hierarchical view.
Items in the server are defined by the programmer of the PLC
A full-fledged PLC may export some 10’000 items, a client needs only a subset.
A client builds groups, populating them with items it is interested in.
Items of a group are expected to have similar real-time requirements
Groups are not hierarchical, but flat.

23. OPC DA: Mapping items to groups
Each client structures its items by groups, independently from the server.
Initially, the client browses the server structure to check if the items it is interested in exist.
A client registers its groups and items at the server.
The server keeps the structure of all its clients.
clients
Client1
Client2
GroupX
GroupZ
Item1
Item2
Item3
Item1
Item2
Server root
server
Area 1
Oven_1
TAG
Temperature
TAG
Heat_On
Tank_1
TAG
Level
Area 2
TAG
Empty_Valve
Fill_Valve
Area 51
TAG

24. OPC DA: Communication Model
4.2.1 OPC Common
Overview: usage and specifications
Clients and Servers: configuration
4.2.2 OPC Data Access
Objects, Types and properties
Communication model
4.2.3 OPC Alarms and Events Specification
Events
Alarm Conditions
Automation Interface
4.2.4 OPC Historical Data Specification
Overview

25. 4.2.2 OPC DA: Example of access to a variable
controller program
OPC application
Reactor_1.Program2
ReadItem ("OPC:Reactor1: Program2.MotorSpeed")
MW%1003
MotorSpeed
Value: 112
MW%1004
Temperature … ….
load
symbol
table
symbols
OPC
server
code
Get () , MW%1003)
Return (MW%1003, 112)
Network
Program 2
Reactor_1
Marker: MW%1003

26. OPC DA: Read Communication Models (per group)
synchronous
asynchronous
client
server
client
server
Call
Call
myGroup.SynchRead()
myGroup.AsyncRead()
Reply
Reply
myGroup_AsyncReadComplete()
client
server
Subscribe
myGroup.IsSubscribed
on change
("subscription-based")
The OPC interface accesses only groups, not individual items.
Notify
myGroup_DataChange()
Notify
myGroup_DataChange()

27. OPC DA: Write Communication Models (per group)
client
server
client
server
Call
Call
myGroup.SynchWrite()
myGroup.AsyncWrite()
Reply
Reply
myGroup_AsyncWriteComplete()
The OPC interface accesses only groups, not individual items.

28. OPC DA: communication paradigm
OPC DA works according to the “shared memory” paradigm.
This means that a newer value overwrites the older one, no queues or history are kept.
The server does not guarantee that different clients see the same snapshot of the plant.
The server does not guarantee that all changes to variables are registered, changes may be missed if the polling period is too long.
OPC DA Client
OPC DA Client
OPC DA Server

29. What are the DA the read and write operations ?
OPC DA: Assessment
How does the OPC server know a) where to fetch an item? b) which items belong to which group?
What are the DA the read and write operations ?
Is communication done by items, by groups or by collection of groups ? Why?
Can a change of an OPC variable be notified as an event, or shall the client poll ?
What are the implications of the shared memory paradigm for the application developer?
a) Loads symbol table from PLC b) it keeps a record of each client and its groups
Read: sync, async, on change Write: sync, async
Client asks for groups (can define what belongs to them by itself), server depends on proprietary protocol
Can be done event-based, depends on application
Developer must no rely on assumption of consistent snapshot of plant

30. Usage and specifications Clients and Servers: configuration
OPC Alarms and Events
4.2.1 OPC Overview
Usage and specifications
Clients and Servers: configuration
4.2.2 OPC Data Access
Objects, Types and properties
Communication model
4.2.3 OPC Alarms and Events Specification
Events
Alarm Conditions
4.2.4 OPC Historical Data Specification
Overview

31. Alarms and Events: Purpose
The controllers (PLC) generate events in response to changes in the plant variables.
together with their precise time of occurrence, type, severity and associated message for the human operator.
An OPC AE server registers these events and makes them available to several clients
Alarms are described as state machines and may require acknowledgement.
The OPC Alarms & Events Interface gives access to the AE server, allowing to:
- browse the OPC AE Server for predefined events.
- enable or disable alarms and events
- subscribe to alarms and events of interest
- receive the event and alarm notifications with the associated attributes
- acknowledge alarms

32. AE: Definitions
An event is a general change of state that is relevant to the OPC server.
An event signal a change:
1) in the field device ("production started")
2) in the OPC server ("alarm acknowledged")
3) in the application ("operator action")
An alarm indicates a state of the process that requires attention and is relevant to the OPC server. An alarm is represented by an alarm condition, which is a state machine determining if the alarm has been enabled, triggered or acknowledged. An alarm rises several events.
An event or an alarm does not transmit process values, but boolean information indicating a change of state, its originator, the time of its occurrence and a message intended for a human operator.
Alarms and events may not get lost (contrarily to OPC DA, which does not guarantee completeness)
Alarms and event are precisely time-stamped by their source, (contrarily to process variables, which are time-stamped by the receiving OPC server).

33. AE: communication paradigm
OPC AE works according to the “message passing” paradigm, contrarily to OPC DA, that works according to the "shared memory" paradigm.
This means that an event is kept in a queue until all clients have read it (or timed out).
The AE server guarantees that different clients will see all events in the same sequence.
OPC AE Client
OPC AE Client
OPC AE Server
12:
12:

34. AE: Displaying Alarms and Events
Alarms and events are usually displayed differently on an operator screen.
- Events are displayed in an event list that can become quite long (typically 1000 entries), entries are not cleared when the source of the event returns to normal
- Alarms are displayed in a short list (typically 50 alarms) appearance changes when the alarm is acknowledged, an alarm line is cleared when the alarm signal is cleared (but remains in the log).
Ack checkbox

35. Usage and specifications Clients and Servers: configuration
AE: Events
4.2.1 OPC Overview
Usage and specifications
Clients and Servers: configuration
4.2.2 OPC Data Access
Objects, Types and properties
Communication model
4.2.3 OPC Alarms and Events Specification
Events
Alarm Conditions
4.2.4 OPC Historical Data Specification
Overview

36. AE: Events kinds
OPC AE defines three kinds of events:
simple: process control system related events (change of a boolean variable)
condition-related: notifies a change of an alarm condition (CLEARED, ACKNOWLEDGED), (see later)
tracking-related: origin outside of the process (e.g. operator intervention)

37. AE: Event- identification
An event is identified by
- its source (the object that generates the event. e.g. Tank1) and
- the event name (which can be the same as in another object, e.g. HiLevelCond)
HiLevelCond
HiLevelCond
event
event
Function Block
LoLevelCond
LoLevelCond
event
event
event name
Tank1
Tank2
event signal (boolean expression)
is an external signal to be used ?(boolean)
signal name for external signal (20 characters)
name of the source (30 characters)
message (60 characters)

38. AE: Events - Notification
Tank1LevelHigh_SimpleEvent (source, timestamp, message, severity, category)
AE Client
specified communication
COM/DCOM
OPC AE Server
queue
unspecified communication network, fieldbus or internal bus
event notification
message
Controller
timestamp
Event FB
Tank1
Plant
Level Switch

39. AE: Events - Time Stamp
There are three places where events can be time-stamped:
- at the device that originally produced the data (external event - low-level event) allowing Sequence-Of-Events with a high accuracy, down to microseconds
- at the controller, (internal event) using the controller's clock to time-stamp messages giving accuracy not greater than the period of the tasks, about 1 ms.
- at the OPC Server, when an event message arrives (tracking events) not more accurate than DA, about 10 ms)
The OPC server can be configured to register the time stamp at the instant of the event transition (positive or negative) and the instant of the acknowledgement.

40. Overview: usage and specifications Clients and Servers: configuration
AE: Alarm conditions
4.2.1 OPC Common
Overview: usage and specifications
Clients and Servers: configuration
4.2.2 OPC Data Access
Objects, Types and properties
Communication model
4.2.3 OPC Alarms and Events Specification
Events
Alarm Conditions
4.2.4 OPC Historical Data Specification
Overview

41. AE: Alarms - Condition Definition
An (alarm) condition is described in a named state machine
The condition state is defined by three variables:
Enabled: the condition is allowed to send event notifications
Active: the alarm signal is true
Acknowledged: the alarm has been acknowledged
Alarm signal (e.g. FIC101.PV > 100 AND FIC101.PV < 150)
Condition
Acknowledgement signal (a positive transition of a boolean variable)
Condition state
Enable (positive transition) Disable (positive transition)

42. AE: Alarms - Condition states and acknowledgement
event notification
alarm signal
acknowledgement
Inactive Acked
Active Unacked
Active Acked
Inactive Acked
Active Unacked
Inactive Unacked
Inactive Unacked
Inactive Acked
condition state
alarm_signal 
condition state transitions
(here: always enabled)
Enabled Inactive Acked
alarm_signal 
EnabledActive Unacked
EnabledActive Acked
Ack 
Ack 
alarm_signal 
Enabled Inactive Unacked
alarm_signal 
An event is generated each time the alarm signal changes state, or is acknowledged

43. AE: Alarms - Acknowledgement
An alarm condition becomes active when the PLC produces an alarm signal describing an abnormal state defined by the application (e.g. the level of the tank is too high).
The operator is expected to acknowledge this condition (client ack) Alternatively, a local operator may use a button or HMI that the PLC reads (field ack)
event notification
Tank1Level_ConditionEvent
AE Client
client ack (acknowledger ID)
COM / DCOM
OPC AE Server
Network, field bus, or internal bus
message
Condition
time- stamp
LevelHigh
controller
AckButton (field ack)
Alarm Signal

44. AE: Summary alarms and events
AE Client
AE Client
COM / DCOM
OPC AE Server
OPC AE Server
event notification (source, timestamp, message)
alarm notification (source, timestamp, message, condition, subcondition, severity, type)
message
Condition
timestamp
Event FB
message
controller
controller
event
alarm
ack

45. What is the difference between Alarms and Events?
OPC A&E: Assessment
What is the difference between Alarms and Events?
Where are Alarms and Events time stamped?
How does the “message passing” paradigm influence the OPC client application developer?
E: state change, A state that can trigger event notifications
Depends, accuracy varies
Ensures the same sequence of events is seen by all clients

46. Usage and specifications Clients and Servers: configuration
OPC Common Overview
4.2.1 OPC Overview
Usage and specifications
Clients and Servers: configuration
4.2.2 OPC Data Access
Objects, Types and properties
Communication model
4.2.3 OPC Alarms and Events Specification
Events
Alarm Conditions
4.2.4 OPC Historical Data Specification
Overview

47. GE Fanuc/Intellution iHistorian (iFix...)
Historian Example
GE Fanuc/Intellution iHistorian (iFix...)
Questions to the historian:
What was the value of FIC101 last week ?
What was the flow average during October ?
Which were the daily averages in October ?
What was the total flow in each month ?
How much fuel did we use for the batch ?
Give the answers in form of tables,
pie diagrams, spreadsheet, reports
Features:
Unlimited Point Collection Sub-Second Data Collection Rates Enhanced Data Compression True Thin Client Administration Fault Tolerant Architecture
Data collection, archiving and retrieval
Report generation
Computations (e.g. VBScript)
Secure access (FDA 21 CFR 11)
20'000 actions/ s, Up to 100'000 data points

48. HDA: Historical Data Access
e.g. Trend Analysis
e.g. Event Logger
HDA Clients
independent processes
OPC HDA Server
hidden
calculations
history
database
raw and ordered data
collector
OPC DA Server
proprietary data acquisition
Field device
Field device

49. HDA: Purpose
An OPC HDA server gives access to a historical data base (logs) in which data from the process have been collected and time-stamped, possibly through an OPC DA interface.
The OPC HDA interface clients, such as trend analysis, product tracking or data mining,
that require ordered access these data logs.
The OPC HDA interface allows to:
- browse the historical data base
- retrieve data through proper filtering, e.g. by date range, by identity, by property
- build aggregates over the retrieved data, such as average, minimum, maximum.
- enter new entries, correct entries or remove entries
- enter / delete annotations in the history data base

50. Data in the historical database are identified by their
HDA: Raw log
:40
Gpcpt2ofpbonne
4824
By definition, values are registered when they change. (even if data are acquired by periodic polling)
:40
Cpt2bac 50
:40
Gpcpt2bac 70
:40
Gpcptbe2 45
:41
Gpcpt1bac
151
:41
Gpcpt1ofpbonne
4826
Data in the historical database are identified by their
itemID (here, represented by their name),
value, (of the respective type)
quality (good, stale, bad), and
timestamp (UTC).
:41
Gpcptae2 45
:41
Cpt1bac 49
:41
Gpdefr2 64
:41
Gpvoydef 2
:41
Gpr3tempscycleprd
318
:42
Gpstn1e1 16
:42
Gpalarme1
:42
Gpalarme2
:43
Gpetatmodemarche 2
:43
Gptpscycle
1346
:43
Gpetatmodemarche 1
:43
Gpdefgene1 16
:43
Gpetatmodemarche
:43
Gptpscycle
317
:43
Gpdefr2
:43
Gpvoydef
:43
Gpdefgene1
:44
Gpetatmodemarche 1
:44
Gpr2tempscycleprd
1992
:44
Gptpscycle
435
:44
Gpalarme3 1
:44
Gpalarme4 1
:44
Gpalarme3
:44
Gpcpt2ofpbonne
4823

51. HDA: How to reduce raw data (even before HDA comes into play)
Data sent to the OPC DA server on change, collector records data in a circular buffer log
Since storage capacity is limited, data are reduced by:
if a variable is received more often than the log's minimal storage interval (e.g. 1s), the log keeps the latest of all values of the interval
if (and only if) a received variable changed by more than the "exception deviation" (e.g. 5%, analog values only), it is entered into the log.
if a variable changed by less than the exception deviation, it may be forced into the log after the log's maximum storage interval (e.g. 4 s) elapsed.
min time
max time
max time
max time
max
exception
deviation
min
time
= process value (event)
= log entry
If the time scale of the log is smaller than that of the trend display in this case, values have to be interpolated to be displayed correctly

52. HDA Application: Trend Display
Parameters:
time scale (with possible offset, zoom, pan)
amplitude scale (low range, high range, scale units)
style: smoothed, stepped, filled (several ways to display the same data)
extrapolate: how to display values not received (e.g. because they did not yet change)
log: how were data sampled

53. HDA: Hierarchical logs
long-term log
1 hour_forever
 5.2 MB / Year 1 2 3 27 28 29 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
september
octobre
october
1 mn_7 days log
 6 MB
yesterday
d-2
d-3
d-4
d-5
d-6
d-7
today
1s_24 hours log:
50 12 B
 52 MB 23 24 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
A hierarchical log is built on the data contained in the parent log.
To reduce the log size, several aggregations can be applied:
- record only maximum, minimum, average over a period, etc...
Actual data

54. Toy Examples: Logging Temperature
Logging Temperature Data using OPC
Task Overview
•  Log Temperature Data to Matrikon OPC Server using
•  Create and use Aliases in the OPC Server
•  Read Temperature Data from MatrikonOPC Server
•  Scaling/ConverCng funcConality inside the OPC Server
•  Read Temperature Data from MatrikonOPC Server in a client App

55. Real-Life Examples and Exercise
Integration of redundant GE Mark V turbine controllers at a power plant
The controllers were going to tie into the plant’s main DCS, an ABB Advant control system. Process Portal B and MicroSCADA client applications were used to visualize process data from the turbine controllers.
Exercise: Draw schema of system and describe which OPC specifications are used and how.
Wind Generation
AES Wind Generation manages 7 different wind farms across the United States, five of which comprise more than 500 turbines, with a total generating capacity of over 700 MW. There are six different turbine models, from four different manufacturers.
Exercise: Explain which OPC concepts are useful in this context and how they can be applied to support a SCADA system
1. What is an OPC server? What does it communicate with?
2. What are the services of a OPC DA server?
3. What are the services of a OPC AE server?
4. What is the paradigm difference between OPC DA and OPC A&E?
5. What are the main purposes of the OPC HDA?
and

56. To probe further….
OPC Foundation: Specifications
SoftwareToolbox Examples in Visual Basic
The Code Project OPC and .NET
Matrikon Free client and server:
WinTech Toolkit for an OPC server
NewAge Automation Toolkit for an OPC server